animal research paper pdf

• Open access
• Published: 06 March 2021

Research perspectives on animal health in the era of artificial intelligence

• Pauline Ezanno   ORCID: orcid.org/0000-0002-0034-8950 1 ,
• Sébastien Picault 1 ,
• Gaël Beaunée 1 ,
• Xavier Bailly 2 ,
• Facundo Muñoz 3 ,
• Raphaël Duboz 3 , 4 ,
• Hervé Monod 5 &
• Jean-François Guégan 3 , 6 , 7  

Veterinary Research volume  52 , Article number:  40 ( 2021 ) Cite this article

14k Accesses

30 Citations

10 Altmetric

Metrics details

Leveraging artificial intelligence (AI) approaches in animal health (AH) makes it possible to address highly complex issues such as those encountered in quantitative and predictive epidemiology, animal/human precision-based medicine, or to study host × pathogen interactions. AI may contribute (i) to diagnosis and disease case detection, (ii) to more reliable predictions and reduced errors, (iii) to representing more realistically complex biological systems and rendering computing codes more readable to non-computer scientists, (iv) to speeding-up decisions and improving accuracy in risk analyses, and (v) to better targeted interventions and anticipated negative effects. In turn, challenges in AH may stimulate AI research due to specificity of AH systems, data, constraints, and analytical objectives. Based on a literature review of scientific papers at the interface between AI and AH covering the period 2009–2019, and interviews with French researchers positioned at this interface, the present study explains the main AH areas where various AI approaches are currently mobilised, how it may contribute to renew AH research issues and remove methodological or conceptual barriers. After presenting the possible obstacles and levers, we propose several recommendations to better grasp the challenge represented by the AH/AI interface. With the development of several recent concepts promoting a global and multisectoral perspective in the field of health, AI should contribute to defract the different disciplines in AH towards more transversal and integrative research.

1 Introduction

Artificial intelligence (AI) encompasses a large range of theories and technologies used to solve problems of high logical or algorithmic complexity. It crosses many disciplines, including mechanistic modelling, software engineering, data science, and statistics (Figure  1 ). Introduced in the 1950s, many AI methods have been developed or extended recently with the improvement of computer performance. Recent developments have been fuelled by the interfaces created between AI and other disciplines, such as bio-medicine, as well as massive data from different fields, particularly those associated with healthcare [ 1 , 2 ].

Interactions between animal health (AH), artificial intelligence (AI), and closely related research domains. This illustration is pinpointing only the links between AH (in blue), AI and its main subfields (in red), and other related fields of research (in black). It can be naturally complexified through the interactions between AH and other research topics (e.g., human medicine) or between core disciplines (e.g., statistics and physics).

AI addresses three challenges that also make sense in animal health (AH): (1) understanding a situation and its dynamics, e.g., epidemic spread; (2) the perception of the environment, which corresponds in AH to the detection of patterns (e.g., repeated sequence of observations), forms (e.g., of a protein) and signals (e.g., increased mortality compared to a baseline) at different scales; (3) computer-based decision making, or, more realistically, human decision support (e.g., expert systems, diagnostic support, resource allocation).

To answer these challenges, a wide range of concepts and methods are developed in AI. This includes machine learning (ML), a widely known AI method nowadays, which has been developing since the 1980s [ 3 ]. Since the 2000s, deep learning is developing with the rise of big data and the continuous increasing of computing capacities, enabling the exploration of massive amount of information that cannot be processed by conventional statistical methods. In addition, this also includes methods and algorithms for solving complex problems, automating tasks or reasoning, integrating information from heterogeneous sources, or decision support (Figure  1 ). These methods are now uprising in the human health sector, but are still rarely used to study animal health issues that they would help to revisit.

Part of the scientific challenges faced in AH can be approached from a new perspective by using some of these AI methods to analyse the ever-increasing data collected on animals, pathogens, and their environment. AH research benefits from advances in machine and deep learning methods, e.g., in predictive epidemiology, individual-based precision medicine, and to study host–pathogen interactions [ 2 , 4 ]. These methods contribute to disease diagnosis and individual case detection, to more reliable predictions and reduced errors, to speed-up decisions and improved accuracy in risk analysis, and to better targeting interventions in AH [ 5 ]. AH research also benefits from scientific advances in other domains of AI. Knowledge representation and modelling of reasoning [ 6 ] allow more realistic representations of complex socio-biological systems such as those encountered in AH. Examples include processes related to decision-making and uncertainty management [ 7 , 8 ], as well as of patient life courses like in human epidemiology [ 9 ]. This contributes to making them more readable by noncomputer experts. In addition, advances in problem solving under constrained resource allocation [ 10 ], in autonomous agents [ 11 ], multi-agent systems [ 12 ], and multi-level systems [ 13 ], as well as on automatic computer code generation [ 14 ] can be mobilised to enhance efficient and reliable epidemiological models. Interestingly, this may aid to anticipate the effect of control and management decisions at different spatial and temporal scales (animal, herd, country…).

Conducting research at the AH/AI interface also leads to identify new challenges for AI, on themes common with human health but considering different contexts and perspectives [ 15 ]. First, taking into account the particular agro- and socio-economic conditions of production systems is crucial when dealing with AH. Animal production systems depend on human activities and decisions. They can be a source of income (e.g., livestock) or labour forces and source of food in family farming. Citizens have also high expectations in terms of ethics and animal welfare [ 16 ]. Conventional measures to control animal diseases may no longer be acceptable by society (e.g., mass culling during outbreaks [ 17 ], antimicrobial usage, [ 18 ]). Alternatives must be identified and assessed, and AI can contribute. For example, individual-based veterinarian medicine is emerging, mobilising both AI methods and new AH data streams, these data differing from data in human health [ 19 ]. The integration of data from deep sequencing in AH, including emerging technologies for studying the metabolome and epigenome, is also a challenge [ 20 , 21 ]. Second, interactions between animal species, in particular between domestic animals and wildlife, lead to specific infectious disease risks (e.g., multi-host pathogens such as for African swine fever, pathogens crossing the species barrier facilitated by frequent contacts and promiscuity). The intensity of such interactions could increase due to separate or synergistic actions of environmental (e.g., landscape homogenisation, land use change for agriculture development, climate change), demographic (e.g., increasing global demand for animal production) and societal (e.g., outdoor livestock management) pressures. In addition, working on multi-species disease networks provides crucial information on the underlying molecular mechanisms favouring interspecific transmission [ 22 ]. Third, animal populations are governed by recurrent decision-making that also impacts health management (e.g., trade, control measures). Economic criteria as consequences on livestock farmers’ incomes are therefore essential indicators for evaluating AH control strategies, which can sometimes be misunderstood or may be at odds with societal expectations. These specificities make the AH/AI interface a theme of interest to stimulate new methodological work and to solve some of old and current locks faced by AH research today. With the development of new concepts in health such as One Health, Ecohealth and Planetary Health, promoting multidisciplinarity, stakeholders’ participation, data sharing, and tackling the complexity of health issues (e.g., multi-host pathogen transmission, short and long-term climatic impacts on disease patterns [ 23 ]), AI could participate in this new development by making it possible to technically solve some of the complex problems posed.

Mobilising the literature published at the AH/AI interface between 2009 and 2019 (Additional file 1 A), focusing our literature search on mainly livestock and wildlife, as well as interviews conducted with French researchers positioned at this interface (Additional file 1 B), we identified the main research areas in AH in which AI is currently involved country-wide. We explored how AI methods contribute to revisiting AH questions and may help remove methodological or conceptual barriers within the field. We also analysed how AH questions interrogate and stimulate new AI technical or scientific developments. In this paper, we first discuss issues related to data collection, organisation and access (Section  1 ), then we discuss how AI methods contribute to revisiting our understanding of animal epidemiological systems (Section  2 ), to improving case detection and diagnosis at different scales (Section  3 ), and to anticipating pathogen spread and control in a wide range of scenarios in order to improve AH management, facilitate decision-making and stimulate innovation (Section  4 ). Finally, we present the possible obstacles and levers to the development of AI in modern AH (Section  5 ), before making recommendations to best address the new challenges represented by this AH/IA interface (Section  6 ).

2 Collect, organise and make accessible quality data

A central point for research in AH remains the quality and availability of data, at the different organization levels of living systems and therefore at different spatial and temporal scales [ 24 ]. Data of interest are diverse. They can be obtained thanks to molecular analysis (e.g., genomic, metagenomics, or metabolic data), from observational data on individuals (e.g., body temperature, behaviour, milk production and composition, weight, feed intake), or from the production system (e.g., herd structure, breeding, management of sanitary issues). They can also be obtained at larger scales, beyond herds or local groups of animals (e.g., epidemiological data, demographic events, commercial movements, meteorological data, land-use occupation).

Even though the acquisition of these massive and heterogeneous data remain challenging (e.g., metabolome data), a large and diverse amount is already collected: (i) through mandatory reporting in accordance with regulations (e.g., commercial movements of cattle, epidemio-surveillance platform), (ii) by automatic devices (e.g., sensors, video surveillance systems), and (iii) on an ad hoc basis as part of research programs. This leads to a very wide diversity of data properties, and therefore of their management, access and possible uses. These data can be specifically obtained for certain animals or herds (e.g., during cohort monitoring programs) or by private companies (e.g., pig trade movements such as in France, milk collection). This can limit accessibility to academics and public research. Globalisation and large-scale animal trade may generate the need to use data obtained at worldwide scale in AH, especially for quantitative epidemiology (e.g., transcontinental spread of pathogens, animal genetics and breed management) leading to standardisation issues [ 25 ].

Consideration should be given to future systems for observing, collecting and managing these data [ 26 ], and to practices aimed at better collaboration between stakeholders. While data management has always been an important element in applied research to facilitate their use and valorisation, it is now a strategic issue both in theoretical and more applied research, coupled with a technical and algorithmic challenge [ 27 , 28 , 29 ]. Indeed, producing algorithms to manage massive data flows and stocks, by optimising calculations, is a challenge, particularly in real time. It seems also necessary to make heterogeneous data sources interoperable, requiring dedicated methodological developments [ 25 ]. In addition, much of the data is private, with ownership often heterogeneous (e.g., multiple owners, non-centralised data, closed data) and sometimes unclear (e.g., lack of knowledge of the real owner of the data between, for example, farmers, the data collector or the farmers’ union). All this tends to considerably complicate access to the data, raises questions about intellectual property, and raises questions in relation to regulations with regards to data protection, e.g., the adaptation of regulation to AH while respecting the confidentiality of the personal data mobilised. What is the relevant business model for data collection or access to existing databases? What about the openness of AH data (e.g., duality between the notion of public good and the private nature of certain data) to make it possible to experiment in real situations and compare the performance of AI algorithms? Answering these questions would facilitate the collection and sharing of ad hoc data. AI, particularly when combining a participatory framework with expert systems and multi-agent systems, helps to build realistic representations of complex socio-biological systems. Thus, it proves to be an effective tool to promote the collaboration of different stakeholders in collective and optimised decision-making, and to assess of the impact of changes in uses and practices [ 30 ].

Encouraging experimentation of AI technologies at a territorial scale becomes crucial to favour their development, validate their performance, and measure their predictive quality. In AH, simplified access to data-generating facilities would allow innovative solutions to be tested on a larger scale and would accelerate their development and evaluation. Substantial expertise exists (e.g., epidemiological data platform, large cohorts, experimental farms) that could be put to good use. In addition, AI could help to revisit sampling methods for field data collection in AH and epidemiological surveillance, by better and more dynamically targeting the data to be collected while avoiding redundant collinear, non-necessary data.

3 Contribution of AI to better understand animal epidemiological systems

Recent technological advances involving AI approaches have made it possible to obtain vast quantities of measurements and observations, as well as to store and share these data more efficiently. This has resulted in an increasing requirement for appropriate data analytical methods. AI methods emerged as the response of the computer-science community to these requirements, leveraging the exponential improvements in computational power. In parallel, statistical methods have greatly evolved in the last few decades as well, e.g., with regards to dimensionality-reduction in the spaces of variables and parameters, variable selection, and model comparison and combination. The rise in computational power has unleashed the development of Bayesian inference through simulation or approximate methods [ 31 ]. Bayesian methods have, in turn, facilitated the integration of data from diverse sources, the incorporation of prior knowledge and allowed for inference on more complex and realistic models while changing the paradigm of statistical inference [ 32 , 33 , 34 ].

3.1 Better understanding the evolution of AH and socio-ecological systems in a One Health context

Learning methods can be used to do phylogenetic reconstructions, contributing in particular to new evolutionary scenarios of pathogens and their transmission pathways. For example, phylogenetic models offer an interesting perspective for identifying environmental bacterial strains with high infectious potentiality [ 35 ], or for predicting the existence of putative host reservoirs or vectors [ 36 ]. The analysis of pathogen sharing among hosts has been used to classify the potential reservoirs of zoonotic diseases using machine learning [ 37 ]. The analysis of pathogen genomes can also be used to identify genotypes of animal pathogens that are more likely to infect humans [ 38 ].

Using phenomenological niche models that rely on data distribution more than on hypotheses about ecological processes at play, disease occurrence data or retrospective serological data coupled with environmental variables can be related to the risk of being exposed to a pathogen. Thus, they can help monitor potential spillovers and emerging risks and anticipate the epidemic pathogen spread [ 39 ]. For instance, Artificial Neural Networks (ANN) have identified the level of genetic introgression between wild and domesticated animal populations in a spatialized context [ 40 ], which may help to understand gene diffusion in host × pathogen systems involving multiple host species, and characterise specimen pools at higher risk to act as pathogen spreaders or sinks. Other AI approaches such as multi-agent models, a more mechanistic approach, have been used in an explicit spatial context for vector-borne pathogen transmissions, and proved to be sufficiently versatile to be adapted to several other particular contexts [ 12 ].

It should be noted here that several studies reveal the relatively ancient nature of AI research in AH. Such AI methods have often made it possible to identify signals (e.g., genetic introgression) or even particular patterns or properties (e.g., importance of density-dependence in the vector-borne transmission) that are less visible or hardly detectable by more conventional statistical treatments.

All these approaches contribute to better understand pathogen transmission in complex system networks as generally observed for emerging infections in tropical, developing regions of the world. On this matter, an improved knowledge is key for protecting humans against these new threats, and AI/AH interfaces development and training in cooperation with the poorest countries would facilitate synergistic effects and actions to predict and tackle new disease threats.

3.2 Reliability, reproducibility and flexibility of mechanistic models in AH

Better understanding and predicting pathogen spread often requires an explicit and integrative representation of the mechanisms involved in the dynamics of AH systems, irrespective of the scale (within-host: [ 41 ]; along a primary production chain: [ 42 ]; in a territory: [ 43 , 44 ]; over a continent: [ 45 ]).

Mathematical (equations) or computer-based (simulations) models can be used. Such mechanistic models (i.e., which represent the mechanisms involved in the infection dynamics), when sufficiently modular to represent contrasted situations, make it possible to anticipate the effects of conventional but also innovative control measures (e.g., new candidate molecules, sensors, genomic selection; [ 46 ]).

However, to assess realistic control measures, mechanistic epidemiological models require the integration of observational data and knowledge from biology, epidemiology, evolution, ecology, agronomy, sociology or economics. Their development can rapidly face challenges of reliability, transparency, reproducibility, and usage flexibility. Moreover, these models are often developed de novo, making little use of previous models from other systems. Finally, these models, even based on realistic biological hypotheses, may be considered negatively as black boxes by end users (health managers), because the underlying assumptions often became hidden in the code or equations.

The integration of multiple modelling perspectives (e.g., disciplines, points of view, spatio-temporal scales) is an important question in the modelling-simulation field. Epidemiological modelling could benefit from existing tools and methods developed in this field [ 47 , 48 , 49 ]. Although essential, good programming practices alone [ 50 ] cannot meet these challenges [ 51 ]. Scientific libraries and platforms accelerate the implementation of the complex models often needed in AH. For example, the R library SimInf [ 52 ] helps integrate observational data into mechanistic models. The BROADWICK framework [ 53 ] provides reusable software components for several scales and modelling paradigms, but still requires modellers to write large amounts of computer code.

New methods at the crossroads between software engineering and AI can enhance transparency and reproducibility in mechanistic modelling, fostering communication between software scientists, modellers and AH researchers throughout the modelling process (e.g., assumption formulation, assessment, and revision). Knowledge representation methods from symbolic AI, formalised using advanced software engineering methods such as domain-specific languages (DSL, e.g., in KENDRICK for differential equation models: [ 54 ]), makes model components accessible in a readable structured text file instead of computer code. Hence, scientists from various disciplines and field managers can be more involved in the model design and evaluation. Scenario exploration and model revision also no longer require rewriting the model code.

Other AI methods can improve model flexibility and modularity. Autonomous software agents enable to represent various levels of abstraction and organisation [ 55 ], helping modellers go more easily back and forth within small and larger scales, and ensure that all relevant mechanisms are adequately formalised at proper scales (i.e., scale-dependency of determinants and drivers in hierarchical living systems). Combining knowledge representation (through a DSL) and such a multi-level agent-based simulation architecture (e.g., in EMULSION, Figure  2 , [ 56 ]) enables to encompass several types of models (e.g., compartmental, individual-based) and scales (e.g., individual, population, territory), and it tackles simultaneously the recurring needs for transparency, reliability and flexibility in modelling contagious diseases. This approach should also facilitate in the future the production of support decision tools for veterinary and public health managers and stakeholders.

AI at the service of mechanistic epidemiological modelling (adapted from [ 51 ] ) . A . Modellers develop each epidemiological model de novo, producing specific codes not easily readable by scientists from other disciplines or by model end-users. B . Using AI approaches to combine a domain-specific language and an agent-based software architecture enhances reproducibility, transparency, and flexibility of epidemiological models. A simulation engine reads text files describing the system to automatically produce the model code. Complementary add-ons can be added if required. Models are easier to transfer to animal health managers as decision support tools.

3.3 Extracting knowledge from massive data in basic AH biology

Supervised, unsupervised and semi-supervised learning methods facilitate basic research development in biology and biomedicine, for example by using morphological analyses to study cell mobility [ 57 ]. The use of classification approaches and smart filters allows nowadays to sort massive molecular data (e.g., data from high throughput sequencing and metagenomics). Metabolic, physiological and immunological signalling pathways are explored, and metabolites are identified and quantified in complex biological mixtures, which was before a major challenge [ 58 ]. In addition, diagnostic time may be reduced by developing image analysis processing (e.g., accelerated detection of clinical patterns; [ 59 , 60 ]), often necessary to study host–pathogen interactions in animal pathology. For example, the use of optimisation methods has improved the understanding of the fragmentation of prion assemblages, contributing to a significant reduction in the time required to diagnose neurodegenerative animal diseases, thus paving the way for identifying potential therapeutic targets [ 61 ]. In livestock breeding, there is a methodological transition underway from traditional prediction strategies to more advanced machine learning approaches including artificial neural networks, deep learning and Bayesian networks which are being used to improve the reliability of genetic predictions and further the understanding of phenotypes biology. [ 62 ].

In human health, new disciplines have emerged in the second half of the 20 th century at the interface between AI and flagship disciplines, such as cell biology and immunology. Interface disciplines have developed, e.g., computational biology and immunology, which today must spread to AH. Current human immunology is based on the description of fine molecular and cellular mechanisms (e.g., the number of known interleukins has increased considerably compared to the 1970s). The desire to understand the processes underlying immune responses has led to a revolution by inviting this discipline to focus on complex systems biology and AI-based approaches [ 63 ]. However, the imbalance between the numbers of immunologists and immunology modellers is hampering the fantastic growth of this new discipline.

As an additional level of complexity, the hierarchical nature of biological systems makes that at the individual level, animals including humans must be considered as holobionts made of myriads of hosted microbial forms that form discrete ecological units (i.e., infracommunities). The potential of AI to grasp such diversity and complexity (e.g., tissue-specific microbiotes) and to scaling-up to higher levels of organization (e.g., component and compound communities of microbes, including pathogens, circulating in herd and in a given region) is certainly tremendous and should be studied with the same vigour as recent development in computational biology and immunology [ 40 ].

4 Revisiting AH case detection methods at different scales

Managing livestock health issues requires effective case detection methods, at the individual or even infra-individual (organ) scale, at the group/herd scale, or at larger scales (e.g., territories, countries). Machine learning methods allow detecting patterns and signals in massive data, e.g., in spatial data or time-series of health syndromes and disease cases, contributing to the development of smart agriculture and telemedicine (Figure  3 ). Alerts can be produced, and contribute to management advice in numerical agriculture [ 64 ] and veterinary practices [ 65 ]. AI may contribute to an earlier detection of infected cases and the rationalisation of treatments (including antimicrobials) in farm animals, by analysing data collected from connected sensors [ 66 ], by targeting individuals or groups of animals [ 59 ], or even by using mechanistic models to predict the occurrence of case detections and their treatment [ 67 ]. Also, machine learning methods enable to discriminate pathogen strains and thus to better understand their respective transmission pathways if different [ 68 ]. Finally, therapeutic strategies can be reasoned through multi-criteria optimisation, by identifying whom to treat in a herd, when, according to what protocol and for how long, in order to maximise the probability of cure while minimising both the risk of drug resistance and the volume or number of doses that are necessary (i.e., individual-based and precision medicine).

Extracting information from massive data to monitor animal health and better rationalise treatments.

Nevertheless, alert quality depends on the quality and representativeness of the datasets used by the learning algorithms. Numerous biases (e.g., hardware, software, human) can affect prediction accuracy. Moreover, alerts produced after training necessarily reflect the specificities of the system from which the data originates (e.g., area, period, rearing practices). Thus, result transposition to other epidemiological systems or to the same system subjected to environmental or regulatory changes remains risky. Furthermore, while machine learning methods (e.g., classification, image analysis, pattern recognition, data mining) provide solutions for a wide range of biomedical and bio-health research questions, it is crucial to demonstrate the performance of these methods by measuring their predictive quality and comparing them to alternative statistical methods whenever possible [ 69 ].

At the population level, case detection is based on direct (detection of syndromes) or indirect surveillance, mobilising syndrome proxies. Hence, the emergence of some animal diseases can be detected by syndromic surveillance, by detecting abnormal or rare signals in routine data (e.g., mortality, reproduction, abortion, behaviour, milk production, increased drug use; [ 70 ]). Also, serological data can be used retrospectively to identify individual characteristics related to a risk of being exposed to a pathogen, and thus orientate management efforts (e.g., in wildlife; [ 71 ]). Statistics and AI are largely complementary to address such issues. Both mobilise the wide range of available data, which are highly heterogeneous, massive and mostly sparse, to detect signals that are often weak or scarce [ 28 , 72 , 73 ]. Such signals can be proxy records (e.g., emergence of infectious diseases following environmental disturbances), health symptoms and syndromes, or even metabolic pathways in cascades which can be precursors of chronic or degenerative diseases. AI also includes methods to mobilise information available on the web. For example, semi-automatic data mining methods enabled to identify emerging signals for international surveillance of epizooties [ 74 ] or to analyse veterinary documents such as necropsy reports [ 75 , 76 ]. Methods from the field of natural language processing can compensate the scarcity of data by extracting syntactic and semantic information from textual records, triggering alerts on new emerging threats that could have been missed otherwise.

On a large to very large scale (i.e., territory, country, continent, global), data analysis of commercial animal movements between farms makes it possible to predict the associated epidemic risk [ 77 , 78 ]. These movements are difficult to predict, particularly since animal trade is based on many factors associated with human activities and decisions. Methods for recognising spatio-temporal patterns and methodological developments for the analysis of oriented and weighted dynamic relational graphs are required in this field because very few of the existing methods allow large-scale systems to be studied, whereas datasets are often very large (e.g., several tens or even hundreds of thousands of interacting operations).

On this topic, the specific frontier between learning methods of AI and statistics is relatively blurred, lying most on the relative prominence of the computational performance of algorithms versus mathematics, probability and rigorous statistical inference. While machine learning methods are more empirical, focused on improving their predictive performance, statistics is more concerned with the quantification and modelling of uncertainties and errors [ 79 , 80 ]. In the last decade, both communities have started to communicate and to mix together. Methods have cross-fertilised, giving birth to statistical models using synthetic variables generated by AI methods, or AI algorithms optimising statistical measures of likelihood or quality. New research areas, such as Probabilistic Machine Learning, have emerged at the interface between the two domains [ 1 , 80 , 81 ]. Meanwhile, machine learning and statistics have kept their specific interests and complementarity; machine learning methods are especially well-suited to processing non-standard data types (e.g., images, sounds), while statistics can draw inference and model processes for which only few data are available, or where the quantities of interest are extreme events.

5 Targeted interventions, model of human decisions, and support of AH decisions

5.1 choosing among alternatives.

A challenge for animal health managers is to identify the most relevant combinations of control measures according to local (e.g., farm characteristics, production objectives) and territorial (e.g., available resources, farm location, management priorities) specificities. They have to anticipate the effects of health, environmental and regulatory changes, and deliver quality health advice. The question also arises of how to promote innovation in AH, such as to anticipate the required characteristics of candidate molecules in vaccine strategies or drug delivery [ 82 , 83 ], or to assess the competitive advantage of new strategies (e.g., genomic selection of resistant animals, new vaccines) over more conventional ones. Private (e.g., farmers, farmers’ advisors) and collective managers (e.g., farmer groups, public authorities) need support decision tools to better target public incentives, identify investments to be favoured by farmers [ 46 ] and target the measures as effectively as possible: who to target (which farms, which animals)?; with which appropriate measure(s)?; when and for how long? These questions become essential to reasoning about input usage (e.g., antimicrobials, pesticides, biocides) within the framework of the agro-ecological transition.

The use of mechanistic modelling is a solution to assess, compare and prioritise ex ante a wide range of options (Figure  4 ; [ 84 ]). However, most of the available models do not explicitly integrate human decision-making, while control decisions are often made by farmers (e.g., unregulated diseases), with sometimes large-scale health and decision-making consequences (e.g., pathogen spread, dissemination of information and rumours, area of influence). Recent work aims to integrate humans and their decisions by mobilising optimal control and adaptive strategies from AI [ 7 , 85 ] or health economics methods [ 86 , 87 ]. A challenge is to propose clear and context-adapted control policies [ 88 ]. Such research is just starting in AH [ 46 ] and must be extended as part of the development of agro-ecology, facing current societal demand for product quality and respect for ecosystems and their biodiversity on one side, animal well-being and ethics on the other side, and more generally international health security.

Identifying relevant strategies to control bovine paratuberculosis at a regional scale (adapted from [ 76 ] ) . Classically, identifying relevant strategies means defining them a priori and comparing them, e.g., by modelling. Only a small number of alternatives can be considered. If all alternatives are considered as in the figure, it results in a multitude of scenarios whose analysis becomes challenging. Here, each point corresponds to the epidemiological situation after 9 years of pathogen spread over a network of 12 500 dairy cattle herds for a given strategy (asterisk: no control). Initially, 10% of the animals are infected on average in 30% of the herds. The blue dots correspond to the most favourable strategies. Mobilizing AI approaches in such a framework, especially optimization under constraints, would facilitate the identification of relevant strategies by exploring the space of possibilities in a more targeted manner.

5.2 Accounting for expectations and fears of animal health managers

Animal health managers should have access to model predictions in a time frame compatible with management needs, which is problematic in the face of unpredictable emerging events (e.g., new epidemiological systems, transmission pathways, trade patterns, control measures). Developing a library of models included in a common framework would strengthen the responsiveness of modellers in animal epidemiology. Relevant models would be developed more quickly and would gain accuracy from real-time modelling as epidemics progress [ 89 , 90 ]. However, if this makes move more quickly from concepts (knowledge and assumptions) to simulations and support decision tools, a gain in performance is still required to perform analyses at a very large scale. The automatic generation of high-performance computer code could be a relevant solution, which however remains a crucial methodological lock to be addressed in AI. In addition, it is often required to perform a very large number of calculations or to analyse very large datasets, which call for a rational use of computing resources. Software transferred to health managers sometimes require the use of private cloud resources (i.e., it does not run on simple individual computers), highlighting the trade-offs between simulation cost, service continuity (e.g., failure management) and time required to obtain simulation results [ 91 ]. These questions are currently related to computer science research, and collaborations are desirable between these researchers and those from AH.

Managers also wish to rely on accurate predictions from realistic representations of the biological systems. Before being used, model behaviour should be analysed, which raises the questions of exploring the space of uncertainties and data, and of optimization under constraints. This often requires intensive simulations, which would benefit from optimization algorithms to explore more efficiently the space of possibilities. In turn, this would allow, for example, the automatic identification of how to achieve a targeted objective (e.g., reducing the prevalence of a disease below an acceptable threshold) while being constrained in resource allocation. While this issue finds solutions in modern statistics for relatively simple systems, it represents a science front for complex systems (e.g., large scale, multi-host/multi-pathogen systems) that are becoming the norm. In addition, optimization goals specific to AH may generate ad hoc methodological needs [ 92 ]. The needs in abstraction and analysis capacity are massive and could benefit from complementarities between AI (e.g., reasoned exploration, intelligent use of computer resources, optimized calculations) and statistics to extract as much information as possible from the data: (1) explore, analyse, predict; (2) infer processes and emergent properties. Methodological developments are still required and would benefit many health issues, particularly in relation to the currently evolving concepts of reservoir-host, edge-host and species barrier [ 93 ]. Furthermore, methodological developments and dissemination of existing methods should be reinforced.

Finally, three barriers have been identified to the development of support decision tools for health managers, related to the societal issue of the acceptability of AI sensu lato, as a major factor of progress. First, ethical issues, which are obvious when it comes to human health, are just as important to consider in AH. Which AI-based tools do we want for modern animal husbandries and trades, and for which objectives? Are these tools not likely to lead to discrimination against farms according to their health status, even when this status cannot be managed by the farmer alone? Second, in AH too, there is a fear that AI-tools may replace human expertise. However, automating does not mean replacing human, his expertise and decision [ 94 ], but rather supporting his capacities for abstraction and analysis, accelerating the global process, making predictions more reliable, guiding complementary research. Nevertheless, a significant development of computer resources and equipment is not without impacting the environment in terms of carbon footprint (e.g., energy-intensive servers, recycling of sensors), which must also be accounted for. Third, the very high complexity of analysing results and acculturating end-users with knowledge issued from academic research, particularly AI, is an obstacle to the appropriation of AI-tools by their users. This may lead to the preference for simpler and more easily accessible methods. However, the latter may not always be the most relevant or reliable. Citizen science projects, also known as community participation in human epidemiology, enable AH to co-design and co-construct the AI-tools of tomorrow with their end-users [ 95 ], to better meet their expectations and needs, and to increase their confidence in the predictions of sometimes obscure research models, especially when they are hard to read (e.g., lines of code). Similarly, these AI-tools could be developed together with public decision-makers, livestock farmers, agro-food industries and sectoral trade unions. Co-construction gives time to explain the science behind the tools and makes it more transparent and useful. This citizen participation, which is nowadays supported in many countries, guarantees decisions more in line with citizens’ expectations and corresponds to a general trend towards structured decision-making. AI must contribute to this democratisation of aid in public decision-making in AH.

6 Barriers to the development of research at the AI/AH interface

Research conducted at the interface between AI and AH requires strong interactions between biological disciplines (e.g., infectiology, immunology, clinical sciences, genetics, ecology, evolution, epidemiology, animal and veterinary sciences) and more theoretical disciplines (e.g., modelling, statistics, computer science), sometimes together with sociology and economics. Conducting research at this interface requires strengthening the few teams already positioned in Western Europe, but also bringing together teams working around the concepts of One Health, Ecohealth and Planetary Health to benefit from recent achievements in infectious disease ecology and modelling, plant health and environmental health [ 96 ]. This work must be based on a wide range of methodological skills (e.g., learning methods, data mining, information systems, knowledge representation, multi-agent systems, problem solving, metamodeling, optimisation, simulation architecture, model reduction, decision models). The need for research, training and support are crucial issues at national, European and international levels. Also, a facilitated and trusted connection is required between academics, technical institutes, and private partners, who are often the holders or collectors of data of interest to solve AH research questions through AI approaches. The construction of better inter-sectoral communication and coordination must be done at supra-institutional level, as this theme seems hyper-competitive and as some current divisions still go against information and data sharing.

An acculturation of researchers to AI, its methods and potential developments, but also its limitations, must be proposed to meet the challenges of 21 st century agriculture. Indeed, there are obstacles to conducting research on this scientific front. Establishing the new collaborations required between teams conducting methodological work and teams in the fields of application remains difficult given the low number of academic staff on these issues, their very high current mobilisation and their low availability to collaborate on new subjects, as well as the difficulty of understanding and mastering these methods. There is a need for watching and training on AI methods available or under development, new softwares/packages, and their applicability. To develop key collaborations and establish a strategic positioning, an interconnection can also be made via transversal teams which appears as a preferential path. Solutions must also be found to encourage method percolation in the community and the development of scientific and engineering skills.

Finally, AI methods, such as classification, machine learning, data mining, and the innovations in AH to which these methods can lead are rarely discussed in veterinary high school education, whereas these students represent the future professionals of AH [ 96 ]. Similarly, there is a quasi-absence of sentinel networks of veterinarians, even if it is developing, although AH questions can arise on a large and collective scale. The scientific community would also benefit from further increasing its skills and experience in the valuation, transfer and protection of intellectual property on these AI methods and associated outcomes.

7 Levers to create a fruitful AI/AH interface

7.1 data sharing and protection.

No innovation at the interface between AI and AH is possible without strong support for the organisation of data storage, management, analysis, calculation, and restitution. The major risk is that demands for AI developments inflate without being supported by available human resources. In addition, an expertise in law, jurisdiction and ethics is required with regard to the acquisition, holding, use and protection of data in AH. This question must be considered at least at the inter-institutional/national level, and could benefit from a similar thinking already engaged in human health. The issue is to be able to support any change with regard to data traceability to their ownership, whether being from public or private domains.

New data are rich and must be valued as much as possible, not by each owner separately, but through data sharing and the mobilisation of multi-disciplinary skills to analyse such heterogeneous and complex data. Hence, data interoperability skills are required and must be developed. Models for making federated data sustainable over decades are required [ 97 ]. In addition, further encouraging the publication of data papers as valuable research products can help to develop the necessary culture of sharing, documentation and metadata.

Finally, to be able to launch ambitious experiments with AI methods on real data, it is necessary to (1) remove unauthorised access to data by negotiating with owners at large scale; (2) analyse and understand the related effect on methodological developments; and (3) if necessary, extend such initiatives to other areas, at national scale, or even across European countries.

7.2 Attract the necessary skills

An undeniable barrier to conduct such research comes from human resources, in particular the current insufficient capacity of supervision by permanent scientists. Collaborations are a solution to attract new skills. However, initiating collaborations at the AH/AI interface becomes very complicated because the qualified teams are already overwhelmed. Skill development at this interface must be supported, the cross-fertilisation of disciplines being essential. A watch on methods must also be carried out, accompanied by explanations for application fields, to train researchers and engineers. Financial incentives for scientist internships in specialised laboratories would increase skill capitalisation in advanced methods, while facilitating future national or international collaborations. In a context of limited resources as observed in many countries nowadays (e.g., new opened positions in national institutions) and limited experts pool (e.g., skills), facilitating post-doctoral fellows and continuing education of researchers becomes crucial. Finally, to consolidate the pool of future researchers in AH, promoting basic AI education in initial training of AH researchers, engineers and veterinarians is paramount.

More specifically concerning current research in immunology, cell biology and infectiology, the contribution of AI has been more widely considered in human health, which could feed a similar reflection in AH as locks and advances are not very specific. Before embarking on the fronts of science (e.g., emerging epigenomics and metabolomics in AH), a few persons from these biological disciplines should acculturate into AI, or even acquire autonomy in the use of methods [ 98 ], which internationally tends to be the trend [ 63 ]. This can be done through the sharing of experiences and basic training on existing methods, their advantages and limitations compared to other methods coming from statistics and mathematical modelling.

7.3 Encourage the development of AH/AI projects

Projects at the AH/AI interface, like any interdisciplinary project, must mobilise teams from both groups of disciplines and allow everyone to progress in their own discipline. However, identifying the issues shared between the most relevant disciplines requires a good acculturation of the disciplines between them, as well as an otherness aimed at better understanding each other [ 95 ], which is not yet the case at the AH/AI interface.

In terms of funding, European project calls offer interesting opportunities, but a significant imbalance persists between the ability to generate data and analyse complex issues, and the availability of human resources and skills to address such issues through AI methods or other modern methods in statistics, mathematics and computer science. The major international foundations (e.g., Bill and Melinda Gates) can also be mobilised on emerging infectious diseases at the animal/human interface (e.g., characterisation of weak signals, phenologies, emergence precursors), with a more significant methodological value. However, risk-taking is rarely allowed by funding agencies, although it is crucial to initiate interdisciplinary work. Dedicated incentive funding would support projects in their initial phase and make larger projects emerge after consolidation of the necessary disciplinary interactions.

Finally, these projects are generally based on the use of significant computing resources. Thus, research institutes and private partners should contribute in a financial or material way to the shared development of digital infrastructures, data centres, supercomputing centres on a national scale, as well as support recognised open-source software platforms on which a large part of the research conducted is based (e.g., Python, R ).

7.4 Promoting innovation and public–private partnership

Encouraging public–private partnership would promote a leverage effect on public funding and would make it possible to place AI research and development on a long-term basis in AH. Mapping the highly changing landscape of companies in the AH/AI sector, whether international structures or start-ups, would provide a better understanding of the possible interactions. Similarly, mapping academic deliverables produced at this interface would increase their visibility and highlight their potential for valorisation or transfer. Finally, considering the production of documented algorithms as scientific deliverables, along with publications, would help support this more operational research. More broadly, it would be advisable to initiate a communication and education/acculturation policy around AI and its development in AH (e.g., links with the society, farmers, agricultural unions, public services).

8 Conclusion

The use of AI methods (e.g., machine learning, expert systems, analytical technologies) converges today with the collecting of massive and complex data, and allows these fields to develop rapidly. However, it is essential not to perceive massive data and AI as the same trend, because the accumulation of data does not always lead to an improvement in knowledge. Nevertheless, the more data are numerous and representative of working concepts and hypotheses, the more important results can be obtained from AI applications. The underlying ethical, deontological and legal aspects of data ownership, storage, management, sharing and interoperability also require that a reflection be undertaken nationally and internationally in AH to better manage these data of multi-sectoral origin and their various uses. Moreover, while the effort to acquire such data is impressive, the development of AI skills within the AH community remains limited in relation to the needs. Opportunities for collaborations with AI teams are limited because these teams are already in high demand. To ensure that AH researchers are well aware of the opportunities offered by AI, but also of the limits and constraints of AI approaches, a training effort must be provided and generalized. Finally, the current boom in AI now makes it possible to integrate the knowledge and points of view of the many players in the field of animal health and welfare further upstream. However, this requires that AI and its actors accept to deal with the specificity and complexity of AH, which is not a simple library of knowledge that can be digitised to search for sequences or informative signals.

Hosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL (2018) Artificial intelligence in radiology. Nat Rev Cancer 18:500–510

Article   CAS   Google Scholar  

Karczewski KJ, Snyder MP (2018) Integrative omics for health and disease. Nat Rev Genet 19:299–310

Murphy KP (2012) Machine learning: a probabilistic perspective. In: Adaptive computation and machine learning series. MIT Press, USA

Zhang W, Chien J, Yong J, Kuang R (2017) Network-based machine learning and graph theory algorithms for precision oncology. NPJ Precis Oncol 1:25

Article   Google Scholar  

Saria S, Butte A, Sheikh A (2018) Better medicine through machine learning: what’s real, and what’s artificial? PLoS Med 15:e1002721. https://doi.org/10.1371/journal.pmed.1002721

Article   PubMed   PubMed Central   Google Scholar  

Bedi G, Carrillo F, Cecchi GA, Fernández Slezak D, Sigman M, Mota NB, Ribeiro S, Javitt DC, Copelli M, Corcoran CM (2015) Automated analysis of free speech predicts psychosis onset in high-risk youths. Schizophrenia 1:15030

Maclachlan MJ, Springborn MR, Fackler PL (2017) Learning about a moving target in resource management: optimal Bayesian disease control. Am J Agri Econ 99:140–162. https://doi.org/10.1093/ajae/aaw033

Lynn LA (2019) Artificial intelligence systems for complex decision-making in acute care medicine: a review. Patient Saf Surg 13:6. https://doi.org/10.1186/s13037-019-0188-2

Pinaire J, Azé J, Bringay S, Landais P (2017) Patient healthcare trajectory. An essential monitoring tool: a systematic review. Health Inf Sci Syst 5:1

Vrakas D, Vlahavas IPL (2008) Artificial intelligence for advanced problem solving techniques. Information Science Reference, Hershey, PA, pp. 369

Shakshuki E, Reid M (2015) Multi-agent system applications in healthcare: current technology and future roadmap. Proc Comput Sci 52:252–261. https://doi.org/10.1016/j.procs.2015.05.071

Roche B, Guégan JF, Bousquet F (2008) Multi-agent systems in epidemiology: a first step for computational biology in the study of vector-borne disease transmission. BMC Bioinform 9:435. https://doi.org/10.1186/1471-2105-9-435

Picault S, Huang Y-L, Sicard V, Ezanno P (2017) Enhancing Sustainability of Complex Epidemiological Models through a Generic Multilevel Agent-based Approach. In: Proceedings of the 26 th International Joint Conference on Artificial Intelligence (IJCAI). pp. 374–380, AAAI. https://doi.org/10.24963/ijcai.2017/53p

Russell S, Norvig P (2010) Artificial intelligence a modern approach. 3 rd edn. Upper Saddle River, New Jersey, pp. 1132

Google Scholar  

Ducrot C, Bed’Hom B, Béringue V, Coulon JB, Fourichon C, Guérin JL, Krebs S, Rainard P, Schwartz-Cornil I, Torny D, Vayssier-Taussat M, Zientara S, Zundel E, Pineau T (2011) Issues and special features of animal health research. Vet Res 42:96

Clark B, Stewart GB, Panzone LA, Kyriazakis I, Frewer LJ (2016) A systematic review of public attitudes, perceptions and behaviours towards production diseases associated with farm animal welfare. J Agric Environ Ethics 29:455–478. https://doi.org/10.1007/s10806-016-9615-x

Miguel E, Grosbois V, Caron A, Pople D, Roche B, Donnelly C (2020) A systemic approach to assess the potential and risks of wildlife culling for infectious disease control. Commun Biol 3:353. https://doi.org/10.1038/s42003-020-1032-z

Hur B, Hardefeldt LY, Verspoor K, Baldwin T, Gilkerson JR (2019) Using natural language processing and VetCompass to understand antimicrobial usage patterns in Australia. Aust Vet J 97:298–300. https://doi.org/10.1111/avj.12836

Article   CAS   PubMed   Google Scholar  

Behmann J, Hendriksen K, Mueller U, Buescher W, Pluemer L (2016) Support vector machine and duration-aware conditional random field for identification of spatio-temporal activity patterns by combined indoor positioning and heart rate sensors. Geoinformatica 20:693–714. https://doi.org/10.1007/s10707-016-0260-3

Suravajhala P, Kogelman LJA, Kadarmideen HN (2016) Multi-omic data integration and analysis using systems genomics approaches: methods and applications in animal production, health and welfare. Genet Sel Evol 48:38. https://doi.org/10.1186/s12711-016-0217-x

Article   CAS   PubMed   PubMed Central   Google Scholar  

Goldansaz SA, Guo AC, Sajed T, Steele MA, Plastow GS, Wishart DS (2017) Livestock metabolomics and the livestock metabolome: a systematic review. PLoS One 12:e0177675. https://doi.org/10.1371/journal.pone.0177675

Anvar SY, Tucker A, Vinciotti V, Venema A, van Ommen GJ, van der Maarel SM, Raz V, ’t Hoen PA (2011) Interspecies translation of disease networks increases robustness and predictive accuracy. PLoS Comput Biol 7:e1002258. https://doi.org/10.1371/annotation/fc0b4192-6427-4fb3-b347-c66651adf855

Morand S, Guégan J-F, Laurans Y (2020) From One Health to Ecohealth, mapping the incomplete integration of human, animal and environmental health. Iddri, Issue Brief No. 04/20

Ezenwa VO, Prieur-Richard A-H, Roche B, Bailly X, Becquart P, Garcia-Peña GE, Hosseini PR, Keesing F, Rizzoli A, Suzán GA, Vignuzzi M, Vittecoq M, Mills JN, Guégan J-F (2015) Interdisciplinarity and infectious diseases: an Ebola case study. PLoS Pathog 11:e1004992. https://doi.org/10.1371/journal.ppat.1004992

Van Boeckel TP, Takahashi S, Liao Q, Xing W, Lai S, Hsiao V, Liu F, Zheng Y, Chang Z, Yuan C, Metcalf CJE, Yu H, Grenfell BT (2016) Hand, foot, and mouth disease in China: critical community size and spatial vaccination strategies. Sci Rep 6:25248. https://doi.org/10.1038/srep25248

Holmstrom LK, Beckham TR (2017) Technologies for capturing and analysing animal health data in near real time. Rev Sci Tech 36:525–538

Neethirajan S (2017) Recent advances in wearable sensors for animal health management. Sens Biosensing Res 12:15–29

Perez AM, Zeng D, Tseng CJ, Chen H, Whedbee Z, Paton D, Thurmond MC (2009) A web-based system for near real-time surveillance and space-time cluster analysis of foot-and-mouth disease and other animal diseases. Prev Vet Med 91:39–45. https://doi.org/10.1016/j.prevetmed.2009.05.006

Article   PubMed   Google Scholar  

Wilkinson MD, Dumontier M, Aalbersberg IJ, Appleton G, Axton M, Baak A, Blomberg N, Boiten J-W, da Silva Santos LB, Bourne PE, Bouwman J, Brookes AJ, Clark T, Crosas M, Dillo I, Dumon O, Edmunds S, Evelo CT, Finkers R, Gonzalez-Beltran A, Gray AJG, Groth P, Goble C, Grethe JS, Heringa J, ’t Hoen PAC, Hooft R, Kuhn T, Kok R, Kok J, et al (2016) The FAIR guiding principles for scientific data management and stewardship. Sci Data 3:160018. https://doi.org/10.1038/sdata.2016.18

Binot A, Duboz R, Promburom P, Phimpraphai W, Cappelle J, Lajaunie C, Goutard FL, Pinyopummintr T, Figuié M, Roger FL (2015) A framework to promote collective action within the One Health community of practice: using participatory modelling to enable interdisciplinary, cross-sectoral and multi-level integration. One Health 1:44–48. https://doi.org/10.1016/j.onehlt.2015.09.001

Robert CP (2014) Bayesian computational tools. Annu Rev Stat Appl 1:153–177. https://doi.org/10.1146/annurev-statistics-022513-115543

Dunson DB (2001) Commentary: practical advantages of Bayesian analysis of epidemiologic data. Am J Epidemiol 153:1222–1226. https://doi.org/10.1093/aje/153.12.1222

Uusitalo L (2007) Advantages and challenges of Bayesian networks in environmental modelling. Ecol Model 203:312–318. https://doi.org/10.1016/j.ecolmodel.2006.11.033

Fokoué E (2019) On the ubiquity of the Bayesian paradigm in statistical machine learning and data science. Math Appl 8:189–209. https://doi.org/10.13164/ma.2019.12

Bailly X (2017) Hidden Markov phylogenetic models offer an interesting perspective to identify “high risk lineages” of environmental pathogens. Infect Genet Evol 55:45–47. https://doi.org/10.1016/j.meegid.2017.08.007

Babayan SA, Orton RJ, Streicker DG (2018) Predicting reservoir hosts and arthropod vectors from evolutionary signatures in RNA virus genomes. Science 362:577–580. https://doi.org/10.1126/science.aap9072

Wardeh M, Sharkey KJ, Baylis M (2020) Integration of shared-pathogen networks and machine learning reveals the key aspects of zoonoses and predicts mammalian reservoirs. Proc Biol Sci 287:20192882. https://doi.org/10.1098/rspb.2019.2882

Li J, Zhang S, Li B, Hu Y, Kang X-P, Wu X-Y, Huang M-T, Li Y-C, Zhao Z-P, Qin C-F, Jiang T (2020) Machine learning methods for predicting human-adaptive influenza A viruses based on viral nucleotide compositions. Mol Biol Evol 37:1224–1236. https://doi.org/10.1093/molbev/msz276

Peters DPC, McVey DS, Elias EH, Pelzel-McCluskey AM, Derner JD, Burruss ND, Schrader TS, Yao J, Pauszek SJ, Lombard J, Rodriguez LL (2020) Big data-model integration and AI for vector-borne disease prediction. Ecosphere 11:e03157. https://doi.org/10.1002/ecs2.3157

Lek S, Guégan J-F (2000) Artificial neuronal networks. In: Application to ecology and evolution. Springer, Berlin. https://doi.org/10.1016/j.it.2016.11.006

Go N, Touzeau S, Islam Z, Belloc C, Doeschl-Wilson A (2019) How to prevent viremia rebound? Evidence from a PRRSv data-supported model of immune response. BMC Syst Biol 13:15

Ferrer Savall J, Bidot C, Leblanc-Maridor M, Belloc C, Touzeau S (2016) Modelling Salmonella transmission among pigs from farm to slaughterhouse: interplay between management variability and epidemiological uncertainty. Intern J Food Microbiol 229:33–43. https://doi.org/10.1016/j.ijfoodmicro.2016.03.020

Widgren S, Engblom S, Bauer P, Frössling J, Emanuelson U, Lindberg A (2016) Data-driven network modelling of disease transmission using complete population movement data: spread of VTEC O157 in Swedish cattle. Vet Res 47:81

Qi L, Beaunée G, Arnoux S, Dutta BL, Joly A, Vergu E, Ezanno P (2019) Neighbourhood contacts and trade movements drive the regional spread of bovine viral diarrhoea virus (BVDV). Vet Res 50:30. https://doi.org/10.1186/s13567-019-0647-x

Buhnerkempe MG, Tildesley MJ, Lindström T, Grear DA, Portacci K, Miller RS, Lombard JE, Werkman M, Keeling MJ, Wennergren U, Webb CT (2014) The impact of movements and animal density on continental scale cattle disease outbreaks in the United States. PLoS One 9:e91724. https://doi.org/10.1371/journal.pone.0091724

Ezanno P, Andraud M, Beaunée G, Hoch T, Krebs S, Rault A, Touzeau S, Vergu E, Widgren S (2020) How mechanistic modelling supports decision 1 making for the control of enzootic infectious diseases. Epidemics 32:100398

Garira W (2018) A primer on multiscale modelling of infectious disease systems. Infect Dis Model 3:176–191. https://doi.org/10.1016/j.idm.2018.09.005

Traoré M, Zacharewicz G, Duboz R, Zeigler B (2018) Modeling and simulation framework for value-based healthcare systems. Simulation 95:481–497. https://doi.org/10.1177/0037549718776765

Childs LM, El Moustaid F, Gajewski Z, Kadelka S, Nikin-Beers R, Smith JW Jr, Walker M, Johnson LR (2019) Multi-scale models and data for infectious diseases: a systematic review. PeerJ Preprints 7:e27485v1. https://doi.org/10.7287/peerj.preprints.27485v1

Sandve GK, Nekrutenko A, Taylor J, Hovig E (2013) Ten simple rules for reproducible computational research. PLoS Comput Biol 9:e1003285. https://doi.org/10.1371/journal.pcbi.1003285

Leek JT, Peng RD (2015) Opinion: reproducible research can still be wrong: adopting a prevention approach. Proc Natl Acad Sci USA 112:1645–1646. https://doi.org/10.1073/pnas.1421412111

Widgren S, Bauer P, Eriksson R, Engblom S (2016) SimInf: an R package for data-driven stochastic disease spread simulations. ArXiv160501421 Q-Bio Stat. http://arxiv.org/abs/1605.01421

O’Hare A, Lycett SJ, Doherty TM, Salvador LC, Kao RR (2016) Broadwick: a framework for computational epidemiology. BMC Bioinform 17:65. https://doi.org/10.1186/s12859-016-0903-2

Bui TMA, Stinckwich S, Ziane M, Roche B, Ho TV (2015) KENDRICK: a domain specific language and platform for mathematical epidemiological modelling. In: proc. IEEE RIVF International Conference on Computing and Communication Technologies, Research, Innovation, and Vision for the Future. pp. 132–7. https://doi.org/10.1109/RIVF.2015.7049888

Mathieu P, Morvan G, Picault S (2018) Multi-level agent-based simulations: four design patterns. Simul Model Pract Theory 83:51–64. https://doi.org/10.1016/j.simpat.2017.12.015

Picault S, Huang Y-L, Sicard V, Arnoux S, Beaunée G, Ezanno P (2019) EMULSION: transparent and flexible multiscale stochastic models in human, animal and plant epidemiology. PLoS Comput Biol 15:e1007342. https://doi.org/10.1371/journal.pcbi.1007342

Sebag AS, Plancade S, Raulet-Tomkiewicz C, Barouki R, Vert J-P, Walter T (2015) Inferring an ontology of single cell motions from high-throughput microscopy data. In: Proc. IEEE International Symposium on Biomedical Imaging, Apr. 2015, New-York, USA, pp. 160–163. https://doi.org/10.1109/ISBI.2015.7163840

Tardivel P, Canlet C, Lefort G, Tremblay-Franco M, Debrauwer L, Concordet D, Servien R (2017) ASICS: an automatic method for identification and quantification of metabolites in complex 1D 1H NMR spectra. Metabolomics 13:109

Dórea FC, Muckle CA, Kelton D, McClure JT, McEwen BJ, McNab WB, Sanchez J, Revie CW (2013) Exploratory analysis of methods for automated classification of laboratory test orders into syndromic groups in veterinary medicine. PLoS One 8:e57334. https://doi.org/10.1371/journal.pone.0057334

Gandia P, Jaudet C, Chatelut E, Concordet D (2017) Population pharmacokinetics of tracers: a new tool for medical imaging? Clin Pharmacokinet 56:101–106

Chyba M, Coron J-M, Mileyko Y, Rezaei H (2016) Optimization of prion assemblies fragmentation. In: Proc. IEEE Conference on Decision and Control (CDC), Las Vegas, USA, 6

Nayeri S, Sargolzaei M, Tulpan D (2019) A review of traditional and machine learning methods applied to animal breeding. Anim Health Res Rev 20:31–46. https://doi.org/10.1017/S1466252319000148

Bassaganya-Riera J, Hontecillas R (2016) Introduction to computational immunology. In: Bassaganya-Riera J (ed) Computational immunology: models and tools. pp. 1–8

Liakos KG, Busato P, Moshou D, Pearson S, Bochtis D (2018) Machine learning in agriculture: a review. Sensors 18:2674. https://doi.org/10.3390/s18082674

Jones-Diette JS, Dean RS, Cobb M, Brennan ML (2019) Validation of text-mining and content analysis techniques using data collected from veterinary practice management software systems in the UK. Prev Vet Med 167:61–67. https://doi.org/10.1016/j.prevetmed.2019.02.015

Morota G, Ventura RV, Silva FF, Koyama M, Fernando SC (2018) Big data analytics and precision animal agriculture symposium: machine learning and data mining advance predictive big data analysis in precision animal agriculture. J Anim Sci 96:1540–1550. https://doi.org/10.1093/jas/sky014

Picault S, Ezanno P, Assié S (2019) Combining early hyperthermia detection with metaphylaxis for reducing antibiotics usage in newly received beef bulls at fattening operations: a simulation-based approach. In: Society of veterinary epidemiology and preventive medicine (SVEPM), pp. 13. Utrecht, The Netherland, 27-30/3/2019

Esener N, Green MJ, Emes RD, Jowett B, Davies PL, Bradley AJ, Dottorini T (2018) Discrimination of contagious and environmental strains of Streptococcus uberis in dairy herds by means of mass spectrometry and machine-learning. Sci Rep 8:17517. https://doi.org/10.1038/s41598-018-35867-6

Hepworth PJ, Nefedov AV, Muchnik IB, Morgan KL (2012) Broiler chickens can benefit from machine learning: support vector machine analysis of observational epidemiological data. J R Soc Interface 9:1934–1942. https://doi.org/10.1098/rsif.2011.0852

Marceau A, Madouasse A, Lehébel A, van Schaik G, Veldhuis A, Van der Stede Y, Fourichon C (2014) Can routinely recorded reproductive events be used as indicators of disease emergence in dairy cattle? An evaluation of 5 indicators during the emergence of bluetongue virus in France in 2007 and 2008. J Dairy Sci 97:6135–6150. https://doi.org/10.3168/jds.2013-7346

Fountain-Jones NM, Machado G, Carver S, Packer C, Recamonde-Mendoza M, Craft ME (2019) How to make more from exposure data? An integrated machine learning pipeline to predict pathogen exposure. J Anim Ecol 88:1447–1461. https://doi.org/10.1111/1365-2656.13076

Charras-Garrido M, Azizi L, Forbes F, Doyle S, Peyrard N, Abrial D (2013) On the difficulty to delimit disease risk hot spots. Int J Appl Earth Obs 22:99–105. https://doi.org/10.1016/j.jag.2012.04.005

Forbes F, Charras-Garrido M, Azizi L, Doyle S, Abrial D (2013) Spatial risk mapping for rare disease with hidden Markov fields and variational EM. Annals Appl Stat 7:1192–1216

Arsevska E, Valentin S, Rabatel J, de Goër de Hervé J, Falala S, Lancelot R, Roche M (2018) Web monitoring of emerging animal infectious diseases integrated in the French Animal Health Epidemic Intelligence System. PLoS One 13:0199960. https://doi.org/10.1371/journal.pone.0199960

Küker S, Faverjon C, Furrer L, Berezowski J, Posthaus H, Rinaldi F, Vial F (2018) The value of necropsy reports for animal health surveillance. BMC Vet Res 14:191. https://doi.org/10.1186/s12917-018-1505-1

Bollig N, Clarke L, Elsmo E, Craven M (2020) Machine learning for syndromic surveillance using veterinary necropsy reports. PLoS One 15:e0228105. https://doi.org/10.1371/journal.pone.0228105

Hoscheit P, Geeraert S, Beaunée G, Monod H, Gilligan CAG, Filipe J, Vergu E, Moslonka-Lefebvre M (2016) Dynamical network models for cattle trade: towards economy-based epidemic risk assessment. J Complex Netw 5:604–624. https://doi.org/10.1093/comnet/cnw026

Moslonka-Lefebvre M, Gilligan CA, Monod H, Belloc C, Ezanno P, Filipe JAN, Vergu E (2016) Market analyses of livestock trade networks to inform the prevention of joint economic and epidemiological risks. J R Soc Interface 13:20151099. https://doi.org/10.1098/rsif.2015.1099

Efron B (2020) Prediction, estimation, and attribution. J Am Stat Ass 115:636–655. https://doi.org/10.1080/01621459.2020.1762613

Ghahramani Z (2012) Probabilistic modelling, machine learning, and the information revolution. MIT Computer Science and Artificial Intelligence Lab, http://mlg.eng.cam.ac.uk/zoubin/talks/mit12csail.pdf , Accessed 17 Oct 2019

Hastie T, Tibshirani R, Friedman JH (2009) The elements of statistical learning: data mining, inference, and prediction. 2 nd edn. Springer Series in Statistics. Springer

Goodswen SJ, Kennedy PJ, Ellis JT (2017) On the application of reverse vaccinology to parasitic diseases: a perspective on feature selection and ranking of vaccine candidates. Int J Parasitol 47:779–790. https://doi.org/10.1016/j.ijpara.2017.08.004

Schneider G (2019) Mind and machine in drug design. Nat Mach Intell 1:128–130. https://doi.org/10.1038/s42256-019-0030-7

Beaunée G, Vergu E, Joly A, Ezanno P (2017) Controlling bovine paratuberculosis at a regional scale: towards a decision modeling tool. J Theor Biol 435:157–183. https://doi.org/10.1016/j.jtbi.2017.09.012

Viet A-F, Krebs S, Rat-Aspert O, Jeanpierre L, Belloc C, Ezanno P (2018) A modelling framework based on MDP to coordinate farmers’ disease control decisions at a regional scale. PLoS One 13:e0197612. https://doi.org/10.1371/journal.pone.0197612

Wang T, Hennessy DA (2015) Strategic interactions among private and public efforts when preventing and stamping out a highly infectious animal disease. Am J Agri Econ 97:435–451. https://doi.org/10.1093/ajae/aau119

Tago D, Hammitt JK, Thomas A, Raboisson D (2016) The impact of farmers’ strategic behavior on the spread of animal infectious diseases. PLoS One 11:e0157450. https://doi.org/10.1371/journal.pone.0157450

Probert WJM, Lakkur S, Fonnesbeck CJ, Shea K, Runge MC, Tildesley MJ, Ferrari MJ (2019) Context matters: using reinforcement learning to develop human-readable, state-dependent outbreak response policies. Phil Trans R Soc B 374:20180277. https://doi.org/10.1098/rstb.2018.0277

Liang R, Lu Y, Qu X, Su Q, Li C, Xia S, Liu Y, Zhang Q, Cao X, Chen Q, Niu B (2020) Prediction for global African swine fever outbreaks based on a combination of random forest algorithms and meteorological data. Transbound Emerg Dis 67:935–946. https://doi.org/10.1111/tbed.13424

Salje H, Tran Kiem C, Lefrancq N, Courtejoie N, Bosetti P, Paireau J, Andronico A, Hozé N, Richet J, Dubost C-L, Le Strat Y, Lessler J, Levy Bruhl D, Fontanet A, Opatowski L, Boelle P-Y, Cauchemez S (2020) Estimating the burden of SARS-CoV-2 in France. Science 369:208–211

Parlavantzas N, Pham LM, Morin C, Arnoux S, Beaunée G, Qi L, Gontier P, Ezanno P (2019) A service-based framework for building and executing epidemic simulation applications in the cloud. Concurr Comp Pract Exper 32:e5554. https://doi.org/10.1002/cpe.5554

Shah N, Malensek M, Shah H, Pallickara S, Pallickara SL (2019) Scalable network analytics for characterization of outbreak influence in voluminous epidemiology datasets. Concurr Comp Pract Exper 31:e4998. https://doi.org/10.1002/cpe.4998

Han BA, Majumdar S, Calmon FP, Glicksberg BS, Horesh R, Kumar A, Perer A, von Marschall EB, Wei D, Mojsilović A, Varshney KR (2019) Confronting data sparsity to identify potential sources of Zika virus spillover infection among primates. Epidemics 27:59–65. https://doi.org/10.1016/j.epidem.2019.01.005

Reddy S, Fox J, Purohit MP (2019) Artificial intelligence-enabled healthcare delivery. J R Soc Med 112:22–28

Duboz R, Echaubard P, Promburom P, Kilvington M, Ross H, Allen W, Ward J, Deffuant G, de Garine-Wichatitsky M, Binot A (2018) Systems thinking in practice: participatory modelling as a foundation for integrated approaches to health. Front Vet Sci 5:303. https://doi.org/10.3389/fvets.2018.00303

Van der Waal K, Morrison RB, Neuhauser C, Vilalta C, Perez AM (2017) Translating big data into smart data for veterinary epidemiology. Front Vet Sci 4:110. https://doi.org/10.3389/fvets.2017.00110

Reichman OJ, Jones MB, Schildhauer MP (2011) Challenges and opportunities of open data in ecology. Science 331:703–705. https://doi.org/10.1126/science.1197962

Schultze JL (2015) Teaching ‘big data’ analysis to young immunologists. Nat Immunol 16:902–905

Download references

Acknowledgements

This work has benefited from interactions with many French researchers (Additional file 1 B) interested in the AI/AH interface, for which we thank them here. We also thank Stéphane Abrioux, Didier Concordet and Human Rezaie for participating to the discussions.

PE is supported by the French Research Agency (project CADENCE: ANR-16-CE32-0007). XB is involved in the project “MOnitoring Outbreak events for Disease surveillance in a data science context” supported by the EU Framework Programme for Research and Innovation H2020 (H2020-SC1-BHC-2018–2019, Grant 874850). JFG is supported by both an “Investissement d’Avenir” managed by the French Research Agency (LABEX CEBA: ANR-10-LABX-25-01) and a US NSF-NIH Ecology of infectious diseases award (NSF#1911457), and is also supported by IRD, INRAE, and Université of Montpellier. The funding bodies had no role in the study design, data analysis and interpretation, and manuscript writing.

Author information

Authors and affiliations.

INRAE, Oniris, BIOEPAR, Nantes, France

Pauline Ezanno, Sébastien Picault & Gaël Beaunée

INRAE, EpiA, Theix, France

Xavier Bailly

ASTRE, Univ Montpellier, CIRAD, INRAE, Montpellier, France

Facundo Muñoz, Raphaël Duboz & Jean-François Guégan

Sorbonne Université, IRD, UMMISCO, Bondy, France

Raphaël Duboz

Université Paris-Saclay, INRAE, Jouy-en-Josas, MaIAGE, France

Hervé Monod

MIVEGEC, IRD, CNRS, Univ Montpellier, Montpellier, France

Jean-François Guégan

Comité National Français Sur Les Changements Globaux, Paris, France

You can also search for this author in PubMed   Google Scholar

Contributions

PE carried out the literature review and analysed the interviews. PE and JFG conducted the interviews, drafted and wrote the manuscript. SP, GB, FM, RD, HM provided complementary views and references in their respective disciplines. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Pauline Ezanno .

Ethics declarations

Competing interests.

The authors declare that they have no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1..

Systematic literature review, interviews, previous publication in French.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ . The Creative Commons Public Domain Dedication waiver ( http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article.

Ezanno, P., Picault, S., Beaunée, G. et al. Research perspectives on animal health in the era of artificial intelligence. Vet Res 52 , 40 (2021). https://doi.org/10.1186/s13567-021-00902-4

Download citation

Received : 16 June 2020

Accepted : 20 January 2021

Published : 06 March 2021

DOI : https://doi.org/10.1186/s13567-021-00902-4

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Animal disease
• Artificial intelligence
• Decision support tool

Veterinary Research

ISSN: 1297-9716

• Submission enquiries: [email protected]
• General enquiries: [email protected]

Loading metrics

Open Access

Essays articulate a specific perspective on a topic of broad interest to scientists.

See all article types »

A guide to open science practices for animal research

Contributed equally to this work with: Kai Diederich, Kathrin Schmitt

Affiliation German Federal Institute for Risk Assessment, German Centre for the Protection of Laboratory Animals (Bf3R), Berlin, Germany

* E-mail: [email protected]

• Kai Diederich, 
• Kathrin Schmitt, 
• Philipp Schwedhelm, 
• Bettina Bert, 
• Céline Heinl

Published: September 15, 2022

• https://doi.org/10.1371/journal.pbio.3001810
• Reader Comments

Translational biomedical research relies on animal experiments and provides the underlying proof of practice for clinical trials, which places an increased duty of care on translational researchers to derive the maximum possible output from every experiment performed. The implementation of open science practices has the potential to initiate a change in research culture that could improve the transparency and quality of translational research in general, as well as increasing the audience and scientific reach of published research. However, open science has become a buzzword in the scientific community that can often miss mark when it comes to practical implementation. In this Essay, we provide a guide to open science practices that can be applied throughout the research process, from study design, through data collection and analysis, to publication and dissemination, to help scientists improve the transparency and quality of their work. As open science practices continue to evolve, we also provide an online toolbox of resources that we will update continually.

Citation: Diederich K, Schmitt K, Schwedhelm P, Bert B, Heinl C (2022) A guide to open science practices for animal research. PLoS Biol 20(9): e3001810. https://doi.org/10.1371/journal.pbio.3001810

Copyright: © 2022 Diederich et al. This is an open access article distributed under the terms of the Creative Commons Attribution License , which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The authors received no specific funding for this work.

Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: All authors are employed at the German Federal Institute for Risk Assessment and part of the German Centre for the Protection of Laboratory Animals (Bf3R) which developed and hosts animalstudyregistry.org , a preregistration platform for animal studies and animaltestinfo.de, a database for non-technical project summaries (NTS) of approved animal study protocols within Germany.

Abbreviations: CC, Creative Commons; CIRS-LAS, critical incident reporting system in laboratory animal science; COVID-19, Coronavirus Disease 2019; DOAJ, Directory of Open Access Journals; DOI, digital object identifier; EDA, Experimental Design Assistant; ELN, electronic laboratory notebook; EU, European Union; IMSR, International Mouse Strain Resource; JISC, Joint Information Systems Committee; LIMS, laboratory information management system; MGI, Mouse Genome Informatics; NC3Rs, National Centre for the Replacement, Refinement and Reduction of Animals in Research; NTS, non-technical summary; RRID, Research Resource Identifier

Introduction

Over the past decade, the quality of published scientific literature has been repeatedly called into question by the failure of large replication studies or meta-analyses to demonstrate sufficient translation from experimental research into clinical successes [ 1 – 5 ]. At the same time, the open science movement has gained more and more advocates across various research areas. By sharing all of the information collected during the research process with colleagues and with the public, scientists can improve collaborations within their field and increase the reproducibility and trustworthiness of their work [ 6 ]. Thus, the International Reproducibility Networks have called for more open research [ 7 ].

However, open science practices have not been adopted to the same degree in all research areas. In psychology, which was strongly affected by the so-called reproducibility crisis, the open science movement initiated real practical changes leading to a broad implementation of practices such as preregistration or sharing of data and material [ 8 – 10 ]. By contrast, biomedical research is still lagging behind. Open science might be of high value for research in general, but in translational biomedical research, it is an ethical obligation. It is the responsibility of the scientist to transparently share all data collected to ensure that clinical research can adequately evaluate the risks and benefits of a potential treatment. When Russell and Burch published “The Principles of Humane Experimental Technique” in 1959, scientists started to implement their 3Rs principle to answer the ethical dilemma of animal welfare in the face of scientific progress [ 11 ]. By replacing animal experiments wherever possible, reducing the number of animals to a strict minimum, and refining the procedures where animals have still to be used, this ethical dilemma was addressed. However, in recent years, whether the 3Rs principle is sufficient to fully address ethical concerns about animal experiments has been questioned [ 12 ].

Most people tolerate the use of animals for scientific purposes only under the basic assumption that the knowledge gained will advance research in crucial areas. This implies that performed experiments are reported in a way that enables peers to benefit from the collected data. However, recent studies suggest that a large proportion of animal experiments are never actually published. For example, scientists working within the European Union (EU) have to write an animal study protocol for approval by the competent authorities of the respective country before performing an animal experiment [ 13 ]. In these protocols, scientists have to describe the planned study and justify every animal required for the project. By searching for publications resulting from approved animal study protocols from 2 German University Medical Centers, Wieschowski and colleagues found that only 53% of approved protocols led to a publication after 6 years [ 14 ]. Using a similar approach, Van der Naald and colleagues determined a publication rate of 60% at the Utrecht Medical Center [ 15 ]. In a follow-up survey, the respective researchers named so-called “negative” or null-hypothesis results as the main cause for not publishing outcomes [ 15 ]. The current scientific system is shaped by publishers, funders, and institutions and motivates scientists to publish novel, surprising, and positive results, revealing one of the many structural problems that the numerous efforts towards open science initiatives are targeting. Non-publication not only strongly contradicts ethical values, but also it compromises the quality of published literature by leading to overestimation of effect sizes [ 16 , 17 ]. Furthermore, publications of animal studies too often show poor reporting that strongly impairs the reproducibility, validity, and usefulness of the results [ 18 ]. Unfortunately, the idea that negative or equivocal findings can also contribute to the gain of scientific knowledge is frequently neglected.

So far, the scientific community using animals has shown limited resonance to the open science movement. Due to the strong controversy surrounding animal experiments, scientists have been reluctant to share information on the topic. Additionally, translational research is highly competitive and researchers tend to be secretive about their ideas until they are ready for publication or patent [ 19 , 20 ]. However, this missing openness could also point to a lack of knowledge and training on the many open science options that are available and suitable for animal research. Researchers have to be convinced of the benefits of open science practices, not only for science in general, but also for the individual researcher and each single animal. Yet, the key players in the research system are already starting to value open science practices. An increasing number of journals request open sharing of data, funders pay for open access publications and institutions consider open science practices in hiring decisions. Open science practices can improve the quality of work by enabling valuable scientific input from peers at the early stages of research projects. Furthermore, the extended communication that open science practices offer can draw attention to research and help to expand networks of collaborators and lead to new project opportunities or follow-up positions. Thus, open science practices can be a driver for careers in academia, particularly those of early career researchers.

Beyond these personal benefits, improving transparency in translational biomedical research can boost scientific progress in general. By bringing to light all the recorded research outputs that until now have remained hidden, the publication bias and the overestimation of effect sizes can be reduced [ 17 ]. Large-scale sharing of data can help to synthesize research outputs in preclinical research that will enable better decision-making for clinical research. Disclosing the whole research process will help to uncover systematic problems and support scientists in thoroughly planning their studies. In the long run, we predict that the implementation of open science practices will lead to the use of fewer animals in unintentionally repeated experiments that previously showed unreported negative results or in the establishment of methods by avoiding experimental dead ends that are often not published. More collaborations and sharing of materials and methods can further reduce the number of animal experiments used for the implementation of new techniques.

Open science can and should be implemented at each step of the research process ( Fig 1 ). A vast number of tools are already provided that were either directly conceptualized for animal research or can be adapted easily. In this Essay, we provide an overview of open science tools that improve transparency, reliability, and animal welfare in translational in vivo biomedical research by supporting scientists to clearly communicate their research and by supporting collaborative working. Table 1 lists the most prominent open science tools we discuss, together with their respective links. We have structured this Essay to guide you through which tools can be used at each stage of the research process, from planning and conducting experiments, through to analyzing data and communicating the results. However, many of these tools can be used at many different steps. Table 1 has been deposited on Zenodo and will be updated continuously [ 21 ].

• PPT PowerPoint slide
• PNG larger image
• TIFF original image

Application of open science practices at each step of the research process can maximize the impact of performed animal experiments. The implementation of these practices will lead to less time pressure at the end of a project. Due to the connection of most of these open science practices, spending more time in the planning phase and during the conduction of experiments will save time during the data analysis and publication of the study. Indeed, consulting reporting guidelines early on, preregistering a statistical plan, and writing down crucial experimental details in an electronic lab notebook, will strongly accelerate the writing of a manuscript. If protocols or even electronic lab notebooks were made public, just citing these would simplify the writing of publications. Similarly, if a data management plan is well designed before starting data collection, analyzing, and depositing data in a public repository, as is increasingly required, will be fast. NTS, non-technical summary.

https://doi.org/10.1371/journal.pbio.3001810.g001

https://doi.org/10.1371/journal.pbio.3001810.t001

Planning the study

Transparent practices can be adopted at every stage of the research process. However, to ensure full effectivity, it is highly recommended to engage in detailed planning before the start of the experiment. This can prevent valuable time from being lost at the end of the study due to careless decisions being made at the beginning. Clarifying data management at the start of a project can help avoiding filing chaos that can be very time consuming to untangle. Keeping clear track of a project and study design will also help if new colleagues are included later on in the project or if entire project parts are handed over. In addition, all texts written on the rationale and hypothesis of the study or method descriptions, or design schemes created during the planning phase can be used in the final publications ( Fig 1 ). Similarly, information required for preregistration of animal studies or for reporting according to the ARRIVE guidelines are an extension of the details required for ethical approval [ 22 , 23 ]. Thus, the time burden within the planning phase is often overestimated. Furthermore, the thorough planning of experiments can avoid the unnecessary use of animals by preventing wrong avenues from being pursued.

Implementing open scientific practices at the beginning of a project does not mean that the idea and study plan must be shared immediately, but rather is critical for making the entire workflow transparent at the end of the project. However, optional early sharing of information can enable peers to give feedback on the study plan. Studies potentially benefit more from this a priori input than they would from the classical a posteriori peer-review process.

Most people perceive guidelines as advice that instructs on how to do something. However, it is sometimes useful to consider the term in its original meaning; “the line that guides us”. In this sense, following guidelines is not simply fulfilling a duty, but is a process that can help to design a sound research study and, as such, guidelines should be consulted at the planning stage of a project. The PREPARE guidelines are a list of important points that should be thought-out before starting a study involving animal experiments in order to reduce the waste of animals, promote alternatives, and increase the reproducibility of research and testing [ 24 ]. The PREPARE checklist helps to thoroughly plan a study and focuses on improving the communication and collaboration between all involved participants of the study (i.e., animal caretakers and scientists). Indeed, open science begins with the communication within a research facility. It is currently available in 33 languages and the responsible team from Norecopa, Norway’s 3R-center, takes requests for translations into further languages.

The UK Reproducibility Network has also published several guiding documents (primers) on important topics for open and reproducible science. These address issues such as data sharing [ 25 ], open access [ 26 ], open code and software [ 27 ], and preprints [ 28 ], as well as preregistration and registered reports [ 27 ]. Consultation of these primers is not only helpful in the relevant phases of the experiment but is also encouraged in the planning phase.

Although the ARRIVE guidelines are primarily a reporting guideline specifically designed for preparing a publication containing animal data, they can also support researchers when planning their experiments [ 22 , 23 ]. Going through the ARRIVE website, researchers will find tools and explanations that can support them in planning their experiments [ 29 ]. Consulting the ARRIVE checklist at the beginning of a project can help in deciding what details need to be documented during conduction of the experiments. This is particularly advisable, given that compliance to ARRIVE is still poor [ 18 ].

Experimental design

To maximize the validity of performed experiments and the knowledge gained, designing the study well is crucial. It is important that the chosen animal species reflects the investigated disease well and that basic characteristics of the animal, such as sex or age, are considered carefully [ 30 ]. The Canadian Institutes of Health Research provides a collection of resources on the integration of sex and gender in biomedical research with animals, including tips and tools for researchers and reviewers [ 31 ]. Additionally, it is advisable to avoid unnecessary standardization of biological and environmental factors that can reduce the external validity of results [ 32 ]. Meticulous statistical planning can further optimize the use of animals. Free to use online tools for calculating sample sizes such as G*Power or the inVivo software package for R can further support animal researchers in designing their statistical plan [ 33 , 34 ]. Randomization for the allocation of groups can be supported with specific tools for scientists like Research Randomizer, but also by simple online random number generators [ 35 ]. Furthermore, it might be advisable when designing the study to incorporate pathological analyses into the experimental plan. Optimal planning of tissue collection, performance of pathological procedures according to accepted best practices, and use of optimal pathological analysis and reporting methods can add some extra knowledge that would otherwise be lost. This can improve the reproducibility and quality of translational biomedicine, especially, but not exclusively, in animal studies with morphological endpoints. In all animal studies, unexpected deaths in experimental animals can occur and be the cause of lost data or missed opportunities to identify health problems [ 36 , 37 ].

To support researchers in designing their animal research, the National Centre for the Replacement, Refinement and Reduction of Animals in Research (NC3Rs) has also developed the Experimental Design Assistant (EDA) [ 38 , 39 ]. This online tool helps researchers to better structure in vivo research by creating detailed schemes of the study design. It provides feedback on the entered design, drawing researcher’s attention to crucial decisions in the project. The resulting schemes can be used to transparently share the study design by uploading it into a study preregistration, enclosing it in a grant application, or submitting it with a final manuscript. The EDA can be used for different study designs in diverse scenarios and helps to communicate researcher plans to others [ 40 ]. The EDA might be particularly of interest to clarify very complex study designs involving multiple experimental groups. Working with the EDA might appear rather complex in the beginning, but the NC3R provides regular webinars that can help to answer any questions that arise.

Preregistration

Preregistration is an effective tool to improve the quality and transparency of research. To preregister their work, scientists must determine crucial details of the study before starting any experiment. Changes occurring during a study can be outlined at the end. A preregistered study plan should include at least the hypothesis and determine all the parameters that are known in advance. A description of the planned study design and statistical analysis will enable reviewers and peers to better retrace the workflow. It can prevent the intentional use of the flexibility of analysis to reach p -values under a certain significance level (e.g., p-hacking or HARKing (Hypothesizing After Results are Known)). With preregistration, scientists can also claim their idea at an early stage of their research with a citable individual identifier that labels the idea as their own. Some open preregistration platforms also provide a digital object identifier (DOI), which makes the registered study citable. Three public registries actively encourage the preregistration of animal studies conducted around the world: OSF registry, preclinicaltrials.eu, and animalstudyregistry.org [ 41 – 45 ]. Scientists can choose the registry according to their needs. Preregistering a study in a public registry supports scientists in planning their study and later to critically reevaluate their own work and assess its limitations and potentials.

As an alternative to public registries, researchers can also submit their study plan to one of hundreds of journals already publishing registered reports, including many journals open to animal research [ 8 ]. A submitted registered report passes 2 steps of peer review. In the first step, reviewers comment on the idea and the study design. After an “in-principle-acceptance,” researchers can conduct their study as planned. If the authors conduct the experiments as described in the accepted study protocol, the journal will publish the final study regardless of the outcome. This might be an attractive option, especially for early career researchers, as a manuscript is published at the beginning of a project with the guarantee of a future final publication.

The benefits of preregistration can already be observed in clinical research, where registration has been mandatory for most trials for more than 20 years. Preregistration in clinical research has helped to make known what has been tested and not just what worked and was published, and the implementation of trial registration has strongly reduced the number of publications reporting significant treatment effects [ 46 ]. In animal research, with its unrealistically high percentage of positive results, preregistration seems to be particularly worthwhile.

Research data management

To get the most out of performed animal experiments, effective sharing of data at the end of the study is essential. Sharing research data optimally is complex and needs to be prepared in advance. Thus, data management can be seen as one part of planning a study thoroughly. Many funders have recognized the value of the original research data and request a data management plan from applicants in advance [ 25 , 47 ]. Various freely available tools such as DMPTool or DMPonline already exist to design a research data management plan that complies to the requirements of different funders [ 48 , 49 ]. The data management plan defines the types of data collected and describes the handling and names responsible persons throughout the data lifecycle. This includes collecting the data, analyzing, archiving, and sharing it. Finally, a data management plan enables long-term access and the possibility for reuse by peers. Developing such a plan, whether it is required by funders or not, will later simplify the application of the FAIR data principle (see section on the FAIR data principle). The Longwood Medical Area Research Data Management Working Group from the Harvard Medical School developed a checklist to assist researchers in optimally managing their data throughout the data lifecycle [ 50 ]. Similarly, the Joint Information Systems Committee (JISC) provides a great research data management toolkit including a checklist for researchers planning their project [ 51 ]. Consulting this checklist in the planning phase of a project can prevent common errors in research data management.

Non-technical project summary

One instrument specifically conceived to create transparency on animal research for the general public is the so-called non-technical project summary (NTS). All animal protocols approved within the EU must be accompanied by these comprehensible summaries. NTSs are intended to inform the public about ongoing animal experiments. They are anonymous and include information on the objectives and potential benefits of the project, the expected harm, the number of animals, the species, and a statement of compliance with the requirements of the 3Rs principle. However, beyond simply informing the public, NTSs can also be used for meta-research to help identify new research areas with an increased need for new 3R technologies [ 52 , 53 ]. NTSs become an excellent tool to appropriately communicate the scientific value of the approved protocol and for meta-scientists to generate added value by systematically analyzing theses summaries if they fulfill a minimum quality threshold [ 54 , 55 ]. In 2021, the EU launched the ALURES platform ( Table 1 ), where NTSs from all member states are published together, opening the opportunities for EU-wide meta-research. NTSs are, in contrast to other open science practices, mandatory in the EU. However, instead of thinking of them as an annoying duty, it might be worth thoroughly drafting the NTS to support the goals of more transparency towards the public, enabling an open dialogue and reducing extreme opinions.

Conducting the experiments

Once the experiments begin, documentation of all necessary details is essential to ensure the transparency of the workflow. This includes methodological details that are crucial for replicating experiments, but also failed attempts that could help peers to avoid experiments that do not work in the future. All information should be stored in such a way that it can be found easily and shared later. In this area, many new tools have emerged in recent years ( Table 1 ). These tools will not only make research transparent for colleagues, but also help to keep track of one’s own research and improve internal collaboration.

Electronic laboratory notebooks

Electronic laboratory notebooks (ELNs) are an important pillar of research data management and open science. ELNs facilitate the structured and harmonized documentation of the data generation workflow, ensure data integrity, and keep track of all modifications made to the original data based on an audit trail option. Moreover, ELNs simplify the sharing of data and support collaborations within and outside the research group. Methodological details and research data become searchable and traceable. There is an extensive amount of literature providing advice on the selection and the implementation process of an ELN depending on the specific needs and research area and its discussion would be beyond the scope of this Essay [ 56 – 58 ]. Some ELNs are connected to a laboratory information management system (LIMS) that provides an animal module supporting the tracking of animal details [ 59 ]. But as research involving animals is highly heterogeneous, this might not be the only decision point and we cannot recommend a specific ELN that is suitable for all animal research.

ELNs are already established in the pharmaceutical industry and their use is on the rise among academics as well. However, due to concerns around costs for licenses, data security, and loss of flexibility, many research institutions still fear the expenses that the introduction of such a system would incur [ 56 ]. Nevertheless, an increasing number of academic institutions are implementing ELNs and appreciating the associated benefits [ 60 ]. If your institution already has an ELN, it might be easiest to just use the option available in the research environment. If not, the Harvard Medical School provides an extensive and updated overview of various features of different ELNs that can support scientists in choosing the appropriate one for their research [ 61 ]. There are many commercial ELN products, which may be preferred when the administrative workload should be outsourced to a large extent. However, open-source products such as eLabFTW or open BIS provide a greater opportunity for customization to meet specific needs of individual research institutions [ 62 – 64 ]. A huge number of options are available depending on the resources and the features required. Some scientists might prefer generic note taking tools such as Evernote or just a simple Word document that offers infinite flexibility, but specific ELNs can further support good record keeping practice by providing immutability, automated backups, standardized methods, and protocols to follow. Clearly defining the specific requirements expected might help to choose an adequate system that would improve the quality of the record compared to classical paper laboratory notebooks.

Sharing protocols

Adequate sharing of methods in translational biomedical sciences is key to reproducibility. Several repositories exist that simplify the publication and exchange of protocols. Writing down methods at the end of the project bears the risk that crucial details might be missing [ 65 ]. On protocols.io, scientists can note all methodological details of a procedure, complete them with uploaded documents, and keep them for personal use or share them with collaborators [ 66 ]. Authors can also decide at any point in time to make their protocol public. Protocols published on protocols.io receive a DOI and become citable; they can be commented on by peers and adapted according to the needs of the individual researcher. Protocol.io files from established protocols can also be submitted together with some context and sample datasets to PLOS ONE , where it can be peer-reviewed and potentially published [ 67 , 68 ]. Depending on the affiliation of the researchers to academia or industry and on an internal or public sharing of files, protocols.io can be free of charge or come with costs. Other journals also encourage their authors to deposit their protocols in a freely accessible repository, such as protocol exchange from Nature portfolio [ 69 ]. Another option might be to separately submit a protocol that was validated by its use in an already published research article to an online and peer-reviewed journal specific for research protocols, such as Bio-Protocol. A multitude of journals, including eLife and Science already collaborate with Bio-Protocol and recommend authors to publish the method in Bio-Protocol [ 70 ]. Bio-Protocol has no submission fees and is freely available to all readers. Both protocols.io and Bio-Protocol allow the illustration of complex scientific methods by uploading videos to published protocols. In addition, protocols can be deposited in a general research repository such as the Open Science Framework (OSF repository) and referenced in appropriate publications.

Sharing critical incidents

Sharing critical or even adverse events that occur in the context of animal experimentation can help other scientists to avoid committing the same mistakes. The system of sharing critical incidents is already established in clinical practice and helps to improve medical care [ 71 , 72 ]. The online platform critical incident reporting system in laboratory animal science (CIRS-LAS) represents the first preclinical equivalent to these clinical systems [ 73 ]. With this web-based tool, critical incidents in animal research can be reported anonymously without registration. An expert panel helps to analyze the incident to encourage an open dialogue. Critical incident reporting is still very marginal in animal research and performed procedures are very variable. These factors make a systemic analysis and a targeted search of incidence difficult. However, it may be of special interest for methods that are broadly used in animal research such as anesthesia. Indeed, a broad feed of this system with data on errors occurring in standard procedures today could help avoid critical incidences in the future and refine animal experiments.

Sharing animals, organs, and tissue

When we think about open science, sharing results and data are often in focus. However, sharing material is also part of a collaborative and open research culture that could help to greatly reduce the number of experimental animals used. When an animal is killed to obtain specific tissue or organs, the remainder is mostly discarded. This may constitute a wasteful practice, as surplus tissue can be used by other researchers for different analyses. More animals are currently killed as surplus than are used in experiments, demonstrating the potential for sharing these animals [ 74 , 75 ].

Sharing information on generated surplus is therefore not only economical, but also an effective way to reduce the number of animals used for scientific purposes. The open-source software Anishare is a straightforward way for breeders of genetically modified lines to promote their surplus offspring or organs within an institution [ 76 ]. The database AniMatch ( Table 1 ) connects scientists within Europe who are offering tissue or organs with scientists seeking this material. Scientists already sharing animal organs can support this process by describing it in publications and making peers aware of this possibility [ 77 ]. Specialized research communities also allow sharing of animal tissue or animal-derived products worldwide that are typically used in these fields on a collaborative basis via the SEARCH-framework [ 78 , 79 ]. Depositing transgenic mice lines into one of several repositories for mouse strains can help to further minimize efforts in producing new transgenic lines and most importantly reduce the number of surplus animals by supporting the cryoconservation of mouse lines. The International Mouse Strain Resource (IMSR) can be used to help find an adequate repository or to help scientists seeking a specific transgenic line find a match [ 80 ].

Analyzing the data

Animal researchers have to handle increasingly complex data. Imaging, electrophysiological recording, or automated behavioral tracking, for example, produce huge datasets. Data can be shared as raw numerical output but also as images, videos, sounds, or other forms from which raw numerical data can be generated. As the heterogeneity and the complexity of research data increases, infinite possibilities for analysis emerge. Transparently reporting how the data were processed will enable peers to better interpret reported results. To get the most out of performed animal experiments, it is crucial to allow other scientists to replicate the analysis and adapt it to their research questions. It is therefore highly recommended to use formats and tools during the analysis that allow a straightforward exchange of code and data later on.

Transparent coding

The use of non-transparent analysis codes have led to a lack of reproducibility of results [ 81 ]. Sharing code is essential for complex analysis and enables other researchers to reproduce results and perform follow-up studies, and citable code gives credit for the development of new algorithms ( Table 1 ). Jupyter Notebooks are a convenient way to share data science pipelines that may use a variety of coding languages, including like Python, R or Matlab, and also share the results of analyses in the form of tables, diagrams, images, and videos. Notebooks contain source code and can be published or collaboratively shared on platforms like GitHub or GitLab, where version control of source code is implemented. The data-archiving tool Zenodo can be used to archive a repository on GitHub and create a DOI for the archive. Thereby contents become citable. Using free and open-source programming language like R or Python will increase the number of potential researchers that can work with the published code. Best practice for research software is to publish the source code with a license that allows modification and redistribution.

Choice of data visualization

Choosing the right format for the visualization of data can increase its accessibility to a broad scientific audience and enable peers to better judge the validity of the results. Studies based on animal research often work with very small sample sizes. Visualizing these data in histograms may lead to an overestimation of the outcomes. Choosing the right dot plots that makes all recorded points visible and at the same time focusses on the summary instead of the individual points can further improve the intuitive understanding of a result. If the sample size is too low, it might not be meaningful to visualize error bars. A variety of freely available tools already exists that can support scientists in creating the most appropriate graphs for their data [ 82 ]. In particular, when representing microscopy results or heat maps, it should be kept in mind that a large part of the population cannot perceive the classical red and green representation [ 83 ]. Opting for the color-blind safe color maps and checking images with free tools such as color oracle ( Table 1 ) can increase the accessibility of graphs. Multiple journals have already addressed flaws in data visualization and have introduced new policies that will accelerate the uptake of transparent representation of results.

Publication of all study outcomes

Open science practices have received much attention in the past few years when it comes to publication of the results. However, it is important to emphasize that although open science tools have their greatest impact at the end of the project, good study preparation and sharing of the study plan and data early on can greatly increase the transparency at the end.

The FAIR data principle

To maximize the impact and outcome of a study, and to make the best long-term use of data generated through animal experiments, researchers should publish all data collected during their research according to the FAIR data principle. That means the data should be findable, accessible, interoperable, and reusable. The FAIR principle is thus an extension of open access publishing. Data should not only be published without paywalls or other access restrictions, but also in such a manner that they can be reused and further processed by others. For this, legal as well as technical requirements must be met by the data. To achieve this, the GoFAIR initiative has developed a set of principles that should be taken into account as early as at the data collection stage [ 49 , 84 ]. In addition to extensively described and machine-readable metadata, these principles include, for example, the application of globally persistent identifiers, the use of open file formats, and standardized communication protocols to ensure that humans and machines can easily download the data. A well-chosen repository to upload the data is then just the final step to publish FAIR data.

FAIR data can strongly increase the knowledge gained from performed animal experiments. Thus, the same data can be analyzed by different researchers and could be combined to obtain larger sample sizes, as already occurs in the neuroimaging community, which works with comparable datasets [ 85 ]. Furthermore, the sharing of data enables other researchers to analyze published datasets and estimate measurement reliabilities to optimize their own data collection [ 86 , 87 ]. This will help to improve the translation from animal research into clinics and simultaneously reduce the number of animal experiment in future.

Reporting guidelines

In preclinical research, the ARRIVE guidelines are the current state of the art when it comes to reporting data based on animal experiments [ 22 , 23 ]. The ARRIVE guidelines have been endorsed by more than 1,000 journals who ask that scientists comply with them when reporting their outcomes. Since the ARRIVE guidelines have not had the expected impact on the transparency of reporting in animal research publications, a more rigorous update has been developed to facilitate their application in practice (ARRIVE 2.0 [ 23 ]). We believe that the ARRIVE guidelines can be more effective if they are implemented at a very early stage of the project (see section on guidelines). Some more specialized reporting guidelines have also emerged for individual research fields that rely on animal studies, such as endodontology [ 88 ]. The equator network collects all guidelines and makes them easily findable with their search tool on their website ( Table 1 ). MERIDIAN also offers a 1-stop shop for all reporting guidelines involving the use of animals across all research sectors [ 89 ]. It is thus worth checking for new reporting guidelines before preparing a manuscript to maximize the transparency of described experiments.

Identifiers

Persistent identifiers for published work, authors, or resources are key for making public data findable by search engines and are thus a prerequisite for compliance to FAIR data principles. The most common identifier for publications will be a DOI, which makes the work citable. A DOI is a globally unique string assigned by the International DOI Foundation to identify content permanently and provide a persistent link to its location on the Internet. An ORCID ID is used as a personal persistent identifier and is recommendable to unmistakably identify an author ( Table 1 ). This will avoid confusions between authors with the same name or in the case of name changes or changes of affiliation. Research Resource Identifiers (RRID) are unique ID numbers that help to transparently report research resources. RRID also apply to animals to clearly identify the species used. RRID help avoid confusion between different names or changing names of genetic lines and, importantly, make them machine findable. The RRID Portal helps scientists find a specific RRID or create one if necessary ( Table 1 ). In the context of genetically altered animal lines, correct naming is key. The Mouse Genome Informatics (MGI) Database is the authoritative source of official names for mouse genes, alleles, and strains ([ 90 ]).

Preprint publication

Preprints have undergone unprecedented success, particularly during the height of the Coronavirus Disease 2019 (COVID-19) pandemic when the need for rapid dissemination of scientific knowledge was critical. The publication process for scientific manuscripts in peer-reviewed journals usually requires a considerable amount of time, ranging from a few months to several years, mainly due to the lengthy review process and inefficient editorial procedures [ 91 , 92 ]. Preprints typically precede formal publication in scientific journals and, thus, do not go through a peer review process, thus, facilitating the prompt open dissemination of important scientific findings within the scientific community. However, submitted papers are usually screened and checked for plagiarism. Preprints are assigned a DOI so they can be cited. Once a preprint is published in a journal, its status is automatically updated on the preprint server. The preprint is linked to the publication via CrossRef and mentioned accordingly on the website of the respective preprint platform.

After initial skepticism, most publishers now allow papers to be posted on preprint servers prior to submission. An increasing number of journals even allow direct submission of a preprint to their peer review process. The US National Institutes of Health and the Wellcome Trust, among other funders, also encourage prepublication and permit researchers to cite preprints in their grant applications. There are now numerous preprint repositories for different scientific disciplines. BioASAP provides a searchable database for preprint servers that can help in identifying the one that best matches an individual’s needs [ 93 ]. The most popular repository for animal research is bioRxiv, which is hosted by the Cold Spring Harbor Laboratory ( Table 1 ).

The early exchange of scientific results is particularly important for animal research. This acceleration of the publication process can help other scientists to adapt their research or could even prevent animal experiments if other scientists become aware that an experiment has already been done before starting their own. In addition, preprints can help to increase the visibility of research. Journal articles that have a corresponding preprint publication have higher citation and Altmetric counts than articles without preprint [ 94 ]. In addition, the publication of preprints can help to combat publication bias, which represents a major problem in animal research [ 16 ]. Since journals and readers prioritize cutting-edge studies with positive results over inconclusive or negative results, researchers are reluctant to invest time and money in a manuscript that is unlikely to be accepted in a high-impact journal.

In addition to the option of publishing as preprint, other alternative publication formats have recently been introduced to facilitate the publication of research results that are hard to publish in traditional peer-reviewed journals. These include micro publications, data repositories, data journals, publication platforms, and journals that focus on negative or inconclusive results. The tool fiddle can support scientists in choosing the right publication format [ 95 , 96 ].

Open access publication

Publishing open access is one of the most established open science strategies. In contrast to the FAIR data principle, the term open access publication refers usually to the publication of a manuscript on a platform that is accessible free of charge—in translational biomedical research, this is mostly in the form of a scientific journal article. Originally, publications accessible free of charge were the answer to the paywalls established by renowned publishing houses, which led to social inequalities within and outside the research system. In translational biomedical research, the ethical aspect of urgently needed transparency is another argument in favor of open access publication, as these studies will not only be findable, but also internationally readable.

There are different ways of open access publishing; the 2 main routes are gold open access and green open access. Numerous journals offer now gold open access. It refers to the immediate and fully accessible publication of an article. The Directory of Open Access Journals (DOAJ) provides a complete and updated list for high-quality, open access, and peer-reviewed journals [ 97 ]. Charité–Universitätsmedizin Berlin offers a specific tool for biomedical open access journals that supports animal researchers to choose an appropriate journal [ 49 ]. In addition, the Sherpa Romeo platform is a straightforward way to identify publisher open access policies on a journal-by-journal basis, including information on preprints, but also on licensing of articles [ 51 ]. Hybrid open access refers to openly accessible articles in otherwise paywalled journals. By contrast, green open access refers to the publication of a manuscript or article in a repository that is mostly operated by institutions and/or universities. The publication can be exclusively on the repository or in combination with a publisher. In the quality-assured, global Directory of Open Access Repositories (openDOAR), scientists can find thousands of indexed open access repositories [ 49 ]. The publisher often sets an embargo during which the authors cannot make the publication available in the repository, which can restrict the combined model. It is worth mentioning that gold open access is usually more expensive for the authors, as they have to pay an article processing charge. However, the article’s outreach is usually much higher than the outreach of an article in a repository or available exclusively as subscription content [ 98 ]. Diamond open access refers to publications and publication platforms that can be read free of charge by anyone interested and for which no costs are incurred by the authors either. It is the simplest and fairest form of open access for all parties involved, as no one is prevented from participating in scientific discourse by payment barriers. For now, it is not as widespread as the other forms because publishers have to find alternative sources of revenue to cover their costs.

As social media and the researcher’s individual public outreach are becoming increasingly important, it should be remembered that the accessibility of a publication should not be confused with the licensing under which the publication is made available. In order to be able to share and reuse one’s own work in the future, we recommend looking for journals that allow publications under the Creative Commons licenses CC BY or CC BY-NC. This also allows the immediate combination of gold and green open access.

Creative commons licenses

Attributing Creative Commons (CC) licenses to scientific content can make research broadly available and clearly specifies the terms and conditions under which people can reuse and redistribute the intellectual property, namely publications and data, while giving the credit to whom it deserves [ 49 ]. As the laws on copyright vary from country to country and law texts are difficult to understand for outsiders, the CC licenses are designed to be easily understandable and are available in 41 languages. This way, users can easily avoid accidental misuse. The CC initiative developed a tool that enables researchers to find the license that best fits their interests [ 49 ]. Since the licenses are based on a modular concept ranging from relatively unrestricted licenses (CC BY, free to use, credit must be given) to more restricted licenses (CC BY-NC-ND, only free to share for non-commercial purposes, credit must be given), one can find an appropriate license even for the most sensitive content. Publishing under an open CC license will not only make the publication easy to access but can also help to increase its reach. It can stimulate other researchers and the interested public to share this article within their network and to make the best future use of it. Bear in mind that datasets published independently from an article may receive a different CC license. In terms of intellectual property, data are not protected in the same way as articles, which is why the CC initiative in the United Kingdom recommends publishing them under a CC0 (“no rights reserved”) license or the Public Domain Mark. This gives everybody the right to use the data freely. In an animal ethics sense, this is especially important in order to get the most out of data derived from animal experiments.

Data and code repositories

Sharing research data is essential to ensure reproducibility and to facilitate scientific progress. This is particularly true in animal research and the scientific community increasingly recognizes the value of sharing research data. However, even though there is increasing support for the sharing of data, researchers still perceive barriers when it comes to doing so in practice [ 99 – 101 ]. Many universities and research institutions have established research data repositories that provide continuous access to datasets in a trusted environment. Many of these data repositories are tied to specific research areas, geographic regions, or scientific institutions. Due to the growing number and overall heterogeneity of these repositories, it can be difficult for researchers, funding agencies, publishers, and academic institutions to identify appropriate repositories for storing and searching research data.

Recently, several web-based tools have been developed to help in the selection of a suitable repository. One example is Re3data, a global registry of research data repositories that includes repositories from various scientific disciplines. The extensive database can be searched by country, content (e.g., raw data, source code), and scientific discipline [ 49 ]. A similar tool to help find a data archive specific to the field is FAIRsharing, based at Oxford University [ 102 ]. If there is no appropriate subject-specific data repository or one seems unsuitable for the data, there are general data repositories, such as Open Science Framework, figshare, Dryad, or Zenodo. To ensure that data stored in a repository can be found, a DOI is assigned to the data. Choosing the right license for the deposited code and data ensures that authors get credit for their work.

Publication and connection of all outcomes

If scientists have used all available open science tools during the research process, then publishing and linking all outcomes represents the well-deserved harvest ( Fig 2 ). At the end of a research process, researchers will not just have 1 publication in a journal. Instead, they might have a preregistration, a preprint, a publication in a journal, a dataset, and a protocol. Connecting these outcomes in a way that enables other scientists to better assess the results that link these publications will be key. There are many examples of good open science practices in laboratory animal science, but we want to highlight one of them to show how this could be achieved. Blenkuš and colleagues investigated how mild stress-induced hyperthermia can be assessed non-invasively by thermography in mice [ 103 ]. The study was preregistered with animalstudyregistry.org , which is referred to in their publication [ 104 ]. A deviation from the originally preregistered hypothesis was explained in the manuscript and the supplementary material was uploaded to figshare [ 105 ].

Application of open science practices can increase the reproducibility and visibility of a research project at the same time. By publishing different research outputs with more detailed information than can be included in a journal article, researchers enable peers to replicate their work. Reporting according to guidelines and using transparent visualization will further improve this reproducibility. The more research products that are generated, the more credit can be attributed. By communicating on social media or additionally publishing slides from delivered talks or posters, more attention can be raised. Additionally, publishing open access and making the work machine-findable makes it accessible to an even broader number of peers.

https://doi.org/10.1371/journal.pbio.3001810.g002

It might also be helpful to provide all resources from a project in a single repository such as Open Science Framework, which also implements other, different tools that might have been used, like GitHub or protocols.io.

Communicating your research

Once all outcomes of the project are shared, it is time to address the targeted peers. Social media is an important instrument to connect research communities [ 106 ]. In particular, Twitter is an effective way to communicate research findings or related events to peers [ 107 ]. In addition, specialized platforms like ResearchGate can support the exchange of practical experiences ( Table 1 ). When all resources related to a project are kept in one place, sharing this link is a straightforward way to reach out to fellow scientists.

With the increasing number of publications, science communication has become more important in recent years. Transparent science that communicates openly with the public contributes to strengthening society’s trust in research.

Conclusions

Plenty of open science tools are already available and the number of tools is constantly growing. Translational biomedical researchers should seize this opportunity, as it could contribute to a significant improvement in the transparency of research and fulfil their ethical responsibility to maximize the impact of knowledge gained from animal experiments. Over and above this, open science practices also bear important direct benefits for the scientists themselves. Indeed, the implementation of these tools can increase the visibility of research and becomes increasingly important when applying for grants or in recruitment decisions. Already, more and more journals and funders require activities such as data sharing. Several institutions have established open science practices as evaluation criteria alongside publication lists, impact factor, and h-index for panels deciding on hiring or tenure [ 108 ]. For new adopters, it is not necessary to apply all available practices at once. Implementing single tools can be a safe approach to slowly improve the outreach and reproducibility of one’s own research. The more open science products that are generated, the more reproducible the work becomes, but also the more the visibility of a study increases ( Fig 2 ).

As other research fields, such as social sciences, are already a step ahead in the implementation of open science practices, translational biomedicine can profit from their experiences [ 109 ]. We should thus keep in mind that open science comes with some risks that should be minimized early on. Indeed, the more open science practices become incentivized, the more researchers could be tempted to get a transparency quality label that might not be justified. When a study is based on a bad hypothesis or poor statistical planning, this cannot be fixed by preregistration, as prediction alone is not sufficient to validate an interpretation [ 110 ]. Furthermore, a boom of data sharing could disconnect data collectors and analysts, bearing the risk that researchers performing the analysis lack understanding of the data. The publication of datasets could also promote a “parasitic” use of a researcher’s data and lead to scooping of outcomes [ 111 ]. Stakeholders could counteract such a risk by promoting collaboration instead of competition.

During the COVID-19 pandemic, we have seen an explosion of preprint publications. This unseen acceleration of science might be the adequate response to a pandemic; however, the speeding up science in combination with the “publish or perish” culture could come at the expense of the quality of the publication. Nevertheless, a meta-analysis comparing the quality of reporting between preprints and peer-reviewed articles showed that the quality of reporting in preprints in the life sciences is at most slightly lower on average compared to peer-reviewed articles [ 112 ]. Additionally, preprints and social media have shown during this pandemic that a premature and overconfident communication of research results can be overinterpreted by journalists and raise unfounded hopes or fears in patients and relatives [ 113 ]. By being honest and open about the scope and limitations of the study and choosing communication channels carefully, researchers can avoid misinterpretation. It should be noted, however, that by releasing all methodological details and data in research fields such as viral engineering, where a dual use cannot be excluded, open science could increase biosecurity risk. Implementing access-controlled repositories, application programming interfaces, and a biosecurity risk assessment in the planning phase (i.e., by preregistration) could mitigate this threat [ 114 ].

Publishing in open access journals often involves higher publication costs, which makes it more difficult for institutes and universities from low-income countries to publish there [ 115 ]. Equity has been identified as a key aim of open science [ 116 ]. It is vital, therefore, that existing structural inequities in the scientific system are not unintentionally reinforced by open science practices. Early career researchers have been the main drivers of the open science movement in other fields even though they are often in vulnerable positions due to short contracts and hierarchical and strongly networked research environments. Supporting these early career researchers in adopting open science tools could significantly advance this change in research culture [ 117 ]. However, early career researchers can already benefit by publishing registered reports or preprints that can provide a publication much faster than conventional journal publications. Communication in social media can help them establish a network enabling new collaborations or follow-up positions.

Even though open science comes with some risks, the benefits easily overweigh these caveats. If a change towards more transparency is accompanied by the implementation of open science in the teaching curricula of the universities, most of the risks can be minimized [ 118 ]. Interestingly, we have observed that open science tools and infrastructure that are specific to animal research seem to mostly come from Europe. This may be because of strict regulations within Europe for animal experiments or because of a strong research focus in laboratory animal science along with targeted research funding in this region. Whatever the reason might be, it demonstrates the important role of research policy in accelerating the development towards 3Rs and open science.

Overall, it seems inevitable that open science will eventually prevail in translational biomedical research. Scientists should not wait for the slow-moving incentive framework to change their research habits, but should take pioneering roles in adopting open science tools and working towards more collaboration, transparency, and reproducibility.

Acknowledgments

The authors gratefully acknowledge the valuable input and comments from Sebastian Dunst, Daniel Butzke, and Nils Körber that have improved the content of this work.

• View Article
• PubMed/NCBI
• Google Scholar
• 6. Cary Funk MH, Brian Kennedy, Courtney Johnson. Americans say open access to data and independent review inspire more trust in research findings. Pew Research Center Website: Pew Research Center; 2019. Available from: https://www.pewresearch.org/science/2019/08/02/americans-say-open-access-to-data-and-independent-review-inspire-more-trust-in-research-findings/ .
• 7. International Reproducibility Networks. International Networks Statement UK Reproducibility Network Website: UK Reproducibility Network. 2021. Available from: https://cpb-eu-w2.wpmucdn.com/blogs.bristol.ac.uk/dist/b/631/files/2021/09/International-Networks-Statement-v1.0.pdf .
• 13. Article 36 of Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 amended by Regilation (EU) 2019/1010 of the European Parliament and of the Council of 5 June 2019. OJEU. 2010;L276:36.
• 19. American Association for Cancer Research. Editorial Policies. 2021. Available from: https://aacrjournals.org/content/authors/editorial-policies .
• 21. Diederich K, Schmitt K, Schwedhelm P, Bert B, Heinl C. Open Science Toolbox for Animal Research. Zenodo. 2022. Available from: https://zenodo.org/record/6497560 .
• 29. NC3R. ARRIVE guidelines. NC3R Website. Available from: https://arriveguidelines.org/ .
• 31. Canadian Institutes of Health Research. How to integrate sex and gender into research. Website of the Canadian Institutes of Health Research: Canadian Institutes of Health Research. 2019 [cited 2019 Aug 21]. Available from: https://cihr-irsc.gc.ca/e/50836.html .
• 33. Simon T, Bate RAC. InVivoStat. Available from: https://invivostat.co.uk/ .
• 35. Urbaniak G, Plous S. Research randomizer (version 4.0) [computer software]. 2013.
• 47. Medical Research Council’s. Data sharing policy. UK Research and Innovation Website 2021. Available from: https://www.ukri.org/publications/mrc-data-sharing-policy/ .
• 48. University of California Curation Center. DMPTool. 2021. Available from: https://dmptool.org/ .
• 49. Digital Curation Centre. DMPOnline. Available from: https://dmponline.dcc.ac.uk/ . Digital Curation Centre; 2021.
• 50. Harvard Longwood Medical Area Research Data Management Working Group. Biomedical Data Lifecycle. Harvard Medical School Website: Harvard Medical School; 2021. Available from: https://datamanagement.hms.harvard.edu/about/what-research-data-management/biomedical-data-lifecycle .
• 51. Joint Information Systems Committee. Research data management toolkit JISC Website: JISC; 2018. Available from: https://www.jisc.ac.uk/guides/rdm-toolkit .
• 54. German Centre for the Protection of Laboratory Animals (Bf3R). NTPs—Nicht Technische Projektzusammenfassungen 3R-SMART; 2020. Available from: https://www.3r-smart.de/index.php?id=6895 .
• 55. Understanding Animal Research. Guide to writing non-technical summaries concordat on openness on animal research in the UK2018. Available from: https://concordatopenness.org.uk/guide-to-writing-non-technical-summaries .
• 56. Gerlach B, Untucht C, Stefan A. Electronic Lab Notebooks and Experimental Design Assistants. In: Bespalov A, Michel MC, Steckler T, editors. Good Research Practice in Non-Clinical Pharmacology and Biomedicine. Cham: Springer International Publishing; 2020. p. 257–75.
• 58. Adam BL, Birte L. ELN Guide: electronic laboratory notebooks in the context of research data management and good research practice–a guide for the life sciences. Cologne, Germany: ZB MED–Information Centre for Life Sciences; 2021.
• 59. AgileBio. LabCollector Website https://labcollector.com/labcollector-lims/features/modules/animals-module/2022 . Available from: https://labcollector.com/labcollector-lims/features/modules/animals-module/ .
• 61. Harvard Longwood Medical Area Research Data Management Working Group. Electronic Lab Notebook Comparison Matrix. Zenodo. 2021.
• 70. Bio-protocol. Collaborating Journals bio-protocol website2021. Available from: https://bio-protocol.org/default.aspx?dw=Collaborating .
• 76. Dinkel H. anishare: GitHub; [updated June 2018]. Available from: https://github.com/hdinkel/anishare .
• 89. O’Connor AM. MERIDIAN: Menagerie of Reporting guidelines Involving Animals. Iowa State University; 2022. Available from: https://meridian.cvm.iastate.edu/ .
• 90. The Jackson Laboratory. Mouse Nomenclature Home Page at the Mouse Genome Informatics website World Wide Web: The Jackson Laboratory,Bar Harbor, Maine. Available from: http://www.informatics.jax.org/mgihome/nomen/index.shtml .
• 97. Directory of Open Access Journals. Find open access journals & articles. Available from: https://doaj.org/ . Directory of Open Access Journals, [DOAJ]; 2021.
• 98. Gold Open Access research has greater societal impact as used more outside of academia [press release]. Springer Nature Website: Springer. Nature. 2020;30:2020.
• 104. Franco NH. Can we use infrared thermography for assessing emotional states in mice?—A comparison between handling-induced stress by different techniques. Available from: animalstudyregistry.org . German Federal Institute for Risk Assessment (BfR); 2020. https://doi.org/10.17590/asr.0000224

Key Topics in Surgical Research and Methodology pp 207–228 Cite as

An Introduction to Animal Research

• James Kinross 3 &
• Lord Ara Darzi 4  

2858 Accesses

Despite advances in computer modelling and bioinformatics, animal models remain a vital component of biomedical research. The growth in this area of work is in part due to the evolution next generation of biotechnologies, which more than ever necessitate the need for in vivo experimentation. An understanding of the principals of animal research therefore remains a necessity for medical researchers as it permits scientific analysis to be interpreted in a more critical and meaningful manner. Initiating and designing an animal experiment can be a daunting process, particularly as the law and legislation governing animal research is complex and new specialist skills must be acquired. This chapter reviews the principles of animal research and provides a practical resource for those researchers seeking to create robust animal experiments that ensure minimal suffering and maximal scientific validity.

• Animal Research
• Human Embryonic Stem Cell
• Occupational Asthma
• Scientific Procedure
• Veterinary Surgeon

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

This is a preview of subscription content, log in via an institution .

Buying options

• Available as PDF
• Read on any device
• Instant download
• Own it forever
• Available as EPUB and PDF
• Compact, lightweight edition
• Dispatched in 3 to 5 business days
• Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Abbreviations

The American College of Laboratory Animal Medicine

Animal and Plant Health Inspection Service

Animal Welfare Act

Control of substances hazardous to health

European Coalition for Biomedical Research

The Food and Drug Agency

Health and Safety Executive

The Human Fertilisation and Embryology Authority

Institutional Animal Care and Use Committee

Individually ventilated cage

In vitro fertilisation

Laboratory animal allergy

Named animal care and welfare officer

National Institute for Clinical Excellence

Named veterinary surgeon

The Office of Laboratory Animal Welfare

Public Health Service

Personal License under the Scientific (Animal Procedures) Act 1986

Project License under the Scientific (Animal Procedures) Act 1986

Royal Society for the Prevention of Cruelty to Animals

Specified pathogen free

The United States Department of Agriculture

(2006) Animals (scientific procedures) inspectorate annual report. Home Office: Science, Research and Statistics, London

Google Scholar  

Davidson N (2006) Davidson review: implementation of EU legislation. http://wwwhm-treasurygovuk/independent_reviews/davidson_review/davidson_indexcfm

Fox JG, Cohen BJ, Loew FM (1984) Laboratory animal medicine. Academic press, New York

RDS, Understanding Animal Research in Medicine, Coalition for Medical Progress (2007) Medical advances and animal research: the contribution of animal science to the medical revolution: some case histories. Available at: http://www.pro-test.org.uk/MAAR.pdf

Foex BA (2007) The ethics of animal experimentation. Emerg Med J 24:750–751

Article   PubMed   Google Scholar  

Matfield M (2002) Animal experimentation: the continuing debate. Nat Rev Drug Discov 1:149–152

Rollin BE (2007) Animal research: a moral science. Talking Point on the use of animals in scientific research. EMBO Rep 8:521–525

Article   PubMed   CAS   Google Scholar  

Singer P (2006) In defense of animals: the second wave. Blackwell, Malden, MA

Croce P (1999) Vivisection or science? An investigation into testing drugs and safeguarding health. Zed Books, London

Schechter AN, Rettig RA (2002) Funding priorities for medical research. JAMA 288:832; author reply 832

Pound P, Ebrahim S, Sandercock P et al (2004) Where is the evidence that animal research benefits humans? BMJ 328:514–517

Perel P, Roberts I, Sena E et al (2007) Comparison of treatment effects between animal experiments and clinical trials: systematic review. BMJ 334:197

Mayor S (2007) UK regulatory body wants public consultation on human—animal hybrid research. BMJ 334:112

Anon (2007) Avoiding a chimaera quagmire. Nature 445:1

Article   Google Scholar  

Karpowicz P, Cohen CB, van der Kooy D (2004) It is ethical to transplant human stem cells into nonhuman embryos. Nat Med 10:331–335

Thompson SG (1994) Why sources of heterogeneity in meta-analysis should be investigated. BMJ 309:1351–1355

Auer JA, Goodship A, Arnoczky S et al (2007) Refining animal models in fracture research: seeking consensus in optimising both animal welfare and scientific validity for appropriate biomedical use. BMC Musculoskelet Disord 8:72

Sarker SK, Patel B (2007) Simulation and surgical training. Int J Clin Pract 61:2120–2125

Dolan K (2007) Laboratory animal law: legal control of the use of animals in research. Blackwell, New York

Book   Google Scholar  

Rice M (2007) Deadline approaches for animal experimentation directive. Eur J Cancer 43:1641

PubMed   Google Scholar  

Field K, Bailey M, Foresman LL et al (2007) Medical records for animals used in research, teaching, and testing: public statement from the American College of Laboratory Animal Medicine. ILAR J 48:37–41

Hauser R, Marczak M, Karaszewski B et al (2008) A preliminary study for identifying olfactory markers of fear in the rat. Lab Anim (NY) 37:76–80

Vermeulen JK, de Vries A, Schlingmann F, Remie R. (1997) Food Deprivation: common sense or nonsense? Animal Tech 48:45–54

Levine S, Saltzman A (2000) Feeding sugar overnight maintains metabolic homeostasis in rats and is preferable to overnight starvation. Lab Anim 34:301–306

Brown C (2007) Endotracheal intubation in the dog. Lab Anim (NY) 36:23–24

Price H (2007) Intubating rabbits. Vet Rec 160:744

Spoelstra EN, Ince C, Koeman A et al (2007) A novel and simple method for endotracheal intubation of mice. Lab Anim 41:128–135

Buscaglia JM (2007) Animal laboratory endoscopic research: a fellow's perspective. Gastrointest Endosc 65:882–883

Fritscher-Ravens A, Patel K, Ghanbari A et al (2007) Natural orifice transluminal endoscopic surgery (NOTES) in the mediastinum: long-term survival animal experiments in transesoph-ageal access, including minor surgical procedures. Endoscopy 39:870–875

Radostits OM, Mayhew IG, Houston DM (2000) Veterinary clinical examination and diagnosis, 2nd edn. W.B. Saunders Company, Philadelphia

James Thomson et al, “Embryonic stem cell lines derived from human blastocysts,” Science 282:1145–1147, November 6, 1998

The National Academies' Guidelines for Human Embryonic Stem Cell Research. 2008. http://books.nap.edu/catalog.php?record_id=12260

Download references

Author information

Authors and affiliations.

Department of Biosurgery and Surgical Technology, Imperial College, 10th floor, QEQM, St. Mary's Hospital, Praed Street, London, W2 1NY, UK

James Kinross

The Department of Biosurgery and Surgical Technology, Imperial College London, 10th Floor, Queen Elizabeth the Queen Mother (QEQM) Building, Imperial College Healthcare NHS Trust, St Mary's Hospital Campus, Praed Street, London, W2 1NY, UK

Lord Ara Darzi

You can also search for this author in PubMed   Google Scholar

Editor information

Editors and affiliations.

Dept. Biosurgery & Surgical Technology, Imperial College London St. Mary's Hospital London, 10th floor QEQM Bldg. Praed Street, London, W2 1NY, United Kingdom

Thanos Athanasiou  & Ara Darzi  & 

UCSF Global Health Sciences, 3333 California Street, Suite 285, San Francisco L, CA, 94143-0443, USA

Haile Debas

Rights and permissions

Reprints and permissions

Copyright information

© 2010 Springer-Verlag Berlin Heidelberg

About this chapter

Cite this chapter.

Kinross, J., Darzi, L.A. (2010). An Introduction to Animal Research. In: Athanasiou, T., Debas, H., Darzi, A. (eds) Key Topics in Surgical Research and Methodology. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-71915-1_17

Download citation

DOI : https://doi.org/10.1007/978-3-540-71915-1_17

Publisher Name : Springer, Berlin, Heidelberg

Print ISBN : 978-3-540-71914-4

Online ISBN : 978-3-540-71915-1

eBook Packages : Medicine Medicine (R0)

Share this chapter

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Publish with us

Policies and ethics

• Find a journal
• Track your research

• Our Program Divisions
• Our Three Academies
• Government Affairs
• Statement on Diversity and Inclusion
• Our Study Process
• Conflict of Interest Policies and Procedures
• Project Comments and Information
• Read Our Expert Reports and Published Proceedings
• Explore PNAS, the Official Scientific Journal of NAS
• Access Transportation Research Board Publications
• Coronavirus Disease 2019 (COVID-19)
• Diversity, Equity, and Inclusion
• Economic Recovery
• Fellowships and Grants
• Publications by Division
• Division of Behavioral and Social Sciences and Education
• Division on Earth and Life Studies
• Division on Engineering and Physical Sciences
• Gulf Research Program
• Health and Medicine Division
• Policy and Global Affairs Division
• Transportation Research Board
• National Academy of Sciences
• National Academy of Engineering
• National Academy of Medicine
• Publications by Topic
• Agriculture
• Behavioral and Social Sciences
• Biography and Autobiography
• Biology and Life Sciences
• Computers and Information Technology
• Conflict and Security Issues
• Earth Sciences
• Energy and Energy Conservation
• Engineering and Technology
• Environment and Environmental Studies
• Food and Nutrition
• Health and Medicine
• Industry and Labor
• Math, Chemistry, and Physics
• Policy for Science and Technology
• Space and Aeronautics
• Surveys and Statistics
• Transportation and Infrastructure
• Searchable Collections
• New Releases

VIEW LARGER COVER

Animals for Research

A directory of sources, tenth edition and supplement.

• Biology and Life Sciences — Laboratory Animal Research

Suggested Citation

National Research Council. 1979. Animals for Research: A Directory of Sources, Tenth Edition and Supplement . Washington, DC: The National Academies Press. https://doi.org/10.17226/1674. Import this citation to: Bibtex EndNote Reference Manager

Publication Info

What is skim.

The Chapter Skim search tool presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter. You may select key terms to highlight them within pages of each chapter.

Copyright Information

The National Academies Press (NAP) has partnered with Copyright Clearance Center's Marketplace service to offer you a variety of options for reusing NAP content. Through Marketplace, you may request permission to reprint NAP content in another publication, course pack, secure website, or other media. Marketplace allows you to instantly obtain permission, pay related fees, and print a license directly from the NAP website. The complete terms and conditions of your reuse license can be found in the license agreement that will be made available to you during the online order process. To request permission through Marketplace you are required to create an account by filling out a simple online form. The following list describes license reuses offered by the NAP through Marketplace:

• Republish text, tables, figures, or images in print
• Post on a secure Intranet/Extranet website
• Use in a PowerPoint Presentation
• Distribute via CD-ROM

Click here to obtain permission for the above reuses. If you have questions or comments concerning the Marketplace service, please contact:

Marketplace Support International +1.978.646.2600 US Toll Free +1.855.239.3415 E-mail: [email protected] marketplace.copyright.com

To request permission to distribute a PDF, please contact our Customer Service Department at [email protected] .

What is a prepublication?

An uncorrected copy, or prepublication, is an uncorrected proof of the book. We publish prepublications to facilitate timely access to the committee's findings.

What happens when I pre-order?

The final version of this book has not been published yet. You can pre-order a copy of the book and we will send it to you when it becomes available. We will not charge you for the book until it ships. Pricing for a pre-ordered book is estimated and subject to change. All backorders will be released at the final established price. As a courtesy, if the price increases by more than $3.00 we will notify you. If the price decreases, we will simply charge the lower price. Applicable discounts will be extended.

Downloading and Using eBooks from NAP

What is an ebook.

An ebook is one of two file formats that are intended to be used with e-reader devices and apps such as Amazon Kindle or Apple iBooks.

Why is an eBook better than a PDF?

A PDF is a digital representation of the print book, so while it can be loaded into most e-reader programs, it doesn't allow for resizable text or advanced, interactive functionality. The eBook is optimized for e-reader devices and apps, which means that it offers a much better digital reading experience than a PDF, including resizable text and interactive features (when available).

Where do I get eBook files?

eBook files are now available for a large number of reports on the NAP.edu website. If an eBook is available, you'll see the option to purchase it on the book page.

View more FAQ's about Ebooks

Types of Publications

Consensus Study Report: Consensus Study Reports published by the National Academies of Sciences, Engineering, and Medicine document the evidence-based consensus on the study’s statement of task by an authoring committee of experts. Reports typically include findings, conclusions, and recommendations based on information gathered by the committee and the committee’s deliberations. Each report has been subjected to a rigorous and independent peer-review process and it represents the position of the National Academies on the statement of task.

• Search Menu
• Advance articles
• Author Guidelines
• Submission Site
• Open Access
• Why Submit?
• About Oxford Journal of Legal Studies
• About University of Oxford Faculty of Law
• Editorial Board
• Advertising and Corporate Services
• Journals Career Network
• Self-Archiving Policy
• Dispatch Dates
• Journals on Oxford Academic
• Books on Oxford Academic

Article Contents

1. introduction: the need for legal animal rights theory, 2. can animals have legal rights, 3. do animals have (simple) legal rights, 4. should animals have (fundamental) legal rights, 5. conclusion.

• < Previous

Towards a Theory of Legal Animal Rights: Simple and Fundamental Rights

• Article contents
• Figures & tables
• Supplementary Data

Saskia Stucki, Towards a Theory of Legal Animal Rights: Simple and Fundamental Rights, Oxford Journal of Legal Studies , Volume 40, Issue 3, Autumn 2020, Pages 533–560, https://doi.org/10.1093/ojls/gqaa007

• Permissions Icon Permissions

With legal animal rights on the horizon, there is a need for a more systematic theorisation of animal rights as legal rights. This article addresses conceptual, doctrinal and normative issues relating to the nature and foundations of legal animal rights by examining three key questions: can, do and should animals have legal rights? It will show that animals are conceptually possible candidates for rights ascriptions. Moreover, certain ‘animal welfare rights’ could arguably be extracted from existing animal welfare laws, even though these are currently imperfect and weak legal rights at best. Finally, this article introduces the new conceptual vocabulary of simple and fundamental animal rights, in order to distinguish the weak legal rights that animals may be said to have as a matter of positive law from the kind of strong legal rights that animals ought to have as a matter of future law.

Legal animal rights are on the horizon, and there is a need for a legal theory of animal rights—that is, a theory of animal rights as legal rights. While there is a diverse body of moral and political theories of animal rights, 1 the nature and conceptual foundations of legal animal rights remain remarkably underexplored. As yet, only few and fragmented legal analyses of isolated aspects of animal rights exist. 2 Other than that, most legal writing in this field operates with a hazily assumed, rudimentary and undifferentiated conception of animal rights—one largely informed by extralegal notions of moral animal rights—which tends to obscure rather than illuminate the distinctive nature and features of legal animal rights. 3 A more systematic and nuanced theorisation of legal animal rights is, however, necessary and overdue for two reasons: first, a gradual turn to legal rights in animal rights discourse; and, secondly, the incipient emergence of legal animal rights.

First, while animal rights have originally been framed as moral rights, they are increasingly articulated as potential legal rights. That is, animals’ moral rights are asserted in an ‘ought to be legal rights’-sense (or ‘manifesto sense’) 4 that demands legal institutionalisation and refers to the corresponding legal rights which animals should ideally have. 5 A salient reason for transforming moral into legal animal rights is that purely moral rights (which exist prior to and independently of legal validation) do not provide animals with sufficient practical protection, whereas legally recognised rights would be reinforced by the law’s more stringent protection and enforcement mechanisms. 6 With a view to their (potential) juridification, it seems advisable to rethink and reconstruct animal rights as specifically legal rights, rather than simply importing moral animal rights into the legal domain. 7

Secondly, and adding urgency to the need for theorisation, legal animal rights are beginning to emerge from existing law. Recently, a few pioneering courts have embarked on a path of judicial creation of animal rights, arriving at them either through a rights-based interpretation of animal welfare legislation or a dynamic interpretation of constitutional (human) rights. Most notably, the Supreme Court of India has extracted a range of animal rights from the Prevention of Cruelty to Animals Act and, by reading them in the light of the Constitution, elevated those statutory rights to the status of fundamental rights. 8 Furthermore, courts in Argentina 9 and Colombia 10 have extended the fundamental right of habeas corpus , along with the underlying right to liberty, to captive animals. 11 These (so far isolated) acts of judicial recognition of animal rights may be read as early manifestations of an incipient formation of legal animal rights. Against this backdrop, there is a pressing practical need for legal animal rights theory, in order to explain and guide the as yet still nascent—and somewhat haphazard—evolution of legal animal rights.

This article seeks to take the first steps towards building a more systematic and nuanced theory of legal animal rights. Navigating the existing theoretical patchwork, the article revisits and connects relevant themes that have so far been addressed only in a scattered or cursory manner, and consolidates them into an overarching framework for legal animal rights. Moreover, tackling the well-known problem of ambiguity and obscurity involved in the generally vague, inconsistent and undifferentiated use of the umbrella term ‘animal rights’, this article brings analytical clarity into the debate by disentangling and unveiling different meanings and facets of legal animal rights. 12 To this end, the analysis identifies and separates three relevant sets of issues: (i) conceptual issues concerning the nature and foundations of legal animal rights, and, more generally, whether animals are the kind of beings who can potentially hold legal rights; (ii) doctrinal issues pertaining to existing animal welfare law and whether it confers some legal rights on animals—and, if so, what kind of rights; and (iii) normative issues as to why and what kind of legal rights animals ought ideally to have as a matter of future law. These thematic clusters will be addressed through three simple yet key questions: can , do and should animals have legal rights?

Section 2 will show that it is conceptually possible for animals to hold legal rights, and will clarify the formal structure and normative grounds of legal animal rights. Moreover, as section 3 will demonstrate, unwritten animal rights could arguably be extracted from existing animal welfare laws, even though such ‘animal welfare rights’ are currently imperfect and weak legal rights at best. In order to distinguish between these weak legal rights that animals may be said to have as a matter of positive law and the kind of strong legal rights that animals ought to have potentially or ideally, the new conceptual categories of ‘ simple animal rights’ and ‘ fundamental animal rights’ will be introduced. Finally, section 4 will explore a range of functional reasons why animals need such strong, fundamental rights as a matter of future law.

As a preliminary matter, it seems necessary to first address the conceptual issue whether animals potentially can have legal rights, irrespective of doctrinal and normative issues as to whether animals do in fact have, or should have, legal rights. Whether animals are possible or potential right holders—that is, the kind of beings to whom legal rights can be ascribed ‘without conceptual absurdity’ 13 —must be determined based on the general nature of rights, which is typically characterised in terms of the structure (or form) and grounds (or ultimate purpose) of rights. 14 Looking at the idea of animal rights through the lens of general rights theories helps clarify the conceptual foundations of legal animal rights by identifying their possible forms and grounds. The first subsection (A) focusses on two particular forms of conceptually basic rights—claims and liberties—and examines their structural compatibility with animal rights. The second subsection (B) considers the two main competing theories of rights—the will theory and interest theory—and whether, and on what grounds, they can accommodate animals as potential right holders.

A. The Structure of Legal Animal Rights

The formal structure of rights is generally explicated based on the Hohfeldian typology of rights. 15 Hohfeld famously noted that the generic term ‘right’ tends to be used indiscriminately to cover ‘any sort of legal advantage’, and distinguished four different types of conceptually basic rights: claims (rights stricto sensu ), liberties, powers and immunities. 16 In the following, I will show on the basis of first-order rights 17 —claims and liberties—that legal animal rights are structurally possible, and what such legal relations would consist of. 18

(i) Animal claim rights

To have a right in the strictest sense is ‘to have a claim to something and against someone’, the claim right necessarily corresponding with that person’s correlative duty towards the right holder to do or not to do something. 19 This type of right would take the form of animals holding a claim to something against, for example, humans or the state who bear correlative duties to refrain from or perform certain actions. Such legal animal rights could be either negative rights (correlative to negative duties) to non-interference or positive rights (correlative to positive duties) to the provision of some good or service. 20 The structure of claim rights seems especially suitable for animals, because these are passive rights that concern the conduct of others (the duty bearers) and are simply enjoyed rather than exercised by the right holder. 21 Claim rights would therefore assign to animals a purely passive position that is specified by the presence and performance of others’ duties towards animals, and would not require any actions by the animals themselves.

(ii) Animal liberties

Liberties, by contrast, are active rights that concern the right holder’s own conduct. A liberty to engage in or refrain from a certain action is one’s freedom of any contrary duty towards another to eschew or undertake that action, correlative to the no right of another. 22 On the face of it, the structure of liberties appears to lend itself to animal rights. A liberty right would indicate that an animal is free to engage in or avoid certain behaviours, in the sense of being free from a specific duty to do otherwise. Yet, an obvious objection is that animals are generally incapable of having any legal duties. 23 Given that animals are inevitably in a constant state of ‘no duty’ and thus ‘liberty’, 24 this seems to render the notion of liberty rights somewhat pointless and redundant in the case of animals, as it would do nothing more than affirm an already and invariably existing natural condition of dutylessness. However, this sort of ‘natural liberty’ is, in and of itself, only a naked liberty, one wholly unprotected against interferences by others. 25 That is, while animals may have the ‘natural liberty’ of, for example, freedom of movement in the sense of not having (and not being capable of having) a duty not to move around, others do not have a duty vis-à-vis the animals not to interfere with the exercise of this liberty by, for example, capturing and caging them.

The added value of turning the ‘natural liberties’ of animals into liberty rights thus lies in the act of transforming unprotected, naked liberties into protected, vested liberties that are shielded from certain modes of interference. Indeed, it seems sensible to think of ‘natural liberties’ as constituting legal rights only when embedded in a ‘protective perimeter’ of claim rights and correlative duties within which such liberties may meaningfully exist and be exercised. 26 This protective perimeter consists of some general duties (arising not from the liberty right itself, but from other claim rights, such as the right to life and physical integrity) not to engage in ‘at least the cruder forms of interference’, like physical assault or killing, which will preclude most forms of effective interference. 27 Moreover, liberties may be fortified by specific claim rights and correlative duties strictly designed to protect a particular liberty, such as if the state had a (negative) duty not to build highways that cut across wildlife habitat, or a (positive) duty to build wildlife corridors for such highways, in order to facilitate safe and effective freedom of movement for the animals who live in these fragmented habitats.

(iii) Animal rights and duties: correlativity and reciprocity

Lastly, some remarks on the relation between animal rights and duties seem in order. Some commentators hold that animals are unable to possess legal rights based on the influential idea that the capacity for holding rights is inextricably linked with the capacity for bearing duties. 28 Insofar as animals are not capable of bearing legal duties in any meaningful sense, it follows that animals cannot have legal (claim) rights against other animals, given that those other animals would be incapable of holding the correlative duties. But does this disqualify animals from having legal rights altogether, for instance, against legally competent humans or the state?

While duties are a key component of (first-order) rights—with claim rights necessarily implying the presence of a legal duty in others and liberties necessarily implying the absence of a legal duty in the right holder 29 —neither of them logically entails that the right holder bear duties herself . As Kramer aptly puts it:

Except in the very unusual circumstances where someone holds a right against himself, X’s possession of a legal right does not entail X’s bearing of a legal duty; rather, it entails the bearing of a legal duty by somebody else. 30

This underscores an important distinction between the conceptually axiomatic correlativity of rights and duties—the notion that every claim right necessarily implies a duty—and the idea of a reciprocity of rights and duties—the notion that (the capacity for) right holding is conditioned on (the capacity for) duty bearing. While correlativity refers to an existential nexus between a right and a duty held by separate persons within one and the same legal relation , reciprocity posits a normative nexus between the right holding and duty bearing of one and the same person within separate, logically unrelated legal relations.

The claim that the capacity for right holding is somehow contingent on the right holder’s (logically unrelated) capacity for duty bearing is thus, as Kramer puts it, ‘straightforwardly false’ from a Hohfeldian point of view. 31 Nevertheless, there may be other, normative reasons (notably underpinned by social contract theory) for asserting that the class of appropriate right holders should be limited to those entities that, in addition to being structurally possible right holders, are also capable of reciprocating, that is, of being their duty bearers’ duty bearers. 32 However, such a narrow contractarian framing of right holding should be rejected, not least because it misses the current legal reality. 33 With a view to legally incompetent humans (eg infants and the mentally incapacitated), contemporary legal systems have manifestly cut the connection between right holding and the capacity for duty bearing. 34 As Wenar notes, the ‘class of potential right holders has expanded to include duty-less entities’. 35 Similarly, it would be neither conceptually nor legally apposite to infer from the mere fact that animals do not belong to the class of possible duty bearers that they cannot belong to the class of possible right holders. 36

B. The Grounds of Legal Animal Rights

While Hohfeld’s analytical framework is useful to outline the possible forms and composition of legal animal rights, Kelch rightly points out that it remains agnostic as to the normative grounds of potential animal rights. 37 In this respect, the two dominant theories of rights advance vastly differing accounts of the ultimate purpose of rights and who can potentially have them. 38 Whereas the idea of animal rights does not resonate well with the will theory, the interest theory quite readily provides a conceptual home for it.

(i) Will theory

According to the will theory, the ultimate purpose of rights is to promote and protect some aspect of an individual’s autonomy and self-realisation. A legal right is essentially a ‘legally respected choice’, and the right holder a ‘small scale sovereign’ whose exercise of choice is facilitated by giving her discretionary ‘legal powers of control’ over others’ duties. 39 The class of potential right holders thus includes only those entities that possess agency and legal competence, which effectively rules out the possibility of animals as right holders, insofar as they lack the sort or degree of agency necessary for the will-theory conception of rights. 40

However, the fact that animals are not potential right holders under the will theory does not necessarily mean that animals cannot have legal rights altogether. The will theory has attracted abundant criticism for its under-inclusiveness as regards both the class of possible right holders 41 and the types of rights it can plausibly account for, and thus seems to advance too narrow a conception of rights for it to provide a theoretical foundation for all rights. 42 In particular, it may be noted that the kinds of rights typically contemplated as animal rights are precisely of the sort that generally exceed the explanatory power of the will theory, namely inalienable, 43 passive, 44 public-law 45 rights that protect basic aspects of animals’ (partially historically and socially mediated) vulnerable corporeal existence. 46 Such rights, then, are best explained on an interest-theoretical basis.

(ii) Interest theory

Animal rights theories most commonly ground animal rights in animal interests, and thus naturally gravitate to the interest theory of rights. 47 According to the interest theory, the ultimate purpose of rights is the protection and advancement of some aspect(s) of an individual’s well-being and interests. 48 Legal rights are essentially ‘legally-protected interests’ that are of special importance and concern. 49 With its emphasis on well-being rather than on agency, the interest theory seems more open to the possibility of animal rights from the outset. Indeed, as regards the class of possible right holders, the interest theory does little conceptual filtering beyond requiring that right holders be capable of having interests. 50 Given that, depending on the underlying definition of ‘interest’, this may cover all animals, plants and, according to some, even inanimate objects, the fairly modest and potentially over-inclusive conceptual criterion of ‘having interests’ is typically complemented by the additional, more restrictive moral criterion of ‘having moral status’. 51 Pursuant to this limitation, not just any being capable of having interests can have rights, but only those whose well-being is not merely of instrumental, but of intrinsic or ‘ultimate value’. 52

Accordingly, under the interest theory, two conditions must be met for animals to qualify as potential right holders: (i) animals must have interests, (ii) the protection of which is required not merely for ulterior reasons, but for the animals’ own sake, because their well-being is intrinsically valuable. Now, whether animals are capable of having interests in the sense relevant to having rights and whether they have moral status in the sense of inherent or ultimate value is still subject to debate. For example, some have denied that animals possess interests based on an understanding of interests as wants and desires that require complex cognitive abilities such as having beliefs and language. 53 However, most interest theories opt for a broader understanding of interests in the sense of ‘being in someone’s interest’, meaning that an interest holder can be ‘made better or worse off’ and is able to benefit in some way from protective action. 54 Typically, though not invariably, the capacity for having interests in this broad sense is bound up with sentience—the capacity for conscious and subjective experiences of pain, suffering and pleasure. 55 Thus, most interest theorists quite readily accept (sentient) animals as potential right holders, that is, as the kind of beings that are capable of holding legal rights. 56

More importantly yet for legal purposes, the law already firmly rests on the recognition of (some) animals as beings who possess intrinsically valuable interests. Modern animal welfare legislation cannot be intelligibly explained other than as acknowledging that the animals it protects (i) have morally and legally relevant goods and interests, notably in their welfare, life and physical or mental integrity. 57 Moreover, it rests on an (implicit or explicit) recognition of those animals as (ii) having moral status in the sense of having intrinsic value. The underlying rationale of modern, non-anthropocentric, ethically motivated animal protection laws is the protection of animals qua animals, for their own sake, rather than for instrumental reasons. 58 Some laws go even further by directly referencing the ‘dignity’ or ‘intrinsic value’ of animals. 59

It follows that existing animal welfare laws already treat animals as intrinsically valuable holders of some legally relevant interests—and thus as precisely the sorts of beings who possess the qualities that are, under an interest theory of rights, necessary and sufficient for having rights. This, then, prompts the question whether those very laws do not only conceptually allow for potential animal rights, but might also give rise to actual legal rights for animals.

Notwithstanding that animals could have legal rights conceptually, the predominant doctrinal opinion is that, as a matter of positive law, animals do not have any, at least not in the sense of proper, legally recognised and claimable rights. 60 Yet, there is a certain inclination, especially in Anglo-American parlance, to speak—in a rather vague manner—of ‘animal rights’ as if they already exist under current animal welfare legislation. Such talk of existing animal rights is, however, rarely backed up with further substantiations of the underlying claim that animal welfare laws do in fact confer legal rights on animals. In the following, I will examine whether animals’ existing legal protections may be classified as legal rights and, if so, what kind of rights these constitute. The analysis will show (A) that implicit animal rights (hereinafter referred to as ‘animal welfare rights’) 61 can be extracted from animal welfare laws as correlatives of explicit animal welfare duties, but that this reading remains largely theoretical so far, given that such unwritten animal rights are hardly legally recognised in practice. Moreover, (B) the kind of rights derivable from animal welfare laws are currently at best imperfect and weak rights that do not provide animals with the sort of robust normative protection that is generally associated with legal rights, and typically also expected from legal animal rights qua institutionalised moral animal rights. Finally, (C) the new conceptual categories of ‘ simple animal rights’ and ‘ fundamental animal rights’ are introduced in order to distinguish, and account for the qualitative differences, between such current, imperfect, weak animal rights and potential, ideal, strong animal rights.

A. Extracting ‘Animal Welfare Rights’ from Animal Welfare Laws

(i) the simple argument from correlativity.

Existing animal welfare laws are not framed in the language of rights and do not codify any explicit animal rights. They do, however, impose on people legal duties designed to protect animals—duties that demand some behaviour that is beneficial to the welfare of animals. Some commentators contend that correlative (claim) rights are thereby conferred upon animals as the beneficiaries of such duties. 62 This view is consistent with, and, indeed, the logical conclusion of, an interest-theoretical analysis. 63 Recall that rights are essentially legally protected interests of intrinsically valuable individuals, and that a claim right is the ‘position of normative protectedness that consists in being owed a … legal duty’. 64 Under existing animal welfare laws, some goods of animals are legally protected interests in exactly this sense of ultimately valuable interests that are protected through the imposition of duties on others. However, the inference from existing animal welfare duties to the existence of correlative ‘animal welfare rights’ appears to rely on a somewhat simplistic notion of correlativity, along the lines of ‘where there is a duty there is a right’. 65 Two objections in particular may be raised against the view that beneficial duties imposed by animal welfare laws are sufficient for creating corresponding legal rights in animals.

First, not every kind of duty entails a correlative right. 66 While some duties are of an unspecific and general nature, only relational, directed duties which are owed to rather than merely regarding someone are the correlatives of (claim) rights. Closely related, not everyone who stands to benefit from the performance of another’s duty has a correlative right. According to a standard delimiting criterion, beneficial duties generate rights only in the intended beneficiaries of such duties, that is, those who are supposed to benefit from duties designed to protect their interests. 67 Yet, animal welfare duties, in a contemporary reading, are predominantly understood not as indirect duties regarding animals—duties imposed to protect, for example, an owner’s interest in her animal, public sensibilities or the moral character of humans—but as direct duties owed to the protected animals themselves. 68 Moreover, the constitutive purpose of modern animal welfare laws is to protect animals for their own sake. Animals are therefore clearly beneficiaries in a qualified sense, that is, they are not merely accidental or incidental, but the direct and intended primary beneficiaries of animal welfare duties. 69

Secondly, one may object that an analysis of animal rights as originating from intentionally beneficial duties rests on a conception of rights precisely of the sort which has the stigma of redundancy attached to it. Drawing on Hart, this would appear to cast rights as mere ‘alternative formulation of duties’ and thus ‘no more than a redundant translation of duties … into a terminology of rights’. 70 Admittedly, as MacCormick aptly puts it:

[To] rest an account of claim rights solely on the notion that they exist whenever a legal duty is imposed by a law intended to benefit assignable individuals … is to treat rights as being simply the ‘reflex’ of logically prior duties. 71

One way of responding to this redundancy problem is to reverse the logical order of rights and duties. On this account, rights are not simply created by (and thus logically posterior to) beneficial duties, but rather the converse: such duties are derived from and generated by (logically antecedent) rights. For example, according to Raz, ‘Rights are grounds of duties in others’ and thus justificationally prior to duties. 72 However, if rights are understood not just as existentially correlative, but as justificationally prior to duties, identifying intentionally beneficial animal welfare duties as the source of (logically posterior) animal rights will not suffice. In order to accommodate the view that rights are grounds of duties, the aforementioned argument from correlativity needs to be reconsidered and refined.

(ii) A qualified argument from correlativity

A refined, and reversed, argument from correlativity must show that animal rights are not merely reflexes created by animal welfare duties, but rather the grounds for such duties. In other words, positive animal welfare duties must be plausibly explained as some kind of codified reflection, or visible manifestation, of ‘invisible’ background animal rights that give rise to those duties.

This requires further clarification of the notion of a justificational priority of rights over duties. On the face of it, the idea that rights are somehow antecedent to duties appears to be at odds with the Hohfeldian correlativity axiom, which stipulates an existential nexus of mutual entailment between rights and duties—one cannot exist without the other. 73 Viewed in this light, it seems paradoxical to suggest that rights are causal for the very duties that are simultaneously constitutive of those rights—cause and effect seem to be mutually dependent. Gewirth offers a plausible explanation for this seemingly circular understanding of the relation between rights and duties. He illustrates that the ‘priority of claim rights over duties in the order of justifying purpose or final causality is not antithetical to their being correlative to each other’ by means of an analogy:

Parents are prior to their children in the order of efficient causality, yet the (past or present) existence of parents can be inferred from the existence of children, as well as conversely. Hence, the causal priority of parents to children is compatible with the two groups’ being causally as well as conceptually correlative. The case is similar with rights and duties, except that the ordering relation between them is one of final rather than efficient causality, of justifying purpose rather than bringing-into-existence. 74

Upon closer examination, this point may be specified even further. To stay with the analogy of (biological) 75 parents and their children: it is actually the content of ‘parents’—a male and a female (who at some point procreate together)—that exists prior to and independently of possibly ensuing ‘children’, whereas this content turns into ‘parents’ only in conjunction with ‘children’. That is, the concepts of ‘parents’ and ‘children’ are mutually entailing, whilst, strictly speaking, it is not ‘parents’, but rather that which will later be called ‘parents’ only once the ‘child’ comes into existence—the pre-existing content—which is antecedent to and causal for ‘children’.

Applied to the issue of rights and duties, this means that it is actually the content of a ‘right’—an interest—that exists prior to and independently of, and is (justificationally) causal for the creation of, a ‘duty’, which, in turn, is constitutive of a ‘right’. The distinction between ‘right’ and its content—an interest—allows the pinpointing of the latter as the reason for, and the former as the concomitant correlative of, a duty imposed to protect the pre-existing interest. It may thus be restated, more precisely, that it is not rights, but the protected interests which are grounds of duties. Incidentally, this specification is consistent with Raz’s definition of rights, according to which ‘having a right’ means that an aspect of the right holder’s well-being (her interest) ‘is a sufficient reason for holding some other person(s) to be under a duty’. 76 Now, the enactment of modern animal welfare laws is in and of itself evidence of the fact that some aspects of animals’ well-being (their interests) are—both temporally and justificationally—causal and a sufficient reason for imposing duties on others. Put differently: animal interests are grounds of animal welfare duties , and this, in turn, is conceptually constitutive of animal rights .

In conclusion, existing animal welfare laws could indeed be analysed as comprising unwritten ‘animal welfare rights’ as implicit correlatives of the explicit animal welfare duties imposed on others. The essential feature of legal rules conferring rights is that they specifically aim at protecting individual interests or goods—whether they do so expressis verbis or not is irrelevant. 77 Even so, in order for a right to be an actual (rather than a potential or merely postulated) legal right, it should at least be legally recognised (if not claimable and enforceable), 78 which is determined by the applicable legal rules. In the absence of unequivocal wording, whether a legal norm confers unwritten rights on animals becomes a matter of legal interpretation. While theorists can show that a rights-based approach lies within the bounds of a justifiable interpretation of the law, an actual, valid legal right hardly comes to exist by the mere fact that some theorists claim it exists. For that to happen, it seems instrumental that some public authoritative body, notably a court, recognises it as such. That is, while animals’ existing legal protections may already provide for all the ingredients constitutive of rights, it takes a court to actualise this potential , by authoritatively interpreting those legal rules as constituting rights of animals. However, because courts, with a few exceptions, have not done so thus far, it seems fair to say that unwritten animal rights are not (yet) legally recognised in practice and remain a mostly theoretical possibility for now. 79

B. The Weakness of Current ‘Animal Welfare Rights’

Besides the formal issue of legal recognition, there are substantive reasons for questioning whether the kind of rights extractable from animal welfare laws are really rights at all. This is because current ‘animal welfare rights’ are unusually weak rights that do not afford the sort of strong normative protection that is ordinarily associated with legal rights. 80 Classifying animals’ existing legal protections as ‘rights’ may thus conflict with the deeply held view that, because they protect interests of special importance, legal rights carry special normative force . 81 This quality is expressed in metaphors of rights as ‘trumps’, 82 ‘protective fences’, 83 protective shields or ‘No Trespassing’ signs, 84 or ‘suits of armor’. 85 Rights bestow upon individuals and their important interests a particularly robust kind of legal protection against conflicting individual or collective interests, by singling out ‘those interests that are not to be sacrificed to the utilitarian calculus ’ and ‘whose promotion or protection is to be given qualitative precedence over the social calculus of interests generally’. 86 Current ‘animal welfare rights’, by contrast, provide an atypically weak form of legal protection, notably for two reasons: because they protect interests of secondary importance or because they are easily overridden.

In order to illustrate this, consider the kind of rights that can be extracted from current animal welfare laws. Given that these are the correlatives of existing animal welfare duties, the substance of these rights must mirror the content laid down in the respective legal norms. This extraction method produces, first, a rather odd subgroup of ‘animal welfare rights’ that have a narrow substantive scope protecting highly specific, secondary interests, such as a (relative) right to be slaughtered with prior stunning, 87 an (absolute) right that experiments involving ‘serious injuries that may cause severe pain shall not be carried out without anaesthesia’ 88 or a right of chicks to be killed by fast-acting methods, such as homogenisation or gassing, and to not be stacked on top of each other. 89 The weak and subsidiary character of such rights becomes clearer when placed within the permissive institutional context in which they operate, and when taking into account the more basic interests that are left unprotected. 90 While these rights may protect certain secondary, derivative interests (such as the interest in being killed in a painless manner ), they are simultaneously premised on the permissibility of harming the more primary interests at stake (such as the interest in not being killed at all). Juxtaposed with the preponderance of suffering and killing that is legally allowed in the first place, phrasing the residual legal protections that animals do receive as ‘rights’ may strike us as misleading. 91

But then there is a second subgroup of ‘animal welfare rights’, extractable from general animal welfare provisions, that have a broader scope, protecting more basic, primary interests, such as a right to well-being, life, 92 dignity, 93 to not suffer unnecessarily, 94 or against torture and cruel treatment. 95 Although the object of such rights is of a more fundamental nature, the substantive guarantee of these facially fundamental rights is, to a great extent, eroded by a conspicuously low threshold for permissible infringements. 96 That is, these rights suffer from a lack of normative force, which manifests in their characteristically high infringeability (ie their low resistance to being overridden). Certainly, most rights (whether human or animal) are relative prima facie rights that allow for being balanced against conflicting interests and whose infringement constitutes a violation only when it is not justified, notably in terms of necessity and proportionality. 97 Taking rights seriously does, however, require certain safeguards ensuring that rights are only overridden by sufficiently important considerations whose weight is proportionate to the interests at stake. As pointed out by Waldron, the idea of rights is seized on as a way of resisting, or at least restricting, the sorts of trade-offs that would be acceptable in an unqualified utilitarian calculus, where ‘important individual interests may end up being traded off against considerations which are intrinsically less important’. 98 Yet, this is precisely what happens to animals’ prima facie protected interests, any of which—irrespective of how important or fundamental they are—may enter the utilitarian calculus, where they typically end up being outweighed by human interests that are comparatively less important or even trivial, notably dietary and fashion preferences, economic profitability, recreation or virtually any other conceivable human interest. 99

Any ‘animal welfare rights’ that animals may presently be said to have are thus either of the substantively oddly specific, yet rather secondary, kind or, in the case of more fundamental prima facie rights, such that are highly infringeable and ‘evaporate in the face of consequential considerations’. 100 The remaining question is whether these features render animals’ existing legal protections non-rights or just particularly unfit or weak rights , but rights nonetheless. The answer will depend on whether the quality of special strength, weight or force is considered a conceptually constitutive or merely typical but not essential feature of rights. On the first view, a certain normative force would function as a threshold criterion for determining what counts as a right and for disqualifying those legal protections that may structurally resemble rights but do not meet a minimum weight. 101 On the second view, the normative force of rights would serve as a variable that defines the particular weight of different types of rights on a spectrum from weak to strong. 102 To illustrate the intricacies of drawing a clear line between paradigmatically strong rights, weak rights or non-rights based on this criterion, let us return to the analogy with (biological) ‘parents’. In a minimal sense, the concept of ‘parents’ may be essentially defined as ‘biological creators of a child’. Typically, however, a special role as nurturer and caregiver is associated with the concept of ‘parent’. Now, is someone who merely meets the minimal conceptual criterion (by being the biological creator), but not the basic functions attached to the concept (by not giving care), still a ‘parent’? And, if so, to what extent? Are they a full and proper ‘parent’, or merely an imperfect, dysfunctional form of ‘parent’, a bad ‘parent’, but a ‘parent’ nonetheless? Maybe current animal rights are ‘rights’ in a similar sense as an absent, negligent, indifferent biological mother or father who does not assume the role and responsibilities that go along with parenthood is still a ‘parent’. That is, animals’ current legal protections may meet the minimal conceptual criteria for rights, but they do not perform the characteristic normative function of rights. They are, therefore, at best atypically weak and imperfect rights.

C. The Distinction between Simple and Fundamental Animal Rights

In the light of the aforesaid, if one adopts the view that animals’ existing legal protections constitute legal rights—that is, if one concludes that existing animal welfare laws confer legal rights on animals despite a lack of explicit legal enactment or of any coherent judicial recognition of unwritten animal rights, and that the kind of rights extractable from animal welfare law retain their rights character regardless of how weak they are—then an important qualification needs to be made regarding the nature and limits of such ‘animal welfare rights’. In particular, it must be emphasised that this type of legal animal rights falls short of (i) our ordinary understanding of legal rights as particularly robust protections of important interests and (ii) institutionalising the sort of inviolable, basic moral animal rights (along the lines of human rights) that animal rights theorists typically envisage. 103 It thus seems warranted to separate the kind of imperfect and weak legal rights that animals may be said to have as a matter of positive law from the kind of ideal, 104 proper, strong fundamental rights that animals potentially ought to have as a matter of future law.

In order to denote and account for the qualitative difference between these two types of legal animal rights, and drawing on similar distinctions as regards the rights of individuals under public and international law, 105 I propose to use the conceptual categories of fundamental animal rights and other, simple animal rights. As to the demarcating criteria, we can distinguish between simple and fundamental animal rights based on a combination of two factors: (i) substance (fundamentality or non-fundamentality of the protected interests) and (ii) normative force (degree of infringeability). Accordingly, simple animal rights can be defined as weak legal rights whose substantive content is of a non-fundamental, ancillary character and/or that lack normative force due to their high infringeability. In contradistinction, fundamental animal rights are strong legal rights along the lines of human rights that are characterised by the cumulative features of substantive fundamentality and normative robustness due to their reduced infringeability.

The ‘animal welfare rights’ derivable from current animal welfare laws are simple animal rights. However, it is worth noting that while the first subtype of substantively non-fundamental ‘animal welfare rights’ belongs to this category irrespective of their infringeability, 106 the second subtype of substantively fundamental ‘animal welfare rights’ presently falls in this category purely in respect of their characteristically high infringeability. Yet, the latter is a dynamic and changeable feature, insofar as these rights could be dealt with, in case of conflict, in a manner whereby they would prove to be more robust. In other words, while the simple animal rights of the second subtype currently lack the normative force of legal rights, they do have the potential to become fundamental animal rights. Why animals need such fundamental rights will be explored in the final section.

Beyond the imperfect, weak, simple rights that animals may be said to have based on existing animal welfare laws, a final normative question remains with a view to the future law: whether animals ought to have strong legal rights proper. I will focus on fundamental animal rights—such as the right to life, bodily integrity, liberty and freedom from torture—as these correspond best with the kind of ‘ought to be legal rights’ typically alluded to in animal rights discourse. Given the general appeal of rights language, it is not surprising that among animal advocates there is an overall presumption in favour of basic human rights-like animal rights. 107 However, it is often simply assumed that, rather than elucidated why, legal rights would benefit animals and how this would strengthen their protection. In order to undergird the normative claim that animals should have strong legal rights, the following subsections will look at functional reasons why animals need such rights. 108 I will do so through a non-exhaustive exploration of the potential legal advantages and political utility of fundamental animal rights over animals’ current legal protections (be they animal welfare laws or ‘animal welfare rights’).

A. Procedural Aspect: Standing and Enforceability

Against the backdrop of today’s well-established ‘enforcement gap’ and ‘standing dilemma’, 109 one of the most practical benefits typically associated with, or expected from, legal animal rights is the facilitation of standing for animals in their own right and, closely related, the availability of more efficient mechanisms for the judicial enforcement of animals’ legal protections. 110 This is because legal rights usually include the procedural element of having standing to sue, the right to seek redress and powers of enforcement—which would enable animals (represented by legal guardians) to institute legal proceedings in their own right and to assert injuries of their own. 111 This would also ‘decentralise’ enforcement, that is, it would not be concentrated in the hands (and at the sole discretion) of public authorities, but supplemented by private standing of animals to demand enforcement. Ultimately, such an expanded enforceability could also facilitate incremental legal change by feeding animal rights questions into courts as fora for public deliberation.

However, while standing and enforceability constitute crucial procedural components of any effective legal protection of animals, for present purposes, it should be noted that fundamental animal rights (or any legal animal rights) are—albeit maybe conducive—neither necessary nor sufficient to this end. On the one hand, not all legal rights (eg some socio-economic human rights) are necessarily enforceable. Merely conferring legal rights on animals will therefore, in itself, not guarantee sufficient legal protection from a procedural point of view. Rather, fundamental animal rights must encompass certain procedural rights, such as the right to access to justice, in order to make them effectively enforceable. On the other hand, animals or designated animal advocates could simply be granted standing auxiliary to today’s animal welfare laws, which would certainly contribute towards narrowing the enforcement gap. 112 Yet, standing as such merely offers the purely procedural benefit of being able to legally assert and effectively enforce any given legal protections that animals may have, but has no bearing on the substantive content of those enforceable protections. Given that the issue is not just one of improving the enforcement of animals’ existing legal protections, but also of substantially improving them, standing alone cannot substitute for strong substantive animal rights. Therefore, animals will ultimately need both strong substantive and enforceable rights, which may be best achieved through an interplay of fundamental rights and accompanying procedural guarantees.

B. Substantive Aspect: Stronger Legal Protection for Important Interests

The aforesaid suggests that the critical function of fundamental animal rights is not procedural in nature; rather, it is to substantively improve and fortify the protection of important animal interests. In particular, fundamental animal rights would strengthen the legal protection of animals on three levels: by establishing an abstract equality of arms, by broadening the scope of protection to include more fundamental substantive guarantees and by raising the burden of justification for infringements.

First of all, fundamental animal rights would create the structural preconditions for a level playing field where human and animal interests are both reinforced by equivalent rights, and can thus collide on equal terms. Generally speaking, not all legally recognised interests count equally when balanced against each other, and rights-empowered interests typically take precedence over or are accorded more weight than unqualified competing interests. 113 At present, the structural makeup of the balancing process governing human–animal conflicts is predisposed towards a prioritisation of human over animal interests. Whereas human interests are buttressed by strong, often fundamental rights (such as economic, religious or property rights), the interests at stake on the animal side, if legally protected at all, enter the utilitarian calculus as unqualified interests that are merely shielded by simple animal welfare laws, or simple rights that evaporate quickly in situations of conflict and do not compare to the sorts of strong rights that reinforce contrary human interests. 114 In order to achieve some form of abstract equality of arms, animals’ interests need to be shielded by strong legal rights that are a match to humans’ rights. Fundamental animal rights would correct this structural imbalance and set the stage for an equal consideration of interests that is not a priori biased in favour of humans’ rights.

Furthermore, as defined above, fundamental animal rights are characterised by both their substantive fundamentality and normative force, and would thus strengthen animals’ legal protection in two crucial respects. On a substantive level , fundamental animal rights are grounded in especially important, fundamental interests. Compared to substantively non-fundamental simple animal rights, which provide for narrow substantive guarantees that protect secondary interests, fundamental animal rights would expand the scope of protection to cover a wider array of basic and primary interests. As a result, harming fundamentally important interests of animals—while readily permissible today insofar as such interests are often not legally protected in the first place 115 —would trigger a justification requirement that initially allows those animal interests to enter into a balancing process. For even with fundamental animal rights in play, conflicts between human and animal interests will inevitably continue to exist—albeit at the elevated and abstractly equal level of conflicts of rights—and therefore require some sort of balancing mechanism. 116

On this justificatory level , fundamental animal rights would then demand a special kind and higher burden of justification for infringements. 117 As demonstrated above, substantively fundamental yet highly infringeable simple animal rights are marked by a conspicuously low threshold for justifiable infringements, and are regularly outweighed by inferior or even trivial human interests. By contrast, the normative force of fundamental animal rights rests on their ability to raise the ‘level of the minimally sufficient justification’. 118 Modelling these more stringent justification requirements on established principles of fundamental (human) rights adjudication, this would, first, limit the sorts of considerations that constitute a ‘legitimate aim’ which can be balanced against fundamental animal rights. Furthermore, the balancing process must encompass a strict proportionality analysis, comprised of the elements of suitability, necessity and proportionality stricto sensu , which would preclude the bulk of the sorts of low-level justifications that are currently sufficient. 119 This heightened threshold for justifiable infringements, in turn, translates into a decreased infringeability of fundamental animal rights and an increased immunisation of animals’ prima facie protected interests against being overridden by conflicting considerations and interests of lesser importance.

Overall, considering this three-layered strengthening of the legal protection of animals’ important interests, fundamental animal rights are likely to set robust limits to the violability and disposability of animals as means to human ends, and to insulate animals from many of the unnecessary and disproportionate inflictions of harm that are presently allowed by law.

C. Fallback Function: The Role of Rights in Non-ideal Societies

Because contemporary human–animal interactions are, for the most part, detrimental to animals, the latter appear to be in particular need of robust legal protections against humans and society. 120 Legal rights, as strong (but not impenetrable) shields, provide an instrument well suited for this task, as they operate in a way that singles out and protects important individual goods against others and the political community as a whole. For this reason, rights are generally considered an important counter-majoritarian institution, but have also been criticised for their overly individualistic, antagonistic and anti-communitarian framing. 121 Certainly, it may be debated whether there is a place for the institution of rights in an ideal society—after all, rights are not decrees of nature, but human inventions that are historically and socially contingent. 122 However, rights are often born from imperfect social conditions, as a ‘response to a failure of social responsibility’ 123 and as corrections of experiences of injustice, or, as Dershowitz puts it: ‘ rights come from wrongs ’. 124 Historical experience suggests that, at least in non-ideal societies, there is a practical need for rights as a safety net—a ‘position of fall-back and security’ 125 —that guarantees individuals a minimum degree of protection, in case or because other, less coercive social or moral mechanisms fail to do so.

Yet, as Edmundson rightly points out, this view of rights as backup guarantees does not quite capture the particular need for rights in the case of animals. 126 It is premised on the existence of a functioning overall social structure that can in some cases, and maybe in the ideal case, substitute for rights. However, unlike many humans, most animals are not embedded in a web of caring, affectionate, benevolent relations with humans to begin with, but rather are caught up in a system of exploitative, instrumental and harmful relations. For the vast majority of animals, it is not enough to say that rights would serve them as fallbacks, because there is nowhere to fall from—by default, animals are already at (or near) the bottom. Accordingly, the concrete need for rights may be more acute in the case of animals, as their function is not merely to complement, but rather to compensate for social and moral responsibility, which is lacking in the first place. 127 To give a (somewhat exaggerated) example: from the perspective of a critical legal scholar, meta-theorising from his office in the ivory tower, it may seem easier, and even desirable, to intellectually dispense with the abstract notion of rights, whereas for an elephant who is actually hunted down for his ivory tusks, concrete rights may make a very real difference, literally between life and death. Therefore, under the prevailing social conditions, animals need a set of basic rights as a primary ‘pull-up’ rather than as a subsidiary backup—that is, as compensatory baseline guarantees rather than as complementary background guarantees.

D. Transformative Function: Rights as ‘Bridges’ between Non-ideal Realities and Normative Ideals

Notwithstanding that animals need fundamental rights, we should not fail to recognise that even the minimum standards such rights are designed to establish and safeguard seem highly ambitious and hardly politically feasible at present. Even a rudimentary protection of fundamental animal rights would require far-ranging changes in our treatment of animals, and may ultimately rule out ‘virtually all existing practices of the animal-use industries’. 128 Considering how deeply the instrumental and inherently harmful use of animals is woven into the economic and cultural fabric of contemporary societies, and how pervasive animal cruelty is on both an individual and a collective level, the implications of fundamental animal rights indeed seem far removed from present social practices. 129 This chasm between normative aspirations and the deeply imperfect empirical realities they collide with is not, however, a problem unique to fundamental animal rights; rather, it is generally in the nature of fundamental rights—human or animal—to postulate normative goals that remain, to some extent, aspirational and unattainable. 130 Aspirational rights express commitments to ideals that, even if they may not be fully realisable at the time of their formal recognition, act as a continuous reminder and impulse that stimulates social and legal change towards a more expansive implementation. 131 In a similar vein, Bilchitz understands fundamental rights as moral ideals that create the pressure for legal institutionalisation and as ‘bridging concepts’ that facilitate the transition from past and present imperfect social realities towards more just societies. 132

This, then, provides a useful lens for thinking about the aspirational nature and transformative function of fundamental animal rights. Surely, the mere formal recognition of fundamental animal rights will not, by any realistic measure, bring about an instant practical achievement of the ultimate goal of ‘abolishing exploitation and liberating animals from enslavement’. 133 They do, however, create the legal infrastructure for moving from a non-ideal reality towards more ideal social conditions in which animal rights can be respected. For example, a strong animal right to life would (at least in industrialised societies) preclude most forms of killing animals for food, and would thus certainly conflict with the entrenched practice of eating meat. Yet, while the current social normality of eating animals may make an immediate prohibition of meat production and consumption unrealistic, it is also precisely the reason why animals need a right to life (ie a right not to be eaten), as fundamental rights help to denormalise (formerly) accepted social practices and to establish, internalise and habituate normative boundaries. 134 Moreover, due to their dynamic nature, fundamental rights can generate successive waves of more stringent and expansive duties over time. 135 Drawing on Bilchitz, the established concept of ‘progressive realisation’ (originally developed in the context of socio-economic human rights) may offer a helpful legal framework for the gradual practical implementation of animal rights. Accordingly, each fundamental animal right could be seen as comprising a minimum core that has to be ensured immediately, coupled with a general prohibition of retrogressive measures , and an obligation to progressively move towards a fuller realisation . 136 Therefore, even if fundamental animal rights may currently not be fully realisable, the very act of introducing them into law and committing to them as normative ideals places animals on the ‘legal map’ 137 and will provide a powerful generative basis—a starting point rather than an endpoint 138 —from which a dynamic process towards their more expansive realisation can unfold.

The question of animal rights has been of long-standing moral concern. More recently, the matter of institutionalising moral animal rights has come to the fore, and attaining legal rights for animals has become an important practical goal of animal advocates. This article started out from the prefatory observation that the process of juridification may already be in its early stages, as judicially recognised animal rights are beginning to emerge from both animal welfare law and human rights law. With legal animal rights on the horizon, the analysis set out to systematically address the arising conceptual, doctrinal and normative issues, in order to provide a theoretical underpinning for this legal development. The article showed that the idea of legal animal rights has a sound basis in both legal theory as well as in existing law. That is, legal animal rights are both conceptually possible and already derivable from current animal welfare laws. However, the analysis has also revealed that the ‘animal welfare rights’ which animals may be said to have as a matter of positive law fall short of providing the sort of strong normative protection that is typically associated with legal rights and that is furthermore expected from legal animal rights qua institutionalised moral animal rights. This discrepancy gave rise to a new conceptual distinction between two types of legal animal rights: simple and fundamental animal rights.

While the umbrella term ‘animal rights’ is often used loosely to refer to a wide range of legal protections that the law may grant to animals, distinguishing between simple and fundamental animal rights helps to unveil important differences between what we may currently call ‘legal animal rights’ based on existing animal welfare laws, which are weak legal rights at best, and the kind of strong, fundamental legal rights that animals should have as a matter of future law. This distinction is further conducive to curbing the trivialisation of the language of animal rights, as it allows us to preserve the normative force of fundamental animal rights by separating out weaker rights and classifying them as other, simple animal rights. Lastly, it is interesting to note that, with courts deriving legal animal rights from both animal welfare law and from constitutional, fundamental or human rights law, first prototypes of simple and fundamental animal rights are already discernible in emerging case law. Whereas Christopher Stone once noted that ‘each successive extension of rights to some new entity has been … a bit unthinkable’ throughout legal history, 139 the findings of this article suggest that we may presently be witnessing a new generation of legal rights in the making—legal animal rights, simple and fundamental.

This article is the first part of my postdoctoral research project ‘Trilogy on a Legal Theory of Animal Rights’, funded by the Swiss National Science Foundation. For helpful comments on earlier versions of this article, I am indebted to William Edmundson, Raffael Fasel, Chris Green, Christoph Krenn, Visa Kurki, Will Kymlicka, Nico Müller, Anne Peters, Kristen Stilt, MH Tse, Steven White, Derek Williams and the anonymous reviewers for the Oxford Journal of Legal Studies.

Seminally, Tom Regan, The Case for Animal Rights (University of California Press 1983); Sue Donaldson and Will Kymlicka, Zoopolis: A Political Theory of Animal Rights (OUP 2011).

See, notably, Matthew H Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (2001) 14 CJLJ 29; Tom L Beauchamp, ‘Rights Theory and Animal Rights’ in Tom L Beauchamp and RG Frey (eds), The Oxford Handbook of Animal Ethics (OUP 2011); William A Edmundson, ‘Do Animals Need Rights?’ (2015) 23 Journal of Political Philosophy 345; Gary L Francione, Animals, Property, and the Law (first printed 1995, Temple UP 2007) 91ff; Steven M Wise, ‘Hardly a Revolution—The Eligibility of Nonhuman Animals for Dignity-Rights in a Liberal Democracy’ (1998) 22 Vt L Rev 793; Anne Peters, ‘Liberté, Égalité, Animalité: Human-Animal Comparisons in Law’ (2016) 5 TEL 25; Thomas G Kelch, ‘The Role of the Rational and the Emotive in a Theory of Animal Rights’ (1999) 27 BC Envtl Aff L Rev 1.

Much legal scholarship deals with animal rights in a rather cursory and incidental manner, because it typically focusses on parallel debates that are closely related to, but seen as preceding, the issue of rights. For example, much has been written about the systemic shortcomings of animal welfare legislation, which—within the entrenched animal welfare/rights-dualism—has served to undergird calls for shifting towards a rights -paradigm for legal protection of animals. Another focal point of legal scholars has been to change the legal status of animals from property to person , which is taken to be a prerequisite for right holding. Yet, even though legal rights for animals may be the ultimate goal informing these debates, surprisingly little detailed attention has been given to such envisaged legal animal rights per se.

Joel Feinberg, Social Philosophy (Prentice-Hall 1973) 67.

See eg Alasdair Cochrane, Animal Rights Without Liberation: Applied Ethics and Human Obligations (Columbia UP 2012) 14–15, 207 (whose ‘account of the moral rights of animals … proposes what the legal rights of animals ought to be ’); cf Joel Feinberg, ‘In Defence of Moral Rights’ (1992) 12 OJLS 149 (describing this indirect way of referencing legal rights as the ‘“There ought to be a law” theory of moral rights’, 156).

As noted by Favre, what is required is ‘that the legal system intervene when personal morals or ethics do not adequately protect animals from human abuse’. David Favre, ‘Integrating Animal Interests into Our Legal System’ (2004) 10 Animal Law Review 87, 88.

Even though moral and legal rights are intimately connected (see HLA Hart, ‘Are There Any Natural Rights?’ (1955) 64 Philosophical Review 175, 177), a somewhat distinct (or at least modified and refined) theorisation is warranted because, unlike moral animal rights, legal animal rights are constituted by legal systems, and their existence and scope have to be determined based on the applicable legal rules. As Wise puts it: ‘philosophers argue moral rights; judges decide legal rights’. Steven M Wise, Drawing the Line: Science and the Case for Animal Rights (Perseus 2002) 34.

Supreme Court of India 7 May 2014, civil appeal no 5387 of 2014 [27] [56] [62ff]; see further Kerala High Court 6 June 2000, AIR 2000 KER 340 (expressing the opinion that ‘legal rights shall not be the exclusive preserve of the humans’, [13]); Delhi High Court 15 May 2015, CRL MC no 2051/2015 [3] [5] (recognizing birds’ ‘fundamental rights to fly in the sky’).

Tercer Juzgado de Garantías de Mendoza 3 November 2016, Expte Nro P-72.254/15; this landmark decision was preceded by an obiter dictum in Cámara Federal de Casación Penal Buenos Aires, 18 December 2014, SAIJ NV9953 [2] (expressing the view that animals are right holders and should be recognized as legal subjects).

Corte Suprema de Justicia 26 July 2017, AHC4806-2017 (MP: Luis Armando Tolosa Villabona). This ruling was later reversed in Corte Suprema de Justicia 16 August 2017, STL12651-2017 (MP: Fernando Castillo Cadena). In January 2020, the Constitutional Court of Colombia decided against granting habeas corpus to the animal in question.

Similar habeas corpus claims on behalf of chimpanzees and elephants, brought by the Nonhuman Rights Project, have not been accepted by US courts. See, notably, Tommy v Lavery NY App Div 4 December 2014, Case No 518336.

On the ambiguity of the term ‘animal rights’, see eg Will Kymlicka and Sue Donaldson, ‘Rights’ in Lori Gruen (ed), Critical Terms for Animal Studies (University of Chicago Press 2018) 320; in using the umbrella term ‘animal rights’ without further specifications, it is often left unclear what exactly is meant by ‘rights’. For example, the term may refer to either moral or legal animal rights—or both. Furthermore, in a broad sense, ‘animal rights’ sometimes refers to any kind of normative protection for animals, whereas in a narrow sense, it is often reserved for particularly important and inviolable, human rights-like animal rights. Moreover, some speak of ‘animal rights’ as if they already existed as a matter of positive law, while others use the same term in a ‘manifesto sense’, to refer to potential, ideal rights.

Joel Feinberg, ‘Human Duties and Animal Rights’ in Clare Palmer (ed), Animal Rights (Routledge 2008) 409; the class of potential right holders comprises ‘any being that is capable of holding legal rights, whether or not he/she/it actually holds such rights’. Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 29.

See generally Alon Harel, ‘Theories of Rights’ in Martin P Golding and William A Edmundson (eds), Philosophy of Law and Legal Theory (Blackwell 2005) 191ff.

Wesley Newcomb Hohfeld, ‘Some Fundamental Legal Conceptions as Applied in Judicial Reasoning’ (1913) 23 Yale LJ 16; Wesley Newcomb Hohfeld, ‘Fundamental Legal Conceptions as Applied in Judicial Reasoning’ (1917) 26 Yale LJ 710.

See Hohfeld, ‘Fundamental Legal Conceptions’ (n 15) 717; these Hohfeldian incidents of rights are merely ‘atomic’ units, whereas many common rights are complex aggregates, clusters or ‘molecular rights’ consisting of combinations thereof. ibid 746; Leif Wenar, ‘The Nature of Rights’ (2005) 33 Philosophy & Public Affairs 223, 225, 234.

First-order rights (claims and liberties) directly concern someone’s actual rather than normative conduct, whereas powers and immunities are second-order rights (‘meta-rights’) that concern other legal relations; by prioritising, for the sake of this analysis, first-order rights regarding (in)actions of and towards animals, this is not to say that second-order rights are not important to accompany and bolster the first-order rights of animals. For instance, just as many complex (eg fundamental) rights contain immunities, that is, the freedom from the legal power of another (the disability bearer) to change the immunity holder’s rights, animals’ claims and liberties may be bolstered by immunity rights that protect those first-order rights from being altered, notably voided, by others. For example, one of the most basic rights frequently discussed for animals, the ‘right not to be property’ (Gary L Francione, Introduction to Animal Rights: Your Child or the Dog? (first printed 2000, Temple UP 2007) 93ff), may be explained as an immunity that would strip away the legal powers that currently go along with the state of legal disposability entailed by animals’ property status, and would thus disable human ‘owners’ to decide over animals’ rights. As passive rights, immunities are quite easily conceivable as animal rights, because they are specified by reference to the correlative position, that is, by what the person disabled by the animal’s immunity right cannot legally do (see generally Matthew H Kramer, ‘Rights Without Trimmings’ in Matthew H Kramer, NE Simmonds and Hillel Steiner, A Debate Over Rights: Philosophical Enquiries (OUP 1998) 22). By contrast, a power refers to one’s control over a given legal relation and entails one’s normative ability to alter another’s legal position (see Hohfeld, ‘Some Fundamental Legal Conceptions’ (n 15) 55). Prima facie , powers may thus seem ill-suited for animals. This is because, unlike passive second-order rights (immunities), powers are active rights that have to be exercised rather than merely enjoyed and, unlike first-order active rights (liberties), powers concern the exercise of legal rather than factual actions and thus require legal rather than mere practical or behavioural agency. Notwithstanding, it may be argued that animals, not unlike children, could hold legal powers (eg powers of enforcement) that are exercisable through human proxies (cf Visa AJ Kurki, ‘Legal Competence and Legal Power’ in Mark McBride (ed), New Essays on the Nature of Rights (Hart Publishing 2017) 46).

For a discussion of Hohfeldian theory in the context of animal rights, see also Wise, ‘Hardly a Revolution’ (n 2) 799ff; Francione, Animals, Property, and the Law (n 2) 96–7; Kelch, ‘The Role of the Rational’ (n 2) 6ff.

Joel Feinberg, ‘The Rights of Animals and Unborn Generations’ in Joel Feinberg, Rights, Justice, and the Bounds of Liberty: Essays in Social Philosophy (Princeton UP 1980) 159; Hohfeld, ‘Some Fundamental Legal Conceptions’ (n 15) 55.

So far, animal rights theory has largely focussed on negative rights. See critically Donaldson and Kymlicka (n 1) 5ff, 49ff.

cf Wenar, ‘The Nature of Rights’ (n 16) 233.

See Hohfeld, ‘Some Fundamental Legal Conceptions’ (n 15) 55; Kramer, ‘Rights Without Trimmings’ (n 17) 10.

See eg Feinberg, ‘The Rights of Animals and Unborn Generations’ (n 19) 162; but see Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 41–2 (arguing that it would not be impossible, though ‘cruel and perhaps silly’, to impose legal duties on animals).

A ‘liberty’ is the negation of ‘duty’ and may thus be redescribed as ‘no-duty’.

On the distinction between naked and vested liberties, see HLA Hart, ‘Legal Rights’ in HLA Hart, Essays on Bentham: Studies in Jurisprudence and Political Theory (OUP 1982) 172.

Hart, ‘Legal Rights’ (n 25) 171, 173.

Hart, ‘Legal Rights’ (n 25) 171.

eg Richard L Cupp, ‘Children, Chimps, and Rights: Arguments from “Marginal” Cases’ (2013) 45 Ariz St LJ 1; see also Christine M Korsgaard, Fellow Creatures: Our Obligations to the Other Animals (OUP 2018) 116ff.

See David Lyons, ‘Rights, Claimants, and Beneficiaries’ (1969) 6 American Philosophical Quarterly 173, 173–4.

Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 42.

Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 42.

In this vein, Tommy v Lavery NY App Div 4 December 2014, Case No 518336, p 4, 6; but see critically New York Court of Appeals, Tommy v Lavery and Kiko v Presti decision of 8 May 2018, motion no 2018-268, concurring opinion Judge Fahey.

For example, the Supreme Court of Colombia explicitly departed from this reciprocity paradigm and held that animals are right holders but not duty bearers. Corte Suprema de Justicia 26 July 2017, AHC4806-2017 (MP: Luis Armando Tolosa Villabona), 14ff; for a refutation of the contractarian reciprocity argument, see also Brief for Philosophers as Amici Curiae Supporting Petitioner-Appellant, Nonhuman Rights Project v Lavery 2018 NY Slip Op 03309 (2018) (Nos 162358/15 and 150149/16), 14ff.

See Peters (n 2) 45–6; David Bilchitz, ‘Moving Beyond Arbitrariness: The Legal Personhood and Dignity of Non-Human Animals’ (2009) 25 SAJHR 38, 42–3; Feinberg, ‘The Rights of Animals and Unborn Generations’ (n 19) 163; but see Tommy v Lavery NY App Div 4 December 2014, Case No 518336, 5.

Leif Wenar, ‘The Nature of Claim Rights’ (2013) 123 Ethics 202, 207.

See Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 43.

See Kelch, ‘The Role of the Rational’ (n 2) 9.

For an overview, see generally Matthew H Kramer, NE Simmonds and Hillel Steiner, A Debate Over Rights: Philosophical Enquiries (OUP 1998).

Hart, ‘Legal Rights’ (n 25) 183, 188–9.

See Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 30; Hart, ‘Legal Rights’ (n 25) 185.

A problematic corollary of the will theory is its conceptual awkwardness, or inability, to accommodate as right holders not just non-human but also human non-agents, such as infants and the mentally incapacitated. As noted by Hart, ‘Are There Any Natural Rights?’ (n 7) 181, the will conception of rights ‘should incline us not to extend to animals and babies … the notion of a right’; see also Kramer, ‘Rights Without Trimmings’ (n 17) 69.

As pointed out by van Duffel, neither the will theory nor the interest theory may be a ‘plausible candidate for a comprehensive theory of rights’, and it may be best to assume that both theories simply attempt to capture the essence of different kinds of rights. See Siegfried van Duffel, ‘The Nature of Rights Debate Rests on a Mistake’ (2012) 93 Pacific Philosophical Quarterly 104, 105, 117 et passim .

Under the will theory, inalienable rights are not ‘rights’ by definition, as they precisely preclude the right holder’s power to waive the correlative duties. See DN MacCormick, ‘Rights in Legislation’ in PMS Hacker and J Raz (eds), Law, Morality, and Society: Essays in Honour of HLA Hart (OUP 1977) 198f; Kramer, ‘Rights Without Trimmings’ (n 17) 73.

The will theory is primarily modelled on active rights (liberties and powers) that directly facilitate individual autonomy and choice, but is less conclusive with regard to passive rights (claims and immunities) which do not involve any action or exercise of choice by the right holder herself. cf Harel (n 14) 194–5.

Hart, ‘Legal Rights’ (n 25) 190, conceded that the will theory does not provide a sufficient analysis of constitutionally guaranteed fundamental rights; legal animal rights, by contrast, are most intelligibly explained as public-law rights held primarily against the state which has correlative duties to respect and protect.

The will theory appears to limit the purpose of rights protection to a narrow aspect of human nature—the active, engaging and self-determining side—while ignoring the passive, vulnerable and needy side. Autonomy is certainly an important good deserving of normative protection, but it is hardly the only such good. See Jeremy Waldron, ‘Introduction’ in Jeremy Waldron (ed), Theories of Rights (OUP 1984) 11; MacCormick, ‘Rights in Legislation’ (n 43) 197, 208.

See Kelch, ‘The Role of the Rational’ (n 2) 10ff; for an interest-based approach to animal rights, see eg Feinberg, ‘The Rights of Animals and Unborn Generations’ (n 19); Cochrane (n 5) 19ff.

Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 29; MacCormick, ‘Rights in Legislation’ (n 43) 192.

J Raz, ‘Legal Rights’ (1984) 4 OJLS 1, 12; Waldron, ‘Introduction’ (n 46) 12, 14.

See William A Edmundson, An Introduction to Rights (2nd edn, CUP 2012) 97; Joseph Raz, The Morality of Freedom (Clarendon Press 1986) 176; Feinberg, ‘The Rights of Animals and Unborn Generations’ (n 19) 167.

See Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 33ff, 39.

Raz, The Morality of Freedom (n 50) 166, 177ff; see also Neil MacCormick, ‘Children’s Rights: A Test-Case for Theories of Right’ in Neil MacCormick, Legal Right and Social Democracy: Essays in Legal and Political Philosophy (OUP 1982) 159–60.

See RG Frey, Interests and Rights: The Case Against Animals (OUP 1980) 78ff; HJ McCloskey, ‘Rights’ (1965) 15 The Philosophical Quarterly 115, 126; but see Tom Regan, ‘McCloskey on Why Animals Cannot Have Rights’ (1976) 26 The Philosophical Quarterly 251.

Harel (n 14) 195; Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 33.

See eg Feinberg, ‘The Rights of Animals and Unborn Generations’ (n 19) 166; Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 39–40; Visa AJ Kurki, ‘Why Things Can Hold Rights: Reconceptualizing the Legal Person’ in Visa AJ Kurki and Tomasz Pietrzykowski (eds), Legal Personhood: Animals, Artificial Intelligence and the Unborn (Springer 2017) 79–80.

See eg Wenar, ‘The Nature of Claim Rights’ (n 35) 207, 227; Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 54; Feinberg, ‘The Rights of Animals and Unborn Generations’ (n 19) 166.

See also Kurki, ‘Why Things Can Hold Rights’ (n 55) 80.

See Thomas G Kelch, ‘A Short History of (Mostly) Western Animal Law: Part II’ (2013) 19 Animal Law Review 347, 348ff; Bilchitz, ‘Moving Beyond Arbitrariness’ (n 34) 44ff; in this vein, the Constitutional Court of South Africa (8 December 2016, CCT 1/16 [57]) noted that ‘the rationale behind protecting animal welfare has shifted from merely safeguarding the moral status of humans to placing intrinsic value on animals as individuals ’ (emphasis added); the well-established German concept of ‘ethischer Tierschutz’ expresses this non-anthropocentric, ethical thrust of animal welfare law. See Margot Michel, ‘Law and Animals: An Introduction to Current European Animal Protection Legislation’ in Anne Peters, Saskia Stucki and Livia Boscardin (eds), Animal Law: Reform or Revolution? (Schulthess 2015) 91–2.

1999 Federal Constitution (Bundesverfassung) (CH), Article 120(2) and 2005 Animal Welfare Act (Tierschutzgesetz) (CH), Article 1 and 3(a); 2010 Animal Welfare Act (Tierschutzgesetz) (LI), Article 1; 2018 Animal Welfare Act (Loi sur la protection des animaux) (LU), Article 1; 1977 Experiments on Animals Act (Wet op de dierproeven) (NL), Article 1a; European Parliament and Council Directive 2010/63/EU of 22 September 2010 on the protection of animals used for scientific purposes [2010] OJ L276/33, Recital 12.

See eg Steven M Wise, ‘Legal Rights for Nonhuman Animals: The Case for Chimpanzees and Bonobos’ (1996) 2 Animal Law Review 179, 179; Richard A Epstein, ‘Animals as Objects, or Subjects, of Rights’ in Cass R Sunstein and Martha C Nussbaum (eds), Animal Rights: Current Debates and New Directions (OUP 2005) 144ff; Francione, Animals, Property, and the Law (n 2) 91ff; Kelch, ‘The Role of the Rational’ (n 2) 18; Court of Appeal of Alberta, Reece v Edmonton (City) , 2011 ABCA 238 [6]; Herrmann v Germany App no 9300/07 (ECtHR, 26 June 2012), separate opinion of Judge Pinto de Albuquerque, 38; Noah v Attorney General HCJ 9232/01 [2002–2003] IsrLR 215, 225, 232, 253.

This type of current legal animal rights will be called ‘animal welfare rights’ in order to indicate their origin in current animal welfare laws.

See eg Cass R Sunstein, ‘Standing for Animals (with Notes on Animal Rights)’ (2000) 47 UCLA Law Review 1333 (claiming that current animal welfare law creates ‘a robust set of animal rights’ or even ‘an incipient bill of rights for animals’. ibid 1334, 1336); Bilchitz, ‘Moving Beyond Arbitrariness’ (n 34) 43ff, 48–9 (concluding that ‘the existing statutory framework can already be seen to confer certain legal rights upon animals’: 50 fn 61); Jerrold Tannenbaum, ‘Animals and the Law: Property, Cruelty, Rights’ (1995) 62 Social Research 539, 581; Beauchamp (n 2) 207; Wise, ‘Hardly a Revolution’ (n 2) 910ff; this view was endorsed by the Supreme Court of India 7 May 2014, civil appeal no 5387 of 2014 [27] (stating that the Prevention of Cruelty to Animals Act ‘deals with duties of persons having charge of animals, which is mandatory in nature and hence confer corresponding rights on animals’).

See eg Joel Feinberg, ‘Human Duties and Animal Rights’ in Feinberg, Rights, Justice, and the Bounds of Liberty (n 19) 193–4 et passim ; Kramer, ‘Do Animals and Dead People Have Legal Rights?’ (n 2) 54; Wenar, ‘The Nature of Claim Rights’ (n 35) 218, 220; Visa AJ Kurki, A Theory of Legal Personhood (OUP 2019) 62–5.

Matthew H Kramer, ‘Legal and Moral Obligation’ in Martin P Golding and William A Edmundson (eds), The Blackwell Guide to the Philosophy of Law and Legal Theory (Blackwell 2005) 188.

eg, for Sunstein correlativity seems to run both ways: ‘Not only do rights create duties, but the imposition of a duty also serves to create a right.’ Cass R Sunstein, ‘Rights and Their Critics’ (1995) 70 Notre Dame L Rev 727, 746.

On this objection, see also Kelch, ‘The Role of the Rational’ (n 2) 8–9.

See Lyons (n 29) 176; Waldron, ‘Introduction’ (n 46) 10; critically Kramer, ‘Rights Without Trimmings’ (n 17) 85ff; Visa AJ Kurki, ‘Rights, Harming and Wronging: A Restatement of the Interest Theory’ (2018) 38 OJLS 430, 436ff.

See eg Beauchamp (n 2) 207; Feinberg, ‘The Rights of Animals and Unborn Generations’ (n 19) 161–2, 166; Bilchitz, ‘Moving Beyond Arbitrariness’ (n 34) 45–6; in this vein, a German high court held that, based on the criminal law justification of necessity (‘rechtfertigender Notstand’), private persons may be authorised to defend the legally protected goods of animals on behalf of the animals, independently of or even against the interests of their owners. OLG Naumburg, judgment of 22 February 2018, case no 2 Rv 157/17, recital II; on why animals need directed rather than indirect duties, see Edmundson, ‘Do Animals Need Rights?’ (n 2) 350ff.

See also Francione, Animals, Property, and the Law (n 2) 100.

Hart, ‘Legal Rights’ (n 25) 181–2, 190.

MacCormick, ‘Rights in Legislation’ (n 43) 199.

Raz, The Morality of Freedom (n 50) 167, 170f; see also Alan Gewirth, ‘Introduction’ in Alan Gewirth, Human Rights: Essays on Justification and Applications (University of Chicago Press 1982) 14.

See Kramer, ‘Rights Without Trimmings’ (n 17) 40.

Gewirth (n 72) 14.

For the sake of the argument, I am only referring to biological parents.

Raz, The Morality of Freedom (n 50) 166, 180–1.

See MacCormick, ‘Rights in Legislation’ (n 43) 191–2; Raz, ‘Legal Rights’ (n 49) 13–14.

According to some scholars, legal rights exist only when they are enforceable. See eg Ronald Dworkin, Justice for Hedgehogs (Harvard UP 2011) 405–6 (stating that legal rights are only those that the right holder is entitled to enforce on demand in directly available adjudicative processes).

A significant practical hurdle to the legal recognition of animal rights is that in virtually any legal order, animals are legal objects rather than legal persons. Because legal personhood and right holding are generally thought to be inextricably linked, many jurists refrain from calling the existing legal protections of animals ‘rights’. See critically Kurki, ‘Why Things Can Hold Rights’ (n 55) 71, 85–6.

See generally Francione, Animals, Property, and the Law (n 2) 91ff.

On this, see Kai Möller, ‘Proportionality and Rights Inflation’ in Grant Huscroft, Bradley W Miller and Grégoire Webber, Proportionality and the Rule of Law: Rights, Justification, Reasoning (CUP 2014) 166; Harel (n 14) 197ff; Waldron, ‘Introduction’ (n 46) 14ff.

Ronald Dworkin, ‘Rights as Trumps’ in Waldron, Theories of Rights (n 46) 153.

Bernard E Rollin, ‘The Legal and Moral Bases of Animal Rights’ in HB Miller and WH Willliams (eds), Ethics and Animals (Humana Press 1983) 106.

Tom Regan, ‘The Day May Come: Legal Rights for Animals’ (2004) 10 Animal Law Review 11, 15–16.

Frederick Schauer, ‘A Comment on the Structure of Rights’ (1993) 27 Ga L Rev 415, 429 et passim .

Jeremy Waldron, ‘Rights in Conflict’ in Jeremy Waldron, Liberal Rights: Collected Papers 1981–1991 (CUP 1993) 209, 215–16 (emphasis added); see also Frederick Schauer, ‘Rights, Constitutions and the Perils of Panglossianism’ (2018) 38 OJLS 635, 637.

Correlative to Council Regulation (EC) 1099/2009 of 24 September 2009 on the protection of animals at the time of killing [2009] OJ L303/1, Article 4 and Annex I.

Correlative to European Parliament and Council Directive 2010/63/EU of 22 September 2010 on the protection of animals used for scientific purposes [2010] OJ L276/33, Article 14(1)(2).

Correlative to 2008 Animal Welfare Ordinance (Tierschutzverordnung) (CH), Article 178a(3).

The permissive character of animal welfare law was highlighted by the Israeli High Court of Justice in a case concerning the force-feeding of geese. Commenting on the ‘problematic’ regulatory language, it noted that the stated ‘purpose of the Regulations is “to prevent the geese’s suffering.” Clearly these regulations do not prevent suffering; at best they minimize, to some extent, the suffering caused’. Noah v Attorney General (n 60) 234–5. See also Shai Lavi, ‘Humane Killing and the Ethics of the Secular: Regulating the Death Penalty, Euthanasia, and Animal Slaughter’ (2014) 4 UC Irvine Law Review 297, 321 (noting the disparity between ‘the resolution to overcome pain and suffering, which exists side-by-side with inhumane conditions that remain unchallenged and are often taken for granted’).

As MacCormick, ‘Children’s Rights’ (n 52) 159, has succinctly put it: ‘Consider the oddity of saying that turkeys have a right to be well fed in order to be fat for the Christmas table’; this is not to minimise the importance of existing animal welfare protections. Even though they are insufficient and weak compared to proper legal rights, that does not mean that they are insignificant. See, on this point, Regina Binder, ‘Animal Welfare Regulation: Shortcomings, Requirements, Perspectives’ in Anne Peters, Saskia Stucki and Livia Boscardin (eds), Animal Law: Reform or Revolution? (Schulthess 2015) 83.

eg correlative to 1972 Animal Welfare Act (Tierschutzgesetz) (DE), § 1 and 17(1).

eg correlative to 2005 Animal Welfare Act (Tierschutzgesetz) (CH), Article 1 and 26(1)(a).

eg derived from Animal Welfare Act 2006 (UK), s 4.

See eg Supreme Court of India 7 May 2014, civil appeal no 5387 of 2014 [62] (extracting from animal welfare law, inter alia , the right to life, to food and shelter, to dignity and fair treatment, and against torture); similarly, Court of Appeal of Alberta, Reece v Edmonton (City) , 2011 ABCA 238, dissenting opinion Justice Fraser [43].

For example, the prima facie right to be free from unnecessary pain and suffering is, in effect, rendered void if virtually any kind of instrumental interest in using animals is deemed necessary and a sufficient justification for its infringement.

See Edmundson, ‘Do Animals Need Rights?’ (n 2) 346; Harel (n 14) 198; Laurence H Tribe, ‘Ten Lessons Our Constitutional Experience Can Teach Us About the Puzzle of Animal Rights: The Work of Steven M Wise’ (2001) 7 Animal Law Review 1, 2.

See Waldron, ‘Rights in Conflict’ (n 86) 209–11.

See Francione, Animals, Property, and the Law (n 2) 17ff, 109.

Francione, Animals, Property, and the Law (n 2) 114.

For Schauer, a certain normative force seems to be constitutive of the concept of rights. He argues that a right exists only insofar as an interest is protected against the sorts of low-level justifications that would otherwise be sufficient to restrict the interest if it were not protected by the right. See Schauer, ‘A Comment on the Structure of Rights’ (n 85) 430 et passim .

In this vein, Sunstein holds that animal welfare laws ‘protect a form of animal rights, and there is nothing in the notion of rights or welfare that calls for much, or little, protection of the relevant interests’. Sunstein, ‘Standing for Animals’ (n 62) 1335.

On the universal basic rights of animals, see eg Donaldson and Kymlicka (n 1) 19ff.

‘Ideal right’ in the sense of ‘what ought to be a positive … right, and would be so in a better or ideal legal system’. Feinberg, Social Philosophy (n 4) 84.

In domestic public law, fundamental or constitutional rights are distinguished from other, simple public (eg administrative) law rights. Likewise, in international law, human rights can be distinguished from other, simple or ordinary international individual rights. See Anne Peters, Beyond Human Rights: The Legal Status of the Individual in International Law (CUP 2016) 436ff.

Indeed, substantively non-fundamental simple animal rights may be quite resistant to being overridden, and may sometimes even be absolute (non-infringeable) rights.

Nonetheless, the usefulness of legal rights is not undisputed within the animal advocacy movement. For an overview of some pragmatic and principled objections against animal rights , see Kymlicka and Donaldson (n 12) 325ff.

See generally Edmundson, ‘Do Animals Need Rights?’ (n 2); Peters (n 2) 46ff.

Today, animals’ legal protections remain pervasively under-enforced by the competent public authorities as well as practically unenforceable by the affected animals or their human representatives for lack of standing. See eg Sunstein, ‘Standing for Animals’ (n 62) 1334ff; Tribe (n 97) 3.

The link between rights and the legal-operational advantage of standing was famously highlighted by Christopher D Stone, ‘Should Trees Have Standing? Toward Legal Rights for Natural Objects’ (1972) 45 S Cal L Rev 450; see further Cass R Sunstein, ‘Can Animals Sue?’ in Cass R Sunstein and Martha C Nussbaum (eds), Animal Rights: Current Debates and New Directions (OUP 2005); Peters (n 2) 47–8.

See Stone (n 110) 458ff; Tribe (n 97) 3.

See eg Constitutional Court of South Africa 8 December 2016, CCT 1/16 (affirming the National Council of Societies for the Prevention of Cruelty to Animals’ statutory power of private prosecution and to institute legal proceedings in case of animal cruelty offences).

See Frederick Schauer, ‘Proportionality and the Question of Weight’ in Grant Huscroft, Bradley W Miller and Grégoire Webber (eds), Proportionality and the Rule of Law: Rights, Justification, Reasoning (CUP 2014) 177–8.

See generally Saskia Stucki, Grundrechte für Tiere (Nomos 2016) 151ff.

For example, under the Swiss 2005 Animal Welfare Act (Tierschutzgesetz), life itself is not a legally protected good, and the (painless, non-arbitrary) killing of an animal does not therefore require any justification.

See also Noah v Attorney General (n 60) 253–4 (pointing out that balancing different interests is ‘part and parcel of our legal system’).

See generally Edmundson, ‘Do Animals Need Rights?’ (n 2) 346; Sunstein, ‘Rights and Their Critics’ (n 65) 736–7.

On this threshold-raising conception of rights, see generally Schauer, ‘A Comment on the Structure of Rights’ (n 85) 430; Ronald Dworkin, Taking Rights Seriously (Harvard UP 1978) 191–2 (noting that a right cannot justifiably be overridden ‘on the minimal grounds that would be sufficient if no such right existed’).

At present, the overwhelming portion of permissible interferences with animals’ interests can hardly be said to be necessary or proportionate in any real sense of the word. See Francione, Introduction to Animal Rights (n 17) 9, 55.

As noted by Teubner, animal rights ‘create basically defensive institutions. Paradoxically, they incorporate animals into human society in order to create defences against the destructive tendencies of human society against animals’. Gunther Teubner, ‘Rights of Non-Humans? Electronic Agents and Animals as New Actors in Politics and Law’ (2006) 33 Journal of Law and Society 497, 521.

See eg Mark Tushnet, ‘An Essay on Rights’ (1984) 62 Tex L Rev 1363; Mary Ann Glendon, Rights Talk: The Impoverishment of Political Discourse (Free Press 1991); for a modern reformulation of the rights critique, see eg Robin L West, ‘Tragic Rights: The Rights Critique in the Age of Obama’ (2011) 53 Wm & Mary L Rev 713.

See generally Alan Dershowitz, Rights from Wrongs: A Secular Theory of the Origins of Rights (Basic Books 2004) 59ff.

See Sunstein, ‘Rights and Their Critics’ (n 65) 754.

Dershowitz (n 122) 9.

Jeremy Waldron, ‘When Justice Replaces Affection: The Need for Rights’ (1988) 11 Harv JL & Pub Pol’y 625, 629.

See Edmundson, ‘Do Animals Need Rights?’ (n 2) 358.

More generally, the practical need for rights as complementary or compensatory guarantees will vary depending on social context, and may be more immediate and pressing for the disempowered, disenfranchised, marginalised, victimised, vulnerable, disadvantaged or even oppressed portions of society. See generally Patricia J Williams, ‘Alchemical Notes: Reconstructing Ideals from Deconstructed Rights’ (1987) 22 Harvard Civil Rights-Civil Liberties Law Review 401.

Donaldson and Kymlicka (n 1) 40, 49; see further Tom Regan, The Case for Animal Rights (University of California Press 2004) 330ff, 348–9; Bilchitz, ‘Moving Beyond Arbitrariness’ (n 34) 69.

See Bilchitz, ‘Moving Beyond Arbitrariness’ (n 34) 69.

On the aspirational dimension of human rights, see generally Philip Harvey, ‘Aspirational Law’ (2004) 52 Buff L Rev 701.

ibid 717–18; Raz, ‘Legal Rights’ (n 49) 14–15, 19; ‘rights are to law what conscious commitments are to the psyche’. Williams (n 127) 424.

See David Bilchitz, ‘Fundamental Rights as Bridging Concepts: Straddling the Boundary Between Ideal Justice and an Imperfect Reality’ (2018) 40 Hum Rts Q 119, 121ff.

Donaldson and Kymlicka (n 1) 49; see also Gary L Francione, Rain Without Thunder: The Ideology of the Animal Rights Movement (Temple UP 2007) 2.

cf Kymlicka and Donaldson (n 12) 331–2.

On the dynamic nature of rights and their generative power, see Raz, The Morality of Freedom (n 50) 171; Waldron, ‘Rights in Conflict’ (n 86) 212, 214.

See David Bilchitz, ‘Does Transformative Constitutionalism Require the Recognition of Animal Rights?’ (2010) 25 Southern African Public Law 267, 291ff.

Bilchitz, ‘Moving Beyond Arbitrariness’ (n 34) 71.

cf Harvey (n 130) 723 (noting that human rights will always remain a ‘work in progress rather than a finished project’); similarly, Kymlicka and Donaldson (n 12) 333.

Stone (n 110) 453.

Email alerts

Citing articles via.

• Recommend to your Library

Affiliations

• Online ISSN 1464-3820
• Print ISSN 0143-6503
• Copyright © 2024 Oxford University Press
• About Oxford Academic
• Publish journals with us
• University press partners
• What we publish
• New features  
• Open access
• Institutional account management
• Rights and permissions
• Get help with access
• Accessibility
• Advertising
• Media enquiries
• Oxford University Press
• Oxford Languages
• University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

• Copyright © 2024 Oxford University Press
• Cookie settings
• Cookie policy
• Privacy policy
• Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• J Med Ethics Hist Med

Bioethics: a look at animal testing in medicine and cosmetics in the UK

Using animals for cosmetics and medical tests has contributed towards a debate based on conflicting interests. Despite the efforts in justifying the value of animals in conducting analyses, this study seeks to elaborate whether or not it is rational to use animals as test subjects in medical and cosmetics fields. The value of animal life is at the core of the emotional conflicts that arise when animals become experimental subjects in medical and cosmetics fields. The aim of this study is to determine if there are ethical differences in the use of animal testing in medicine versus cosmetics. The research, through review and content analysis of the existing literature, compares and provides the outcomes of using animals in medical and cosmetics tests by examining studies conducted in the UK. The findings of this research indicated that animal testing is considered acceptable in the medical field only if there are no other alternatives, but is completely unacceptable in the cosmetics field. The study also provides recommendations in the form of alternatives that protect animals from cruelty and may benefit the different stakeholders and the society at large.

Introduction

Throughout history, animals have been the subject of experimentation to improve our understanding of anatomy and pathology ( 1 ). However, animal testing only became significant in the twentieth century ( 2 ).

Animal experiments are used extensively when developing new medicines and for testing the safety of certain products. Recently, the use of animals for biomedical research has been severely criticized by animal rights and protection groups. Similarly, many nations have established laws to make the practice of animal testing more humane. There are two positions in animal testing. One is that animal testing is acceptable if suffering is minimized and there are human benefits that could not have been achieved using any other means ( 3 ). The second position considers animal testing unacceptable because it causes suffering, and the benefits to human beings are either not proven or could be obtained using other methods.

As such, animal testing is a highly controversial subject that often elicits conflicting emotions from supporters and critics alike. It is also a divisive subject as some people support animal testing only in certain cases and oppose its use in other areas. For example, scientists note that significant medical breakthroughs have only been made possible through drug testing on animals. To them and other like-minded people, such achievements are reason enough to keep using animals in the lab ( 4 ). Animal tests determine if experimental drugs are effective or ineffective on human beings. Eventually, the medicine is tried out on a small group of humans through clinical trials before declaring the medicine safe to use.

Badyal and DesaI ( 5 ) note that these treatments are as beneficial to humans as they are to animals, since some human diseases are found in animals too. Therefore, some who support animal testing only advocate its use for medical (but not cosmetics) purposes, arguing that the advancement in human medicine may lead to advancement in animal medicine.

While a significant population completely disapproves of animal testing, a faction of people only disagrees with the use of animals for cosmetics testing, arguing that it is despicable and cruel to use animal life merely so that humans can advance their beauty technology. The concern extends to animals used for science, and people want animal suffering to be minimized ( 6 ). The discovery of new drugs has for a long time been based on a number of interactions among aspects such as data collected from patients, tissues, organs or cell culture and varied animal species ( 7 ). Those who oppose the use of animal testing for cosmetics believe it is outrageous and cruel to use animal life for the simple reason of making humans look better, and that the benefits to human beings do not validate the harms done to animals ( 7 ).

For such reasons, the use of animals for testing cosmetics products has been banned in the UK and all other member states of the European Union since 2013 ( 8 ). However, other countries like China and the United States of America still continue with the practice ( 9 ). Linzey adds that about 50 - 100 million animals are used for experiments every year, and that over 1.37 million animals were used for drug experimentation in America in the year 2010 ( 9 ). In the meantime, the number of experiments conducted on animals has declined in Britain but is increasing in other countries. While experiments involving vertebrates are regulated in most countries, experiments on invertebrates are not ( 5 ).

The aim of this study is to examine whether or not animal testing is still useful and necessary in the present time, and whether there are ethical differences between animal testing in medical and cosmetics fields. We use the UK as our case study and provide alternatives that can be recommended in place of animal testing.

This review was based on a cross-sectional survey by Clemence and Leaman ( 11 ) that analysed the importance of animal testing from two different aspects: medicine and cosmetics. The population consisted of individuals residing in the UK, and the sample size was 987 (= 0.03). The research included 496 men and 491 women. The report compared public views with the responses from a similar study in 2014 that had 969 participants (477 men and 492 women). The inclusion criteria were based on numerous strata such as gender, social grade definitions (i.e., professionals such as doctors and architects, people with responsible jobs such as professors, middle rank public servants such as nurses and clerics, skilled manual workers, etc.), respondents’ working status (fulltime, part-time, not working), ethnicity (white, non-white), and educational background. This report measured public perception on whether it is ethical to use animal testing for medical or cosmetics purposes. Participants were required to state whether they found it acceptable, mostly unacceptable, unacceptable, or were undecided. Consequently, the same participants were also tasked to indicate whether they saw conducting animal testing for scientific experimentation as completely necessary, somewhat necessary, not very necessary, completely unnecessary, or they did not know.

The study also utilized data from the UK Home Office ( 12 ) to determine which animals were most frequently used for medical and cosmetics research around the world. This report also provided crucial information as to the purposes of animal testing, for instance for medical research, biological testing, regulatory testing, etc.

According to the UK Home Office ( 12 ), in the year 2016, 48.6% of the animal tests in medical research were conducted for genetically oriented studies. Moreover, 28.5% of the medical research involving animal testing was for basic biological research, 13.5% was for regulatory

testing, 8.6% was for translating research from animals to humans, and 0.8% for other trainings. This is summarized in Figure 1 below.

Purposes of Animal Testing in Medicine

Data from the UK Home Office ( 10 ) indicates that the most commonly used animals for medical and cosmetics research are mice and rabbits (72.8%), fish (13.6%), rats (6.3%), birds (3.9%) and other animal species representing 3.4% of the total test animal population, as indicated in Figure 2 below.

Types of Animals Used in Testing

A published report ( 12 ) indicated that 17% of the sampled group viewed animal testing for medical research as ‘mostly unacceptable’ if there were no alternative, 17% as ‘not acceptable’, and 65% as ‘acceptable’. This was in stark contrast with testing for cosmetics purposes, to which an overwhelming 80% of the participants responded as ‘unacceptable’. The summary of the results is provided in Figure 3 and Figure 4 below.

Animal Testing for Medical Research

Animal Testing for Cosmetics Research

 In the same study ( 12 ), the participants were asked about the necessity of conducting scientific experiments on animals, which 38% of the respondents viewed as ‘completely necessary’, 23% as ‘somewhat necessary’, 20% as ‘not very necessary’, and 16% as ‘completely unnecessary’. The results are summarized in Figure 5 below.

Necessity of Conducting Scientific Experiments on Animals

The application of these methods to evaluate the safety of cosmetics was the most detested as stated by about 80% of the people who were interviewed during the investigation. The sensitivity to human life, on the other hand, reduces the strictness towards utilization of animals to find anti-viruses and antibiotics for various diseases.

The outcome portrays the essentiality of using animals to determine materials that would help the population to live healthily ( 13 ). However, in the past few decades, the number of animals used for testing drugs has been steadily decreasing ( 14 ).

The data indicates that most of the medical research processes involving animal testing emanate from genetically oriented studies, which constitute 48.6% of the medical research animal testing. Experimentation on human genetics presents various legal and ethical challenges to medical and biological researchers, alongside problems in creating experimental procedures using human test subjects. These problems occur partially due to the fact that the experimentation processes involved in these types of studies often lead to extensive gene and physiological damages to the test subjects. Such experiments typically involve deliberate presentation of diseases and other gene modifications to the test subjects, usually requiring the euthanizing of the involved subjects ( 15 ). The animal testing experimentations involving genetic processes include studies in gene modification and examine diseases believed to hold genetic components, such as cancer and diabetes ( 16 ). These experimentation processes typically involve some sort of gene modification that can simulate the presentation of genetically based disorders manifested in human beings to allow researchers to better understand those disorders.

The data also indicate that another major application of animal testing in the medical field is in basic research in biological systems and processes, which accounts for 28.5% of the testing categories. This application of animal testing in medical research involves studies in how biological systems function, and the nature and manner of disease transmission in living organisms. The findings accrued through these kinds of studies translate to advancements in the scientific knowledge of human pathology and present opportunities for the derivation and testing of cures, as noted by Festing and Wilkinson ( 17 ).

The findings further present that regulatory testing (13.5%) and animal to human translation research (8.6%) account for significant portions of the application of animal testing in the medical field. The use of animal testing for regulatory testing purposes involves applying new medical findings, procedures and products to animals to see if they meet the thresholds mandated by the medical regulatory bodies. Translation of research findings from animals to humans involves conducting research into the possibility of animal pathogens becoming infectious to humans, and identifying potential ways of applying non-human physiology to the improvement of human health. Other forms of medical and biological trainings and studies that also engage the use of animals in experimentation in the medical field include elements such as basic physiology and pathogen studies, typically conducted in educational institutions.

Animal testing in the field of cosmetics generally involves the use of animal subjects in testing new cosmetics products and ingredients. The practice essentially involves the application or forced ingestion or injection of these substances to various parts of test animals to examine their toxicity, irritation of the eyes and/or skin, ultraviolet light-triggered toxicity, and their potential for causing unwanted gene mutations ( 18 ).

The use of animal testing in the field of cosmetics research and production presents an unethical viewpoint since the findings do not advance human health, and the practice leads to the torture and killing of animals. The Humane Society ( 18 ) also notes that at the conclusion of the experimentation, the animals are usually killed through methods such as decapitation, neck twisting and asphyxiation, often without pain relief.

With regard to the ethical principles of animal testing in both fields, a convincing argument should first be presented to the Institutional Animal Care and Use Committee (IACUC). This is to justify the need for a researcher to conduct animal studies, and to ensure that the research is conducted using the smallest possible number of animals and with minimal suffering. Additionally, Naderi et al. ( 19 ) noted an increased level of legislation on the matter of animal testing, with researchers being required to submit comprehensive proposals to the IACUC to demonstrate procedural compliance with the guiding principles of the organization before conducting animal tests. Furthermore, Holden ( 20 ) highlighted the fact that researchers need to justify to review and ethics committees the use of mice rather than other alternatives in experiments. These issues indicate that researchers should look for alternatives to animal testing before proceeding with animal trials.

The issue then remains on the nature and availability of alternatives to animal testing in the medical research field. Researchers have undertaken measures to introduce some levels of such alternatives in medical studies. The accrued data indicate that a significant number of people agree with animal testing for medical research, especially when compared to those who agree with animal testing for cosmetics purposes. The data obtained from the studies indicate a slow but perceptible shift in the public opinions regarding animal testing for medical research purposes. People are increasingly finding it unacceptable to use animal test subjects even in medical research. However, the majority of the sampled people believed that medical testing procedures should use animal test subjects, but only when there is no other alternative. This indicates that people view animal testing for medical research as ethical, but under certain conditions.

The use of animals in research is still relevant because the process is useful in veterinary medicine as it helps the students understand the physiology and anatomy and improves surgical skills ( 21 ). The study by Badyal and Desai ( 5 ) supports this perception by highlighting the fact that animal use in laboratory investigation will make new discoveries possible. However, researchers should apply ethical concepts to reduce the amount of pain and unnecessary procedures for the animals. Moreover, animal testing to develop new drugs will continue to protect the future existence of humanity. Cheluvappa, et al. ( 22 ) reiterate that animal experimentation will remain essential to testing future medicine because it helps scientists understand the changes of behaviour, embryology and genetics through dissections that are conducted on the genetically produced animals.

Animals play an important role in testing human drugs as they have a large number of medical reactions similar to those of human beings. Specifically, animals such as dogs, mice and rabbits have an identical DNA that cannot be replicated through artificial models. Public concern for the increasing use of animals in terms of ethics and safety provokes anxiety among the population. Conversely, these uncertainties and unavailability of trustable alternatives show the importance of using animals in medical research as the scientists aim to protect the human race ( 23 ).

However, the use of animals to test cosmetics is highly limited due to the availability of alternative sources. For instance, The Laboratory Animals Veterinary Association (LAVA) claims that the UK government prohibits any individual from using animals to determine the suitability of cosmetics to the human body ( 13 , 24 ). In its circular, The European Union states that they have succeeded in developing alternative measures that cosmetics firms can apply to test their products without using laboratory animals ( 25 ).

Recommendations: Alternatives to Animal Testing

To improve business ethics in cosmetics companies, it is necessary for alternatives to be integrated instead of animals. Companies can employ assessment of scientific barriers to find replacements for animal test subjects and to procure the knowledge of correctly using animals for medical and cosmetics tests. Sophisticated tests on human cells or tissues, computer-modelling techniques, and experiments on people who volunteer are some measures that can limit acts of animal cruelty by cosmetics companies. Companies need to integrate tests that minimize involvement of animals in order to limit the possibility of animal cruelty, and consequently improve their business ethics. Some of the recommended alternatives are listed here.

Computer Simulation

The concept was developed by Denis Noble, and the system is currently enrolled in clinical settings. These simulations are used to test heart replacements, and are also applied to explore human behavior. Various scholars provide that this model is more accurate than animal experiments because it uses human data to analyse diseases and make predictions ( 26 ).

Stem cells are proper alternatives to the in vitro systems of disease testing and toxin evaluations ( 27 ). The experiments involve evaluation of embryonic stem cells that can be grown in Petri dishes. The Petri dishes can be placed in the cells, and after that the resulting components are placed under evaluation to help in the discovery of new medications. Stem cells are essential because they can differentiate into human tissues and make it possible to screen the suspected diseases ( 26 ).

These materials are majorly utilized in the cosmetics industry to minimize the number of animals used to test the level of toxicity in a product. Significantly, investigations showed that human tissues developed in laboratories can be used to assess the allergic responses to the available chemicals ( 28 ). These results can then be analysed by comparing reactions, and a bio signature of genes is used to make appropriate interventions.

Notably, scientists can take high-resolution pictures of human tissues, which are then analyzed with the help of various computer systems. The advantage of this model is characterized by its ability to customize the parts of the organism under consideration. Moreover, 3D images also develop prototype designs and materials that can be used to investigate the existing and future ailments ( 29 ).

This study indicates that it is justifiable to use animals in experimentations only when there are no alternatives, and the tests have significant benefits to humans. Many researchers are working towards finding options that will help eliminate the use of animals for medical and cosmetics tests. The different natures of tests conducted on animals in the fields of medicine and cosmetics tend to have clear negative implications. For such reasons, it is imperative for organizations to develop practices that endorse business ethics. Although animal tests are ideal in establishing whether drugs can be effective in treating humans for various ailments, entities that conduct these tests need to be educated about the gravity of the situation. Animals have been extremely useful in conducting genetic studies and for biological systems investigations. However, a comparison between animal tests in medicine and cosmetics reveals that their benefits in the field of medicine outweigh those in cosmetics. Therefore, animals are essential contributors to scientific experiments that are affiliated with the medical industry. The effects that medical products may have on humans make it ethical to carry out the tests on animals first.

After analysing the arguments of both the supporters and opponents involved in the controversial subject of animal testing, it is difficult to determine which direction is right or wrong. However, the agreement is that animal suffering be minimized at all costs. This research concludes that cosmetics companies should adhere to the established laws and principles against the use and abuse of animals in tests and should seek alternative methods to test their products.

Acknowledgements

Citation to this article:

Kabene S, Baadel S. Bioethics: a look at animal testing in medicine and cosmetics in the UK. J Med Ethics Hist Med. 2019; 12: 15.

Conflict of Interests

Authors declare having no conflict of interest.