research papers on machine learning and artificial intelligence

• Reference Manager
• Simple TEXT file

People also looked at

Perspective article, artificial intelligence and machine learning in sport research: an introduction for non-data scientists.

• Institute for Health and Sport, Victoria University, Melbourne, VIC, Australia

In the last two decades, artificial intelligence (AI) has transformed the way in which we consume and analyse sports. The role of AI in improving decision-making and forecasting in sports, amongst many other advantages, is rapidly expanding and gaining more attention in both the academic sector and the industry. Nonetheless, for many sports audiences, professionals and policy makers, who are not particularly au courant or experts in AI, the connexion between artificial intelligence and sports remains fuzzy. Likewise, for many, the motivations for adopting a machine learning (ML) paradigm in sports analytics are still either faint or unclear. In this perspective paper, we present a high-level, non-technical, overview of the machine learning paradigm that motivates its potential for enhancing sports (performance and business) analytics. We provide a summary of some relevant research literature on the areas in which artificial intelligence and machine learning have been applied to the sports industry and in sport research. Finally, we present some hypothetical scenarios of how AI and ML could shape the future of sports.

Introduction

It was in Moneyball ( Lewis, 2004 ), the famous success storey of the Major League Baseball team “Oakland Athletics,” that using in-game play statistics came under focus as a means to assemble an exceptional team. Despite Oakland Athletics' relatively small budget, the adoption of a rigorous data-driven approach to assemble a new team led to the playoffs in the year 2002. An economic evaluation of the Moneyball hypothesis ( Hakes and Sauer, 2006 ) describes how, at the time, a baseball hitters' salary was not truly explained by the contribution of a player's batting skills to winning games. Oakland Athletics gained a big advantage over their competitors by identifying and exploiting this information gap. It's been almost two decades since Moneyball principles, or SABRmetrics ( Lewis, 2004 ) was introduced to baseball. SABR stands for Society for American Baseball Research and SABRmetricians are those scientists who gather the in-game data and analyse it to answer questions that will lead to improving team performance. Since the success of the Oakland Athletics, most MLB teams started employing SABRmetricians. The ongoing and exponential increase of computer processing power has further accelerated the ability to analyse “big data,” and indeed, computers increasingly are taking charge of the deeper analysis of data sets, through means of artificial intelligence (AI). Likewise, the surge in high-quality data collection and data aggregation (accomplished by organisations like Baseball Savant/StatCast, ESPN and others) are key ingredients to the spike in the accuracy and breadth of analytics that was observed in the MLB in recent years.

The adoption of AI and statistical modelling in sports has become therefore more prominent in recent years as new technologies and research applications are impacting professional sports at various levels of sophistication. The wide applicability of machine learning algorithms, combined with increasing computing processing power as well as access to more and new sources of data in recent years, has made sports organisations hungry for new applications and strategies. The overriding aim is still to make them more competitive on and off the field–in athletic and business performance. The benefits of leveraging the power of AI can, in that regard, take different forms from optimising business or technical decision making to enhancing athlete/team performance but also increasing demand for attendance at sporting events, as well as promoting alternative entertainment formats of the sport.

We next list some areas where AI and machine learning (ML) have left their footprints in the world of sports ( Beal et al., 2019 ) and provide some examples of applications in each (some of the listed applications could overlap with one or more of the areas).

• Game activity/analytics: match outcome modelling, player/ball Tracking, match event (e.g., shot) classification, umpire assistance, sports betting .

• Talent identification and acquisition: player recruitment, player performance measurement, biomechanics .

• Training and coaching: assessment of team formation efficacy, tactical planning, player injury modelling .

• Fan and business focused: measurement of a player's economic value, modelling demand for event attendance, ticket pricing optimisation (variable and dynamic), wearable and sensor design, highlight packaging, virtual and augmented reality sport applications, etc .

The field of AI (particularly ML) offers new methodologies that have proven to be beneficial for tackling the above challenges. In this perspective paper we aim to provide sports business professionals and non-technical sports audiences, coaches, business leaders, policy makers and stakeholders with an overview of the range of AI approaches used to analyse sport performance and business centric problems. We also discuss perspectives on how AI could shape the future of sports in the next few years.

Research on AI and ML in Sports

In this section, we will not be reviewing examples of how AI has been applied to sports for a specific application, but rather, we will look at the intersection of AI and sports at a more abstract level, discussing some research that either surveyed or summarised the application of AI and ML in sports.

One of the earliest works discussing the potential applications of artificial intelligence in sports performance, and its positive impact on improving decision-making is by Lapham and Bartlett (1995) . The paper discusses how expert systems (i.e., a knowledge-based database used for reasoning) can be used for sports biomechanics purposes. Bartlett (2006) reviewed developments in the use of AI in sports biomechanics (e.g., throwing, shot putting, football kicking, …) to show that, at the time of writing, expert systems were marginally used in sports biomechanics despite being popular for “gait analysis” whereas Artificial Neural Networks were used for applications such as performance patterns in training and movement patterns of sports performers. An Artificial Neural Network (ANN) is a system that mimics the functionality of a human brain. ANNs are used to solve computational problems or estimate functions from a given data input, by imitating the way neurons are fired or activated in the human brain. Several (layers of) artificial neurons, known as perceptrons, are connected to perform computations which return an output as a function of the provided input ( Anderson, 1995 ).

Bartlett (2006) predicted that multi-layer ANNs will play a big role in sports technique analysis in the future. Indeed, as we discuss later, multi-layer ANNs, now commonly referred to as Deep Learning, have become one of the most popular techniques in sports related analytics. Last but not least Bartlett (2006) described the applications of Evolutionary Computation and hybrid systems in the optimization of sports techniques and skill learning. Further discussion around the applications of AI in sports biomechanics can be found in Ratiu et al. (2010) . McCabe and Trevathan (2008) discussed the use of artificial intelligence for prediction of sporting outcomes, showing how the behaviour of teams can be modelled in different sporting contests using multi-layer ANNs.

Between 2006 and 2010, machine learning algorithms, particularly ANNs were becoming more popular amongst computer scientists. This was aided by the impressive improvements in computer hardware, but also due to a shift in mindset in the AI community. Large volumes of data were made public amongst researchers and scientists (e.g., ImageNet a visual database delivered by Stanford University), and new open-source machine learning competitions were organised (such as Netflix Prize and Kaggle). It is these types of events that have shaped the adoption of AI and machine learning in many different fields of study from medicine to econometrics and sports, by facilitating access to training data and offering free open-source tools and frameworks for leveraging the power of AI. Note that, in addition to ANN, other machine learning techniques are utilised in such competitions, and sometimes these can be used in combination with one another. For instance, some of the techniques that went into the winning of the Netflix prize include singular value decomposition combined with restricted Boltzmann machines and gradient boosted decision trees.

Other examples discussing ANNs in sports include Novatchkov and Baca (2013) who discuss how ANNs can be used for understanding the quality of execution, assisting athletes and coaches, and training optimisation. However, the applications of AI to sports analytics go beyond the use of ANNs. For example, Fister et al. (2015 ) discussed how nature-inspired AI algorithms can be used to investigate unsolved research problems regarding safe and effective training plans. Their approach ( Fister et al., 2015 ) relies on the notion of artificial collective intelligence ( Chmait et al., 2016 ; Chmait, 2017 ) and the adaptability of algorithms to adapt to a changing environment. The authors show how such algorithms can be used to develop an artificial trainer to recommend athletes with an informed training strategy after taking into consideration various factors related to the athlete's physique and readiness. Other types of scientific methods that include Bayesian approaches have been applied to determining player abilities ( Whitaker et al., 2021 ) but also predicting match outcomes ( Yang and Swartz, 2004 ). Bayesian analysis and learning is an approach for building (statistical and inference) models by updating the probability for a hypothesis as more evidence or information becomes available by using Bayes' theorem ( Ghosh et al., 2007 ).

There are numerous research papers in which AI and ML is applied to sport, and it is not our aim to comprehensively discuss these works here 1 . However, we refer to a recent survey that elaborates on this topic. Beal et al. (2019) surveyed the applications of AI in team sports. The authors summarised existing academic work, in a range of sports, tackling issues such as match outcome modelling, in-game tactical decision making, player performance in fantasy sport games, and managing professional players' sport injuries. Work by Nadikattu (2020) presents, at an abstract level, discussions on how AI can be implemented in (American) sports from enhancing player performance, to assisting coaches to come up with the right formations and tactics, to developing automated video highlights of sports matches and supporting referees using computer vision applications.

We emphasise that the application of AI in sports is not limited to topics of sports performance, athlete talent identification or the technical analysis of the game. The (off the field) business side of sports organisations is rapidly shifting towards a data driven culture led by developing profiles of their fans and their consumer preferences. As fans call for superior content and entertainment, sport organisations must react by delivering a customised experience to their patrons. This is often achieved by the use of statistical modelling as well as other machine learning solutions, for example, to understand the value of players from an economic perspective. As shown in Chmait et al. (2020a) , investigating the relationship between the talent and success of athletes (to determine the existence of what is referred to as superstardom phenomenon or star power) is becoming an important angle to explore value created in sport. To provide an idea of the extent of such work, we note some sports in which the relationship between famous players/teams and their effect on audience attendance or sport consumption has been studied:

• In soccer ( Brandes et al., 2008 ; Jewell, 2017 ),

• In Major League Baseball ( Ormiston, 2014 ; Lewis and Yoon, 2016 )

• In the National Basketball Association ( Berri et al., 2004 ; Jane, 2016 )

• In tennis: superstar player effect in demand for tennis tournament attendance ( Chmait et al., 2020a ), the presence of a stardom effect in social media ( Chmait et al., 2020b ), player effect on German television audience demand for live broadcast tennis matches ( Konjer et al., 2017 )

• And similarly, in Cricket ( Paton and Cooke, 2005 ), Hockey ( Coates and Humphreys, 2012 ), and in the Australian Football League ( Lenten, 2012 ).

AI algorithms are being used in Formula 1 (F1) to improve the racing tactics of competing teams by analysing data from hundreds of sensors in the F1 car. Recent work by Piccinotti (2021) shows how artificial intelligence can provide F1 with automated ways for identifying tyre replacement strategies by modelling pit-stop timing and frequency as sequential decision-making problems.

Researchers from Tennis Australia and Victoria University devised a racket recommendation technique based on real HawkEye (computer vision system) data. An algorithm was used to recommend a selection of rackets based on movement, hitting pattern and style of the player with the aim to improve the player's performance ( Krause, 2019 ).

Accurate and fair judging of sophisticated skills in sports like gymnastics is a difficult task. Recently, a judging system was developed by Fujitsu Ltd. The system scores a routine based on the angles of a gymnast's joints. It uses AI to analyse 3D laser sensors that capture the gymnasts' movements ( Atiković et al., 2020 ).

Finally, it is important to note the exceptionally successful adoption of AI in board games like Chess, Checkers, Shogi and the Chinese game of GO, as well as virtual games (like Dota2 and StarCraft). In the last couple of decades, AI has delivered a staggering rise in performance in such areas to the point that machines (almost) constantly defeat human world champions. We refer to some notable solutions like Schaeffer et al. (2007) Checkers artificial algorithm, DeepBlue defeating Kasparov in Chess ( Campbell et al., 2002 ), AlphaGo Zero defeating Lee Sedol in Go ( Silver et al., 2017 ) (noting that AlphaZero is also unbeatable in chess) and Vinyals et al. (2019) AlphaStar in StarcraftII as well as superhuman AI for multiplayer poker ( Brown and Sandholm, 2019 ). Commonly, in these types of games or sports, AI algorithms rely on a Reinforcement Learning approach (which we will describe later) as well as using techniques like the Monte-Carlo Search Trees to explore the game and devise robust strategies to solve and play these games. Some of the recent testbeds used to evaluate AI agents and algorithms are discussed in Hernández-Orallo et al. (2017 ). For a broader investigation of AI in board and virtual/computer games refer to Risi and Preuss (2020) .

The rise of applying AI and ML is unstoppable and to that end, one might be wondering how AI an ML tools work and why are they different from traditional summary analytics. We touch upon these considerations in the next section.

The Machine Learning Paradigm

To understand why ML is used in a wide range of applications, we need to take a look into the difference between recent AI approaches to learning and traditional analytics approaches. At a higher conceptual level, one can describe old or traditional approaches to sports analytics, as starting off with some set of rules that constitute the problem definition, some data that is to be processed using a program/application which will then deliver answers to the given problem. In contrast, in a machine learning/predictive analytics paradigm, the way this process works is fundamentally different. For instance, in some approaches of the ML paradigm, one typically starts by feeding the program with answers and corresponding data to a specific problem, with an algorithm narrowing down the rules of the problem. These rules are later used for making predictions and they are evaluated or validated by testing their accuracy over new (unseen) data.

To that end, machine learning is an area of AI that is concerned with algorithms that learn from data by performing some form of inductive learning. In simple terms, ML prediction could be described as a function 2 from a set of inputs i 1 , i 2 , …, i n , to forecast an unknown value y , as follows f ( w 1 * i 1 , w 2 * i 2 , …, w n * i n ) = y , where w t is the weight of input t .

Different types or approaches of ML are used for different types of problems. Some of the most popular are supervised learning, unsupervised learning , and reinforcement learning :

• In supervised learning, we begin by observing and recording both inputs (the i 's) and outputs (the y 's) of a system, for a given period of time. This data (collection of correct examples of inputs and their corresponding outputs) is then analysed to derive the rules that underly the dynamics of the observed system, i.e., the rules that map a given input to its correct output.

• Unlike the above, in unsupervised learning, the correct examples or outputs from a given system are not available. The task of the algorithm is to discover (previously unnoticed) patterns in the input data.

• In reinforcement learning, an algorithm (usually referred to as an agent) is designed to take a series of actions that maximise its cumulative payoff or rewards over time. The agent then builds a policy (a map of action selection rules) that return a probability of taking a given action under different conditions of the problem.

For a thorough introduction to the fundamentals of machine learning and the popular ML algorithms see Bonaccorso (2017) . The majority of AI applications in sports are based on one or more of the above approaches to ML. In fact, in most predictive modelling applications, the nature of the output y that needs to be predicted or analysed could influence the architecture of the learning algorithm.

Explaining the details of how different ML techniques work is outside the scope of this paper. However, to provide an insight into how such algorithms function in layman's terms and the differences between them, we briefly present (hypothetical) supervised, unsupervised and reinforcement learning problems in the context of sports. These examples will assist the professionals but also applied researchers who work in sport to better understand the way that data scientists think so to facilitate talking to them about their approach and methodology, without requiring to dive deep into the details of the underlying analytics.

Supervised Learning: Predicting Player Injury

Many sports injuries (e.g., muscle strain) can be effectively treated or prevented if one is able to detect them early or predict the likelihood of sustaining them. There could be many different (combinations of) reasons/actions leading to injuries like muscle strain. For example, in the Australian Football League, some of hypotheses put forward leading to muscle strain include: muscle weakness and lack of flexibility, fatigue, inadequate warm-up, and poor lumbar posture ( Brockett et al., 2004 ). Detecting the patterns that can lead to such injuries is extremely important both for the safety of the players, and for the success and competitiveness of the team.

In a supervised learning scenario, data about the players would be collected from previous seasons including details such as the number of overall matches and consecutive matches they played, total time played in each match, categorised by age, number of metres run, whether or not they warmed up before the match, how many times they were tackled by other players, and so on , but more importantly, whether or not the players ended up injured and missed their next match.

The last point is very important as it is the principal difference between supervised learning and other approaches: the outcome (whether or not the player was injured) is known in the historical data that was collected from previous seasons. This historical data is then fed (with the outcome) to a machine learning algorithm with the objective of learning the patterns (combination of factors) which led to an injury (and usually assigning a probability of the likelihood of an injury given these patterns). Once these patterns are learnt, the algorithm or model is then tested on new (unseen data) to see if it performs well and indeed predicts/explains injury at a high level of accuracy (e.g., 70% of the time). If the accuracy of the model is not as required, the model is tuned (or trained with slightly different parameters) until it reaches the desired or acceptable accuracy. Note here that we did not single out a specific algorithm or technique to achieve the above. Indeed, this approach can be applied using many different ML algorithms such as Neural Networks, Decision Trees and regression models.

Unsupervised Learning: Fan Segmentation

We will use a sport business example to introduce the unsupervised learning approach. Most sports organisations keep track of historical data about their patrons who attended their sporting events, recording characteristics such as their gender, postcode, age, nationality, education, income, marital status, etc. A natural question of interest here is to understand if the different segments of customers/patrons will purchase different categories (e.g., price, duration, class etc.) of tickets.

Some AI algorithms are designed to help split the available data, so that each data point (historical ticket sale) sits in a group/class that is similar to the other data points (other sales) in that same class given the recorded features. The algorithm will then use some sort of a similarity or distance metric to classify the patrons according to the category of tickets that they might purchase.

This is different from how supervised learning algorithms, like those discussed in the previous section, work. As we described before, in supervised learning we instruct the algorithm with the outcome in advance while training it (i.e., we classify/label each observation based on the outcome: injury or no injury, cheap or expensive seats, …). In the unsupervised learning approach, there is no such labelling or classification of existing historical data. It is the mission of the unsupervised learning algorithm to discover (previously unnoticed) patterns in the input data and group it into (two or more) classes.

Imagine the following use case where an Australian Football League club aims to identify a highly profitable customer segment within its entire set of stadium attendees, with the aim to enhance its marketing operations. Mathematical models can be used to discover (segments of) similar customers based on variations in some customer attributes within and across each segment. A popular unsupervised learning algorithm to achieve such goal is the K-means clustering algorithm which finds the class labels from the data. This is done by iteratively assigning the data points (e.g., customers) from the input into a group/class based on the characteristics of this input. The essence is that the groups or classes to which the data points are assigned to are not defined prior to exploring the input data (although the number of groups or segments can be pre-defined) but are rather dynamically formed as the K-means algorithm iterates over the data points. In the context of customer segmentation, when presenting the mathematical model (K-means algorithm) with customer data, there is no requirement to label a portion (or any of) of this data into groups in advance in order to train the model as usually done in supervised models.

Reinforcement Learning: Simulations and Fantasy Sports

As mentioned before, in reinforcement learning, an algorithm (such as Q-learning and SARSA algorithms) learns how to complete a series of tasks (i.e., solve a problem) by interacting with an (artificial) environment that was designed to simulate the real environment/problem at hand. Unlike the case with supervised learning, the algorithm is not explicitly instructed about the right/accurate action in different states/conditions of the environment (or steps of problem it is trying to solve). But rather it incrementally learns such a protocol through reward maximisation.

In simple terms, reinforcement learning approaches represent problems using what are referred to as: an agent (a software algorithm), and a table of states and actions . When the agent executes an action, it transitions from one state to another and it receives a reward or a penalty (a positive or negative numerical score respectively) as a result. The reward/penalty associated with the action-state combination is then stored in the agent's table for future reference and refinement. The agent's goal is to take the action that maximises its reward. When the agent is still unaware of the expected rewards from executing a given action when at a given state, it takes a random action and updates its table following that action. After many (thousands of) iterations over the problem space, the agent's table holds (a weighted sum of) the expected values of the rewards of all future actions starting from the initial state.

Reinforcement learning has been applied to improve the selection of team formations in fantasy sports ( Matthews et al., 2012 ). Likewise, the use of reinforcement learning is prominent in online AI bots and simulators like chess, checkers, Go, poker, StarCraft, etc.

Finally, it is important to also note the existence of genetic or evolutionary algorithms, sometimes referred to as nature/bio-inspired algorithms. While such algorithms are not typically considered to be ML algorithms (but rather search techniques and heuristics), they are very popular in solving similar types of problems tackled by ML algorithms. In short, the idea behind such algorithms is to run (parallel) search, selection and mutation techniques, by going over possible candidate solutions of a problem. The solutions are gradually optimised until reaching a local (sub-optimal) or global maximum (optimal solution). To provide a high-level understanding of evolutionary algorithms, consider the following sequence of steps:

• We start by creating (a population of) initial candidate or random strategies/solutions to the problem at hand.

• We assess these candidate solutions (using a fitness function) and assign scores to each according to how well they solve the problem at hand.

• We then pick a selection of these candidate solutions that performed best at stage two above. We then combine ( crossbreed ) these together to generate ( breed) new solutions (e.g., take some attributes from one candidate solution and others from another candidate solution in order to come up with a new solution).

• We then apply random changes ( mutations ) to the resulting solutions from the previous step.

• We repeat the solution combination/crossbreeding process until a satisfactory solution is reached.

Evolutionary algorithms can be used as alternative means for training machine learning algorithms such as reinforcement learning algorithms and deep neural networks.

The Future of AI in Sport

There is no doubt that AI will continue to transform sports, and the ways in which we play, watch and analyse sports will be innovative and unexpected. In fact, machine learning has drastically changed the way we think about match strategies, player performance analytics but also how we track, identify and learn about sport consumers. A Pandora's box of ethical issues is emerging and will increasingly need to be considered when machines invade the traditionally human centred and naturally talented athlete base of sport. It is unlikely that AI will completely replace coaches and human experts, but there is no doubt that leveraging the power of AI will provide coaches and players with a big advantage and lead over those who only rely on human expertise. It will also provide sport business managers with deeper, real time insights into the behaviours, needs and wants of sport consumers and in turn AI will become a main producer of sport content that is personalised and custom made for individual consumers. But human direction and intervention seems to be, at least in the near future, still essential working towards elite sport performance and strategic decision making in sport business. The sporting performance on the field is often produced as an entertainment spectacle, where the sporting context is the platform for generating the business of sport. Replacing referees with automated AI is clearly possible and increasingly adopted in various sports, because it is more accurate and efficient, but is it what the fans want?

What might the future of sport with increasingly integrated AI look like? Currently, most of the research in AI and sports is specialised. That is to provide performance or business solutions and solve specific on and off field problems. For instance, scientists have successfully devised solutions to tackle problems like player performance measurement, and quantifying the effect of a player/team on demand for gate attendance. Nevertheless, our research has not identified studies (yet) that provide a 360-degree analysis on, for example, the absolute value of an athlete by taking into account all the dimensions of his or her performance on how much business can be developed, for example in regard to ticket sales or endorsement deals.

One of the main challenges to achieve such a comprehensive analysis is mainly due to the fact that data about players and teams, and commercial data such as ticket sales and attendance numbers, are kept proprietary and are not made public to avoid providing other parties with competitive information. Moreover, privacy is an important consideration as well. Regulations about data privacy and leakage of personal identification details must be put in place to govern the use and sharing of sports (performance and consumption) data. Data ownership, protection, security, privacy and access will all drive the need for comprehensive and tight legislation and regulation that will strongly influence the speed and comprehensiveness of the adoption of AI in sport. To that end, it is worth considering privacy and confidentiality implications independently when studying the leagues' journey of AI adoption compared to that of individual teams and ultimately the individual players. Eventually, the successful adoption of AI in a sports league will likely depend on the teams in that league and their players to be willing to share proprietary data or insights with other teams in the league. Performance data of players in particular is becoming a hot topic of disputation. It may well be AI that will determine the bargaining power of players and their agents in regard to the value of their contracts. As an extension of this it will then also be AI providing the information that will determine if players are achieving the performance objectives set by coaches and as agreed to in contracts. In other words, confidentiality and ownership of league, team or player level data will become an increasing bone of legal contention and this will be reflected in the complexity of contractual agreements and possible disputes in the change rooms and on the field of play. Being in control of which data can or cannot, and will or will not, be used is at stake.

From an economic perspective, relying on artificial algorithms could increase the revenue of sports organisations and event organisers when enabled to apply efficient variable and dynamic pricing strategies and build comprehensive and deep knowledge consumer platforms. Different types of ML algorithms can be adopted to deliver more effective customer marketing via personalisation and to increase sales funnel conversion rates.

Finally, for a window on the future of data privacy, it might be useful to return to baseball where the addiction to big data started its spread across the high-performance sport industry. Hattery (2017 , p. 282) explains that in baseball “using advanced data collection systems … the MLB teams compete to create the most precise injury prediction models possible in order to protect and optimise the use of their player-assets. While this technology has the potential to offer tremendous value to both team and player, it comes with a potential conflict of interest. Players' goals are not always congruent with those of the organisation: the player strives to protect his own career while the team is attempting to capitalise on the value of an asset. For this reason, the player has an interest in accessing data that analyses his potential injury risk. This highlights a greater problem in big data: what rights will individuals possess regarding their own data points?”

This privacy issue can be further extended to the sport business space Dezfouli et al. (2020) have shown how AI can be designed to manipulate human behaviour. Algorithms learned from humans' responses who were participating in controlled experiments. The algorithms identified and targeted vulnerabilities in human decision-making. The AI succeeded in steering participants towards executing particular actions. So, will AI one day be shaping the spending behaviour of sports fans by exploiting their fan infused emotional vulnerabilities and monitoring their (for example) gambling inclinations? Will AI sacrifice the health of some athletes in favour of the bigger team winning the premiership? Or is this already happening? Time will tell.

Data Availability Statement

The original contributions presented in the study are included in the article, further inquiries can be directed to the corresponding author.

Author Contributions

NC and HW had major contribution to the writing of this manuscript. NC contributed to the writing of the parts around artificial intelligence and machine learning and provided examples of these. HW shaped the scope of the manuscript and wrote and edited many of its sections particularly the introduction and the discussion. Both authors contributed to the article and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

1. ^ For conferences and published articles on AI and sports analytics see Swartz (2020) .

2. ^ Note that such function is also found in regression techniques where the weights/coefficients are unknown. In ML, it is usually the case where both the function and its weights are unknown and are determined using various search techniques and algorithms.

Anderson, J. A. (1995). An introduction to Neural Networks . Cambridge, MA: MIT Press.

Google Scholar

Atiković, A., Kamenjašević, E., Nožinović, M. A., Užičanin, E., Tabaković, M., and Curić, M. (2020). Differences between all-around results in women's artistic gymnastics and ways of minimizing them. Balt. J. Health Phys. Act. 12, 80–91. doi: 10.29359/BJHPA.12.3.08

CrossRef Full Text | Google Scholar

Bartlett, R. (2006). Artificial intelligence in sports biomechanics: new dawn or false hope? J. Sports Sci. Med. 5, 474–479.

PubMed Abstract | Google Scholar

Beal, R., Norman, T. J., and Ramchurn, S. D. (2019). Artificial intelligence for team sports: a survey. Knowl. Eng. Rev. 34. doi: 10.1017/S0269888919000225

PubMed Abstract | CrossRef Full Text | Google Scholar

Berri, D. J., Schmidt, M. B., and Brook, S. L. (2004). Stars at the gate: the impact of star power on NBA gate revenues. J. Sports Econom. 5, 33–50. doi: 10.1177/1527002503254051

Bonaccorso, G. (2017). Machine Learning Algorithms . Birmingham: Packt Publishing Ltd.

Brandes, L., Franck, E., and Nuesch, S. (2008). Local heroes and superstars: an empirical analysis of star attraction in German soccer. J. Sports Econom. 9, 266–286. doi: 10.1177/1527002507302026

Brockett, C. L., Morgan, D. L., and Proske, U. W. E. (2004). Predicting hamstring strain injury in elite athletes. Med. Sci. Sports Exerc. 36, 379–387. doi: 10.1249/01.MSS.0000117165.75832.05

Brown, N., and Sandholm, T. (2019). Superhuman AI for multiplayer poker. Science 365, 885–890. doi: 10.1126/science.aay2400

Campbell, M., Hoane Jr, A. J., and Hsu, F. H. (2002). Deep blue. Artif. Intell. 134, 57–83. doi: 10.1016/S0004-3702(01)00129-1

Chmait, N. (2017). Understanding and measuring collective intelligence across different cognitive systems: an information-theoretic approach. in IJCAI (Melbourne), 5171–5172.

Chmait, N., Dowe, D. L., Li, Y. F., Green, D. G., and Insa-Cabrera, J. (2016). Factors of collective intelligence: how smart are agent collectives? in Proceedings of the Twenty-second European Conference on Artificial Intelligence (Prague), 542–550.

Chmait, N., Robertson, S., Westerbeek, H., Eime, R., Sellitto, C., and Reid, M. (2020a). Tennis superstars: the relationship between star status and demand for tickets. Sport Manag. Rev . 23, 330–347. doi: 10.1016/j.smr.2019.03.006

Chmait, N., Westerbeek, H., Eime, R., Robertson, S., Sellitto, C., and Reid, M. (2020b). Tennis influencers: the player effect on social media engagement and demand for tournament attendance. Telemat Inform. 50:101381. doi: 10.1016/j.tele.2020.101381

Coates, D., and Humphreys, B. R. (2012). Game attendance and outcome uncertainty in the National Hockey League. J. Sports Econom. 13, 364–377. doi: 10.1177/1527002512450260

Dezfouli, A., Nock, R., and Dayan, P. (2020). Adversarial vulnerabilities of human decision-making. Proc. Nat. Acad. Sci. U.S.A. 117, 29221–29228. doi: 10.1073/pnas.2016921117

Fister Jr, I., Ljubič, K., Suganthan, P. N., Perc, M., and Fister, I. (2015). Computational intelligence in sports: challenges and opportunities within a new research domain. Appl. Math. Comput. 262, 178–186. doi: 10.1016/j.amc.2015.04.004

Ghosh, J. K., Delampady, M., and Samanta, T. (2007). An Introduction to Bayesian Analysis: Theory and Methods . Berlin: Springer Science and Business Media.

Hakes, J. K., and Sauer, R. D. (2006). An economic evaluation of the Moneyball hypothesis. J. Econ. Perspect. 20, 173–186. doi: 10.1257/jep.20.3.173

Hattery, M. (2017). Major League Baseball players, big data, and the right to know: the duty of Major League Baseball teams to disclose health modelling analysis to their players. Marquette Sports Law Rev. 28, 257–283. https://scholarship.law.marquette.edu/sportslaw/vol28/iss1/9/

Hernández-Orallo, J., Baroni, M., Bieger, J., Chmait, N., Dowe, D. L., Hofmann, K., et al. (2017). A new AI evaluation cosmos: ready to play the game? AI Mag. 38, 66–69. doi: 10.1609/aimag.v38i3.2748

Jane, W.-J. (2016). The effect of star quality on attendance demand: the case of the National Basketball Association. J. Sports Econom. 17, 396–417. doi: 10.1177/1527002514530405

Jewell, R. T. (2017). The effect of marquee players on sports demand: the case of US Major League Soccer. J. Sports Econom. 18, 239–252. doi: 10.1177/1527002514567922

Konjer, M., Meier, H. E., and Wedeking, K. (2017). Consumer demand for telecasts of tennis matches in Germany. J. Sports Econom. 18, 351–375. doi: 10.1177/1527002515577882

Krause, L. (2019). Exploring the influence of practice design on the development of tennis players (Doctoral dissertation). Victoria University, Footscray, VIC, Australia.

Lapham, A. C., and Bartlett, R. M. (1995). The use of artificial intelligence in the analysis of sports performance: a review of applications in human gait analysis and future directions for sports biomechanics. J. Sports Sci. 13, 229–237. doi: 10.1080/02640419508732232

Lenten, L. J. (2012). Comparing attendances and memberships in the Australian Football League: the case of hawthorn. Econ Labour Relat. Rev. 23, 23–38. doi: 10.1177/103530461202300203

Lewis, M. (2004). Moneyball: The Art of Winning an Unfair Game . New York, NY: WW Norton and Company.

Lewis, M., and Yoon, Y. (2016). An empirical examination of the development and impact of star power in Major League Baseball. J. Sports Econom. 19, 155–187. doi: 10.1177/1527002515626220

Matthews, T., Ramchurn, S., and Chalkiadakis, G. (2012). Competing with humans at fantasy football: Team formation in large partially-observable domains. in Proceedings of the AAAI Conference on Artificial Intelligence , Vol. 26 (Vancouver, BC), 1394–1400.

McCabe, A., and Trevathan, J. (2008). Artificial intelligence in sports prediction. in Fifth International Conference on Information Technology: New Generations (IEEE: Las Vegas, NV), 1194–1197. doi: 10.1109/ITNG.2008.203

Nadikattu, R. R. (2020). Implementation of new ways of artificial intelligence in sports. J. Xidian Univ. 14, 5983–5997. doi: 10.2139/ssrn.3620017

Novatchkov, H., and Baca, A. (2013). Artificial intelligence in sports on the example of weight training. J. Sports Sci. Med. 12, 27–37.

Ormiston, R. (2014). Attendance effects of star pitchers in major league baseball. J. Sports Econom. 15, 338–364. doi: 10.1177/1527002512461155

Paton, D., and Cooke, A. (2005). Attendance at county cricket: an economic analysis. J. Sports Econom. 6, 24–45. doi: 10.1177/1527002503261487

Piccinotti, D. (2021). Open Loop Planning for Formula 1 Race Strategy Identification. Menlo Park, CA: Association for the Advancement of Artificial Intelligence.

Ratiu, O. G., Badau, D., Carstea, C. G., Badau, A., and Paraschiv, F. (2010). Artificial intelligence (AI) in sports, in Proceedings of the 9th WSEAS International Conference on Artificial Intelligence, Knowledge Engineering, and Data Bases (Cambridge, UK), 93–97.

Risi, S., and Preuss, M. (2020). From chess and Atari to StarCraft and beyond: how game AI is driving the world of AI. KI-Künstliche Intell. 34, 7–17. doi: 10.1007/s13218-020-00647-w

Schaeffer, J., Burch, N., Björnsson, Y., Kishimoto, A., Müller, M., Lake, R., et al. (2007). Checkers is solved. science 317, 1518–1522. doi: 10.1126/science.1144079

Silver, D., Schrittwieser, J., Simonyan, K., Antonoglou, I., Huang, A., Guez, A., et al. (2017). Mastering the game of go without human knowledge. Nature 550, 354–359. doi: 10.1038/nature24270

Swartz, T. B. (2020). Where should I publish my sports paper? Am. Stat. 74, 103–108. doi: 10.1080/00031305.2018.1459842

Vinyals, O., Babuschkin, I., Czarnecki, W. M., Mathieu, M., Dudzik, A., Chung, J., et al. (2019). Grandmaster level in StarCraft II using multi-agent reinforcement learning. Nature 575, 350–354. doi: 10.1038/s41586-019-1724-z

Whitaker, G. A., Silva, R., Edwards, D., and Kosmidis, I. (2021). A Bayesian approach for determining player abilities in football. J. R. Stat. Soc. Series C 70, 174–201. doi: 10.1111/rssc.12454

Yang, T. Y., and Swartz, T. (2004). A two-stage Bayesian model for predicting winners in major league baseball. J. Data Sci. 2, 61–73. doi: 10.6339/JDS.2004.02(1).142

Keywords: artificial intelligence, machine learning, sports business, sports analytics, sport research, future of sports

Citation: Chmait N and Westerbeek H (2021) Artificial Intelligence and Machine Learning in Sport Research: An Introduction for Non-data Scientists. Front. Sports Act. Living 3:682287. doi: 10.3389/fspor.2021.682287

Received: 18 March 2021; Accepted: 15 November 2021; Published: 08 December 2021.

Reviewed by:

Copyright © 2021 Chmait and Westerbeek. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Nader Chmait, nader.chmait@vu.edu.au

This article is part of the Research Topic

The Future of Sport Business

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 Glob Health
• v.8(2); 2018 Dec

Artificial intelligence, machine learning and health systems

Trishan panch.

1 Wellframe Inc., Boston, Massachusetts, USA

Peter Szolovits

2 Department of Electrical Engineering and Computer Science, Computer Science and Artificial Intelligence Laboratory, Massachussetts Institute of Technology, Cambridge, Massachusetts, USA

3 Department of Global Health and Population, Harvard TH Chan School of Public Health, Harvard University, Boston, Massachusetts, USA

4 Department of Global Health and Social Medicine, Harvard Medical School, Harvard University, Boston, Massachusetts, USA

Associated Data

Globally, health systems face multiple challenges: rising burden of illness, multimorbidity and disability driven by ageing and epidemiological transition, greater demand for health services, higher societal expectations and increasing health expenditures [ 1 ]. A further challenge relates to inefficiency, with poor productivity [ 2 ]. These health system challenges exist against a background of fiscal conservatism, with misplaced economic austerity policies that are constraining investment in health systems.

Fundamental transformation of health systems is critical to overcome these challenges and to achieve universal health coverage (UHC) by 2030. Machine learning, the most tangible manifestation of artificial intelligence (AI) – and the newest growth area in digital technology – holds the promise of achieving more with less, and could be the catalyst for such a transformation [ 3 ]. But the nature and extent of this promise has not been systematically assessed.

To date, the impact of digital technology on health systems has been equivocal [ 4 ]. Is AI the ingredient for such a transformation, or will it face the same fate as earlier attempts at introducing digital technology? In this paper, we explore potential applications of AI in health systems and the ways in which AI could transform health systems to achieve UHC by improving efficiency, effectiveness, equity and responsiveness of public health and health care services.

EVOLUTION OF AI AND MACHINE LEARNING

AI is a broad discipline that aims to understand and design systems that display properties of intelligence ( Box 1 ) – emblematic of which is the ability to learn: to derive knowledge from data. This is a broad definition that arguably has some cross over with existing statistical techniques [ 6 ]. The recent explosion in progress in this field is attributable to a subset of AI – machine learning and one family of techniques in particular, deep learning, where computers are programmed to learn associations based on large quantities of raw data such as the pixels of digital images. Deep learning systems have been applied extensively and set new benchmarks in areas of the economy where high quality digital data are plentiful and there is a strong economic incentive to automate prediction tasks [ 5 ].

Evolution of Artificial Intelligence and Machine Learning.

Artificial Intelligence (AI)

A broad scientific discipline with its roots in philosophy, mathematics and computer science that aims to understand and develop systems that display properties of intelligence.

Machine Learning

A sub discipline of AI, where computers programs (algorithms) learn associations of predictive power from examples in data. Machine learning is most simply the application of statistical models to data using computers. Machine learning uses a broader set of statistical techniques than those typically used in medicine. Newer techniques such as Deep Learning are based on models with less assumptions about the underlying data and are therefore able to handle more complex data.

Deep Learning

Deep learning methods allow a machine to be fed with large quantities of raw data and to discover the representations necessary for detection or classification. Deep learning methods rely on multiple layers of representation of the data with successive transformations that amplify aspects of the input that are important for discrimination and suppress irrelevant variations . Deep learning may be supervised or unsupervised. Deep learning methods have been responsible for many of the recent foundational advances in machine learning [ 5 ].

Supervised Learning

Training computer programs to learn associations between inputs and outputs in data through analysis of outputs of interest defined by a (typically human) supervisor. Once associations have been learned based on existing data they can be used to predict future examples. This is one of the most established areas of machine learning with multiple examples inside and outside health care.

Unsupervised Learning

Computer programs that learn associations in data without external definition of associations of interest. Unsupervised learning is able to identify previously undiscovered predictors, as opposed to simply relying on known associations (s30).

Reinforcement Learning

Computer programs that learn actions based on their ability to maximize a defined reward. This approach is influenced by behavioural psychology and has been applied with considerable success in gaming where there is perfect information, many possible options and no real world cost of failure.

In these cases, deep learning algorithms have been able to uncover associations of predictive value, typically for a single use case, with large amounts of data and human expertise to curate the data and tune the algorithms involved [ 7 ]. These advances in machine learning are not a prototype for “artificial general intelligence”: a broad general-purpose intelligence that can, like the human brain, independently be deployed across use cases and independently incorporate learned concepts together in a self-reinforcing cycle.

AI AND DECISION MAKING IN HEALTH SYSTEMS

Effective management of health systems, like the provision of public health or health care, is in essence a lattice of information processing tasks. Policy makers modify health system functions of organisation and governance, financing and resource management to achieve health system outputs (health care services and public health) and system goals [ 8 ],.

The provision of health care itself involves two core information processing tasks: first, screening and diagnosis, which is the classification of cases based on history, examination and investigations, and second treatment and monitoring, which involves the planning, implementation and monitoring of a multistep process to deliver a future outcome.

The essential form of these processes across the domains of health system management and the provision of care involve hypothesis generation, hypothesis testing and action. Machine learning has the potential to improve hypothesis generation and hypothesis testing tasks within a health system by revealing previously hidden trends in data, and thus has the potential for substantial impact both at the individual patient and system level.

Machine learning expands on existing statistical techniques [ 6 ], utilising methods that are not based on a priori assumptions about the distribution of the data, and can find patterns in the data that can in turn be used to formulate hypotheses and hypothesis tests. Thus, whilst machine learning models are more difficult to interpret, they can incorporate many more variables and are generalizable across a much broader array of data types, and can produce results in more complex situations [ 9 ]. These methods have been deployed in the research context in screening and diagnosis and prediction of future events ( Table 1 ). These deployments are in disparate areas, typically in hospital rather than community setting, and in the vast majority of cases based on data from single centers, with implications for reproducibility [ 11 ] and generalizability [ 12 ]. However, the rapid pace of development of machine learning continues both within health care and more broadly across all information processing tasks in society [ 13 ].

Applications of artificial intelligence in diagnosis and prediction

*References s21 to s61 are available in the Online Supplementary Document (Online Supplementary Document) .

POTENTIAL EFFECT OF AI ON CLINICAL CARE AND HEALTH WORKFORCE

Machine learning has become a “General Purpose Technology”, in that it is pervasive, can be improved over time and has the potential to spawn complementary innovations [ 14 ]. The implementation of such technologies tends to result in “widespread economic disruption, with concomitant winners and losers” [ 15 ]. Economists Acemoglu and Restrepo, who studied the historical effects of automation – the process of substitution of mechanization for human labour – argue that automation exerts a displacement effect where human labour is displaced by machines in areas where machines have differential advantage [ 16 ]. However, countervailing forces that increase demand for labour offset this displacement effect: a productivity effect, as operations become more efficient and less costly. This in turn allows savings to be invested on existing non-automatable tasks and on the creation of new non-automatable tasks, some of which involve directly working on the automating technology.

To see how this general trend might apply to the health care workforce it is useful to examine the clinical area that is currently best represented in machine learning literature, diagnostic radiology.

As deep learning algorithms have set new performance benchmarks in diagnostic image analysis, some commentators have forecast the inevitable demise of radiologists and questioned the need for training new radiologists [ 17 ]. It is plausible that machine learning will enable existing radiologists to handle more cases and then, as machine learning systems are able to work more autonomously, to transfer responsibility for diagnostic image analysis to non-radiologists supported by machine learning systems. Such reorientation of tasks would create an opportunity for health systems to recalibrate the skill mix of radiology teams and their distribution, with more tasks done at the primary care level and non-automatable work and rarer cases handled by a smaller number of radiologists at secondary and tertiary centers.

The researchers behind a machine learning system responsible for pneumonia diagnosis [ 18 ] have developed a tool where the technology system “reads” the image first and highlights areas for the human radiologist to focus on – thereby improving workflow efficiency by allowing a human decision maker to focus her limited attention where it can be most effectively deployed and deal with many more cases [ 10 ]. One would expect the same applications to also transform pathology and other specialties reliant upon image analysis [ 19 , 20 ].

Machine learning will thus create processes performed by a hybrid of human and computer . These instances offer the potential to achieve optimal combination of leveraging human ability to generate hypotheses, collaborate and oversee AI systems to harness AI ability to analyse large volumes of data to find associations with predictive power or optimise against a success criterion. Jha and Topol propose that radiology and pathology should be amalgamated into a new specialty called an “information specialist”, whose responsibility will not be so much to extract information from images and histology but to manage the information extracted by artificial intelligence in the clinical context of the patient (reference s21 in Online Supplementary Document (Online Supplementary Document) ). It is plausible that the quality and scope of care will increase considerably whilst costs may stay relatively constant unless other specialties can be amalgamated in this way and more work shifted to primary health care.

Photo: iStock

CREATING A RECEPTIVE CONTEXT FOR HARNESSING THE BENEFITS OF AI IN HEALTH SYSTEMS

Machine learning is a rapidly advancing technology. Whilst there are significant technical leaps forward to come, as for any new technology, it will not just be technical challenges that limit the application of machine learning in health, but rather the absence of a receptive context for their adoption and diffusion (s22,s23).

A receptive context for AI requires, among others, availability of curated data, an enabling regulatory environment, legal provisions to safeguard citizens’ rights, clear rules on accountability, and capacity to manage strategic change to enable appropriate introduction and application of machine learning.

Curation of data

A perennial problem in health information technology is interoperability – the absence of a common data schema for health care organisations has made it difficult to combine data across a health system. Current approaches (s24) in health systems do not support existing data aggregation needs and should be reconsidered to support the additional data requirements of machine learning.

However, a recent unpublished study from Google and three academic medical centers demonstrated how data could be combined across three teaching hospitals without translation into a common format first. In this study deep learning was applied to data in its native format, without transformation into a common data structure first, and produced results exceeding existing performance benchmarks in predicting in-hospital mortality, 30-day unplanned readmission, prolonged length of stay and all of a patient’s final diagnoses (s25). Whilst advances in machine learning, such as this example, may reduce the need for data harmonization across a health system, they do not eliminate it. It is still important that health systems support initiatives to aggregate data both for existing functions and to support machine learning. In fact, the potential impact of machine learning should redouble motivation behind these initiatives.

It is likely that machine learning will deliver technically superior performance, but it will not be perfect. If successful, many people will benefit, but undoubtedly some will also be worse off. It is plausible that those negatively affected will be from marginalized groups who might be underrepresented in the data sets used to build machine learning algorithms. As such, while machine learning may deliver superior technical performance, it could compound inequities. To tackle the risk of further widening inequities, it is essential that there is adequate diversity of individuals represented in data sets, data are used from different clinical sites, and diversity is present in those developing machine learning algorithms. Without these provisions, examples of “algorithmic bias” (s26) and reinforcement of social inequities demonstrated elsewhere in society will be demonstrated in health care applications.

Trust and data management

An important barrier to the creation of data sets necessary for the development of machine learning systems is the lack of trust regarding how data will be used. Recent examples of over-zealous data sharing, where data have been shared transgressing legislative safeguards, have further eroded trust betwee citizens whose data are used and those using the data (s27).

Whilst in the consumer digital economy it is broadly accepted that as citizens we offer our data in return for better search results or a more relevant social network feed, for patients it is not clear that the same implicit understanding of the value of this quid pro quo is present. This could be because patients have simply not witnessed the benefits of data sharing in the way that retail consumers have. Further, retail and health sectors are qualitatively different, and the core concerns regarding trust in governments, privacy or fears regarding discrimination based on health status loom larger in patients’ minds than any potential distal benefits of data sharing.

Working with the technology industry

The inconvenient truth is that advances in machine learning are going to originate from or require co-operation with a handful of technology companies that have already invested billions of dollars to aggregate the intellectual capital and necessary computing and storage resources for machine learning. However such concentration accentuates broader concentration in the economy and introduces additional legislative complexity for national governments, which could end up being reliant on a handful of private technology companies for core infrastructure for AI.

As such, novel contracting mechanisms are necessary to work with these private technology companies to enable capture and use of health data at a national scale whilst maintaining privacy and fair attribution of intellectual property created. No such agreement is in place currently and there is no consensus about how to develop one. The absence of an agreed framework for contracting and intellectual property rights provides a significant opportunity for international public health organisations to display leadership and for the private corporate entities to demonstrate corporate social responsibility and to safeguard social solidarity to ensure the benefits of AI are broadly shared.

A broader conversation is needed between citizens, curators of data in health systems and private companies involved in AI and machine learning in health care. This conversation should seek to resolve issues pertaining to intellectual property related to health data, and the trade off between health data as a public good and as private capital and address patient concerns regarding privacy.

Accountability

Machine learning systems, in particular deep neural nets, are effectively ‘black boxes’ – their operations involve millions of data points used to calibrate models that can generate thousands of classifications – where the byzantine inner representation of said data are not typically intelligible to a human observer. As such, the internal process of generating an inference from data are not describable in the same way as traditional statistical models are.

The European Union is in the process of enforcing new General Data Protection Regulation which will effectively create a “right to explanation,” where an individual has the right to request an explanation of a decision that was made about them using “automated processing” (s28). This will generate challenges for decisions where processes generating them cannot be clearly explained.

In a recent paper on “Accountability of AI Under the Law”, Harvard University’s Berkman Klein working group on Explanation and the Law analyse the issue of accountability of machine learning systems (s29). They define mechanisms for accountability that are dependent on the type of problem being addressed by the machine learning system. For more defined problems theoretical constraints or statistical evidence from trials of machine learning systems might be sufficient, but in the type of problems that are experienced in clinical practice, where the objectives are not always clear and there is a high likelihood of external factors, explanation is necessary for accountability. Explanation is defined as permitting “an observer to determine the extent to which a particular input was determinative or influential to an output.” They recommend that an AI system should be expected to provide an explanation in situations where a human decision maker would be expected to do the same. In order to achieve this, however, technology investments are necessary to create distinct explanation systems that are able to communicate the inner workings of machine learning (in particular deep learning) algorithms, assuming that this is possible to do.

Capacity for managing strategic change

For machine learning based diagnosis, care management and monitoring to be adopted in practice, demonstration of algorithmic superiority alone will not be sufficient. To convince clinicians and policy makers, machine learning enabled systems will have to deliver outcomes of interest in practice through experimental trials or through real world observations of performance. Machine learning, however, is a moving target, and such initiatives may need to be repeated as algorithms improve with greater availability of data and better techniques. This may incur a significant cost to health systems, which will need to offset these costs by improvements in performance and health workforce efficiency.

The uncertainty around how machine learning will impact on the workforce – within health care and more broadly – is a concern to policy makers. Most likely, the impact of the aforementioned ‘displacement effect’ will be felt most acutely by those in lower skilled manual and non-manual occupations. In health systems with currently adequate numbers of health workforce, this displacement may generate a larger pool of workers seeking employment, particularly those involving psychological and emotional well-being and caring for the elderly and disabled – typically occupations that are considered skilled and non-automatable. With potentially a greater supply of front line care workers and machine learning systems, there is an opportunity for improved chronic disease management and community based care for ageing populations. However, this may not be sufficient to counteract the broader effects of automation on labour markets, and governments will need to make proactive investments in retraining workers displaced to prepare them for alternative opportunities, including new roles in the development and curation of data sets and machine learning algorithms.

Conversely, in low- and middle-income countries where there is an acute shortage of health workforce, machine learning offers the real opportunity to expand health care service coverage and increase the likelihood of achieving universal health coverage.

CONCLUSIONS

In this paper we have discussed the direct impact of machine learning on health systems, but have not explored the indirect effects of machine learning in basic sciences, drug discovery and other enabling technologies on health systems.

Prediction is inherently difficult: technology modifies its environment and the environment then generates further opportunities and new constraints for the technology. Ultimately, general purpose intelligence will be possible, as a version of it already exists in human brains. However, an extrapolation of existing techniques to re-create general intelligence artificially appears unlikely in the next 5-10 years.

However, what is immediately plausible, and should therefore be planned for, is a federation of ‘narrow’ and ‘targeted’ machine learning systems that are able to tackle core information processing problems across a health system by augmenting capabilities of human decision makers, and in so doing establishing new standards of effectiveness and efficiency in clinical and management operations. This is a significant opportunity for health system transformation as the cost of augmenting decision-making capabilities across the health system is unlikely to be large. There is no other approach that offers such potential impact without commensurate scaling of cost. The fixed cost involved in developing machine learning solutions: the cost of research and development and of re-tooling a health system is considerable, but given the potential scalability, the rationale to invest is clear.

An opportunity exists to seed growth in machine learning through the creation of high resolution clinical data sets and the necessary mechanisms for sharing of data and collaborative investigation to establish both efficacy and safety. What is currently missing in health systems is the leadership to do so. Whilst the issues raised are being actively discussed among the academic AI community, the academic AI community alone will not be able to solve them – it will require leadership from policy makers and the engagement of citizens, patients and clinicians. The fear of wholesale displacement of health workforce by AI is overstated, but where fear is warranted is in considering the opportunity cost of not embracing AI, of continuing business as usual with piecemeal implementation of AI that does not realize its potential for transformation of health systems.

Funding: None

Authorship contributions: RA and TP conceived the study. TP wrote the first draft with input from RA. All authors contributed to the subsequent drafts and the final manuscript.

Competing interests: The authors have completed the Unified Competing Interest form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare no competing interests.

Additional Material

Artificial intelligence, machine learning and deep learning

Ieee account.

• Change Username/Password
• Update Address

Purchase Details

• Payment Options
• Order History
• View Purchased Documents

Profile Information

• Communications Preferences
• Profession and Education
• Technical Interests
• US & Canada: +1 800 678 4333
• Worldwide: +1 732 981 0060
• Contact & Support
• About IEEE Xplore
• Accessibility
• Terms of Use
• Nondiscrimination Policy
• Privacy & Opting Out of Cookies

A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity. © Copyright 2024 IEEE - All rights reserved. Use of this web site signifies your agreement to the terms and conditions.

Help | Advanced Search

Computer Science > Artificial Intelligence

Title: dominion: a new frontier for ai research.

Abstract: In recent years, machine learning approaches have made dramatic advances, reaching superhuman performance in Go, Atari, and poker variants. These games, and others before them, have served not only as a testbed but have also helped to push the boundaries of AI research. Continuing this tradition, we examine the tabletop game Dominion and discuss the properties that make it well-suited to serve as a benchmark for the next generation of reinforcement learning (RL) algorithms. We also present the Dominion Online Dataset, a collection of over 2,000,000 games of Dominion played by experienced players on the Dominion Online webserver. Finally, we introduce an RL baseline bot that uses existing techniques to beat common heuristic-based bots, and shows competitive performance against the previously strongest bot, Provincial.

Submission history

Access paper:.

• HTML (experimental)
• Other Formats

References & Citations

• Google Scholar
• Semantic Scholar

BibTeX formatted citation

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

• Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .

CS50's Introduction to Artificial Intelligence with Python

Learn to use machine learning in Python in this introductory course on artificial intelligence.

Associated Schools

Harvard School of Engineering and Applied Sciences

What you'll learn.

Graph search algorithms

Reinforcement learning

Machine learning

Artificial intelligence principles

How to design intelligent systems

How to use AI in Python programs

Course description

AI is transforming how we live, work, and play. By enabling new technologies like self-driving cars and recommendation systems or improving old ones like medical diagnostics and search engines, the demand for expertise in AI and machine learning is growing rapidly. This course will enable you to take the first step toward solving important real-world problems and future-proofing your career.

CS50’s Introduction to Artificial Intelligence with Python explores the concepts and algorithms at the foundation of modern artificial intelligence, diving into the ideas that give rise to technologies like game-playing engines, handwriting recognition, and machine translation. Through hands-on projects, students gain exposure to the theory behind graph search algorithms, classification, optimization, reinforcement learning, and other topics in artificial intelligence and machine learning as they incorporate them into their own Python programs. By course’s end, students emerge with experience in libraries for machine learning as well as knowledge of artificial intelligence principles that enable them to design intelligent systems of their own.

Enroll now to gain expertise in one of the fastest-growing domains of computer science from the creators of one of the most popular computer science courses ever, CS50. You’ll learn the theoretical frameworks that enable these new technologies while gaining practical experience in how to apply these powerful techniques in your work.

Instructors

David J. Malan

You may also like

Using Python for Research

Take your introductory knowledge of Python programming to the next level and learn how to use Python 3 for your research.

CS50: Introduction to Computer Science

An introduction to the intellectual enterprises of computer science and the art of programming.

CS50's Understanding Technology

This is CS50’s introduction to technology for students who don’t (yet!) consider themselves computer persons.

Original Investigation

Robert J. Gallo, MD, Lisa Shieh, MD, PhD et al.

Annette Flanagin, RN, MA, Romain Pirracchio, MD, MPH, PhD et al.

• Related Article

Pierre Rinder, MSc, Théo Marcille, MSc et al.

Elif Baran, HBSc, Melissa Lee, MD et al.

Zhuo Chen, PhD, Jean-Eudes Dazard, PhD et al.

Jessica Ellis, BSA, Mika K. Hamer, PhD, MPH et al.

Dennis Ren, MD, Andrew James Tagg, MBBS, MRCSEd et al.

Corinne Andriola, PhD, Randall P. Ellis, PhD et al.

Ming Tai-Seale, PhD, MPH, Sally L. Baxter, MD, MSc et al.

Sara Habib, MD, Haroon Butt, MD et al.

Amulya Yalamanchili, MD, Bishwambhar Sengupta, PhD et al.

Timothy A. Rengers, BS, Cornelius A. Thiels, DO, MBA et al.

Michael D. Howell, MD, MPH, Greg S. Corrado, PhD et al.

Daniel A. Hashimoto, MD, Julian Varas, MD et al.

Nigam H. Shah, MBBS, PhD, John D. Halamka, MD, MS et al.

Robert M. Wachter, MD, Erik Brynjolfsson, PhD

Bradley D. Menz, BPharm(Hons), Natansh D. Modi, BPharm(Hons) et al.

Ramsey M. Wehbe, MD, MSAI, Aggelos K. Katsaggelos, PhD et al.

James Henderson, PhD, Anthony Cuttitta, MPH et al.

Nigam H. Shah, MBBS, PhD, David Entwistle, BS, MHSA et al.

Research Letter

Hawazin W. Elani, BDS, PhD, William V. Giannobile, DDS, DMSc

Jonah Zaretsky, MD, Jeong Min Kim, MD et al.

Invited Commentary

Charumathi Raghu Subramanian, MD, Daniel A. Yang, MD et al.

Raj M. Ratwani, PhD, David W. Bates, MD et al.

Yen Nhi Hoang, MSc, Ya-Ling Chen, PhD et al.

Alaa Youssef, PhD, Madelena Y. Ng, DrPH, MPH et al.

Special Communication

Marshall H. Chin, MD, MPH, Nasim Afsar-Manesh, MD, MBA, MHM et al.

_RightRail_Socials

Samantha Anderer, Yulin Hswen, ScD, MPH

Jennifer Abbasi, Yulin Hswen, ScD, MPH

Christine J. Colasacco, BS, Hayley L. Born, MD

Kirsten Bibbins-Domingo, PhD, MD, MAS, Derek C. Angus, MD, MPH et al.

Theodore J. Strange, MD, Mario R. Castellanos, MD

Brittany M. Gouse, MD, MPH, Janice M. Weinberg, ScD et al.

Nicholas Pratt, MD, RN, Ricky Madhavan, MD et al.

Sharon K. Inouye, MD, MPH

Julia Adler-Milstein, PhD, Donald A. Redelmeier, MD et al.

New machine learning algorithm promises advances in computing

Digital twin models may enhance future autonomous systems.

Systems controlled by next-generation computing algorithms could give rise to better and more efficient machine learning products, a new study suggests.

Using machine learning tools to create a digital twin, or a virtual copy, of an electronic circuit that exhibits chaotic behavior, researchers found that they were successful at predicting how it would behave and using that information to control it.

Many everyday devices, like thermostats and cruise control, utilize linear controllers -- which use simple rules to direct a system to a desired value. Thermostats, for example, employ such rules to determine how much to heat or cool a space based on the difference between the current and desired temperatures.

Yet because of how straightforward these algorithms are, they struggle to control systems that display complex behavior, like chaos.

As a result, advanced devices like self-driving cars and aircraft often rely on machine learning-based controllers, which use intricate networks to learn the optimal control algorithm needed to best operate. However, these algorithms have significant drawbacks, the most demanding of which is that they can be extremely challenging and computationally expensive to implement.

Now, having access to an efficient digital twin is likely to have a sweeping impact on how scientists develop future autonomous technologies, said Robert Kent, lead author of the study and a graduate student in physics at The Ohio State University.

"The problem with most machine learning-based controllers is that they use a lot of energy or power and they take a long time to evaluate," said Kent. "Developing traditional controllers for them has also been difficult because chaotic systems are extremely sensitive to small changes."

These issues, he said, are critical in situations where milliseconds can make a difference between life and death, such as when self-driving vehicles must decide to brake to prevent an accident.

The study was published recently in Nature Communications.

Compact enough to fit on an inexpensive computer chip capable of balancing on your fingertip and able to run without an internet connection, the team's digital twin was built to optimize a controller's efficiency and performance, which researchers found resulted in a reduction of power consumption. It achieves this quite easily, mainly because it was trained using a type of machine learning approach called reservoir computing.

"The great thing about the machine learning architecture we used is that it's very good at learning the behavior of systems that evolve in time," Kent said. "It's inspired by how connections spark in the human brain."

Although similarly sized computer chips have been used in devices like smart fridges, according to the study, this novel computing ability makes the new model especially well-equipped to handle dynamic systems such as self-driving vehicles as well as heart monitors, which must be able to quickly adapt to a patient's heartbeat.

"Big machine learning models have to consume lots of power to crunch data and come out with the right parameters, whereas our model and training is so extremely simple that you could have systems learning on the fly," he said.

To test this theory, researchers directed their model to complete complex control tasks and compared its results to those from previous control techniques. The study revealed that their approach achieved a higher accuracy at the tasks than its linear counterpart and is significantly less computationally complex than a previous machine learning-based controller.

"The increase in accuracy was pretty significant in some cases," said Kent. Though the outcome showed that their algorithm does require more energy than a linear controller to operate, this tradeoff means that when it is powered up, the team's model lasts longer and is considerably more efficient than current machine learning-based controllers on the market.

"People will find good use out of it just based on how efficient it is," Kent said. "You can implement it on pretty much any platform and it's very simple to understand." The algorithm was recently made available to scientists.

Outside of inspiring potential advances in engineering, there's also an equally important economic and environmental incentive for creating more power-friendly algorithms, said Kent.

As society becomes more dependent on computers and AI for nearly all aspects of daily life, demand for data centers is soaring, leading many experts to worry over digital systems' enormous power appetite and what future industries will need to do to keep up with it.

And because building these data centers as well as large-scale computing experiments can generate a large carbon footprint, scientists are looking for ways to curb carbon emissions from this technology.

To advance their results, future work will likely be steered toward training the model to explore other applications like quantum information processing, Kent said. In the meantime, he expects that these new elements will reach far into the scientific community.

"Not enough people know about these types of algorithms in the industry and engineering, and one of the big goals of this project is to get more people to learn about them," said Kent. "This work is a great first step toward reaching that potential."

This study was supported by the U.S. Air Force's Office of Scientific Research. Other Ohio State co-authors include Wendson A.S. Barbosa and Daniel J. Gauthier.

• Electronics
• Telecommunications
• Energy Technology
• Computers and Internet
• Neural Interfaces
• Information Technology
• Computer Science
• Artificial intelligence
• Alan Turing
• Computing power everywhere
• Grid computing
• Data mining

Story Source:

Materials provided by Ohio State University . Original written by Tatyana Woodall. Note: Content may be edited for style and length.

Journal Reference :

• Robert M. Kent, Wendson A. S. Barbosa, Daniel J. Gauthier. Controlling chaos using edge computing hardware . Nature Communications , 2024; 15 (1) DOI: 10.1038/s41467-024-48133-3

Cite This Page :

Explore More

• Toward a Successful Vaccine for HIV
• Highly Efficient Thermoelectric Materials
• Toward Human Brain Gene Therapy
• Whale Families Learn Each Other's Vocal Style
• AI Can Answer Complex Physics Questions
• Otters Use Tools to Survive a Changing World
• Monogamy in Mice: Newly Evolved Type of Cell
• Sustainable Electronics, Doped With Air
• Male Vs Female Brain Structure
• Breeding 'Carbon Gobbling' Plants

Trending Topics

Strange & offbeat.

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• My Account Login
• Explore content
• About the journal
• Publish with us
• Sign up for alerts
• Open access
• Published: 17 June 2020

Ethical principles in machine learning and artificial intelligence: cases from the field and possible ways forward

• Samuele Lo Piano   ORCID: orcid.org/0000-0002-2625-483X 1 , 2  

Humanities and Social Sciences Communications volume  7 , Article number:  9 ( 2020 ) Cite this article

97k Accesses

96 Citations

176 Altmetric

Metrics details

• Science, technology and society

Decision-making on numerous aspects of our daily lives is being outsourced to machine-learning (ML) algorithms and artificial intelligence (AI), motivated by speed and efficiency in the decision process. ML approaches—one of the typologies of algorithms underpinning artificial intelligence—are typically developed as black boxes. The implication is that ML code scripts are rarely scrutinised; interpretability is usually sacrificed in favour of usability and effectiveness. Room for improvement in practices associated with programme development have also been flagged along other dimensions, including inter alia fairness, accuracy, accountability, and transparency. In this contribution, the production of guidelines and dedicated documents around these themes is discussed. The following applications of AI-driven decision-making are outlined: (a) risk assessment in the criminal justice system, and (b) autonomous vehicles, highlighting points of friction across ethical principles. Possible ways forward towards the implementation of governance on AI are finally examined.

Similar content being viewed by others

Determinants of behaviour and their efficacy as targets of behavioural change interventions

Augmenting large language models with chemistry tools

AI-enabled electrocardiography alert intervention and all-cause mortality: a pragmatic randomized clinical trial

Introduction.

Artificial intelligence (AI) is the branch of computer science that deals with the simulation of intelligent behaviour in computers as regards their capacity to mimic, and ideally improve , human behaviour. To achieve this, the simulation of human cognition and functions, including learning and problem-solving, is required (Russell, 2010 ). This simulation may limit itself to some simple predictable features, thus limiting human complexity (Cowls, 2019 ).

AI became a self-standing discipline in the year 1955 (McCarthy et al., 2006 ) with significant development over the last decades. AI resorts to ML to implement a predictive functioning based on data acquired from a given context. The strength of ML resides in its capacity to learn from data without need to be explicitly programmed (Samuel, 1959 ); ML algorithms are autonomous and self-sufficient when performing their learning function. This is the reason why they are ubiquitous in AI developments. Further to this, ML implementations in data science and other applied fields are conceptualised in the context of a final decision-making application, hence their prominence.

Applications in our daily lives encompass fields, such as (precision) agriculture (Sennaar, 2019 ), air combat and military training (Gallagher, 2016 ; Wong, 2020 ), education (Sears, 2018 ), finance (Bahrammirzaee, 2010 ), health care (Beam and Kohane, 2018 ), human resources and recruiting (Hmoud and Laszlo, 2019 ), music composition (Cheng, 2009/09 ), customer service (Kongthon et al., 2009 ), reliable engineering and maintenance (Dragicevic et al., 2019 ), autonomous vehicles and traffic management (Ye, 2018 ), social-media news-feed (Rader et al., 2018 ), work scheduling and optimisation (O’Neil, 2016 ), and several others.

In all these fields, an increasing amount of functions are being ceded to algorithms to the detriment of human control, raising concern for loss of fairness and equitability (Sareen et al., 2020 ). Furthermore, issues of garbage-in-garbage-out (Saltelli and Funtowicz, 2014 ) may be prone to emerge in contexts when external control is entirely removed. This issue may be further exacerbated by the offer of new services of auto-ML (Chin, 2019 ), where the entire algorithm development workflow is automatised and the residual human control practically removed.

In the following sections, we will (i) detail a series of research questions around the ethical principles in AI; (ii) take stock of the production of guidelines elaborated in the field; (iii) showcase their prominence in practical examples; and (iv) discuss actions towards the inclusion of these dimensions in the future of AI ethics.

Research questions on the ethical dimensions of artificial intelligence

Critical aspects in AI deployment have already gained traction in mainstreaming literature and media. For instance, according to O’Neil ( 2016 ), a main shortcoming of ML approaches is the fact these resort to proxies for driving trends, such as person’s ZIP code or language in relation with the capacity of an individual to pay back a loan or handle a job, respectively. However, these correlations may be discriminatory, if not illegal.

Potential black swans (Taleb, 2007 ) in the code should also be considered. These have been documented, for instance, in the case of the Amazon website, for which errors, such as the quotation of plain items (often books) up to 10,000 dollars (Smith, 2018 ) have been reported. While mistakes about monetary values may be easy to spot, the situation may become more complex and less intelligible when incommensurable dimensions come to play. That is the reason why a number of guidelines on the topic of ethics in AI have been proliferating over the last few years.

While reflections around the ethical implications of machines and automation deployment were already put forth in the ’50s and ’60s (Samuel, 1959 ; Wiener, 1988 ), the increasing use of AI in many fields raises new important questions about its suitability (Yu et al., 2018 ). This stems from the complexity of the aspects undertaken and the plurality of views, stakes, and values at play. A fundamental aspect is how and to what extent the values and the perspectives of the involved stakeholders have been taken care of in the design of the decision-making algorithm (Saltelli, 2020 ). In addition to this ex-ante evaluation, an ex-post evaluation would need to be put in place so as to monitor the consequences of AI-driven decisions in making winners and losers.

To wrap up, it is fundamental to assess how and if ethical aspects have been included in the AI-driven decision-making implemented by asking questions such as:

What are the most prominent ethical concerns raised by large-scale deployment of AI applications?

How are these multiple dimensions interwoven?

What are the actions the involved stakeholders are carrying out to address these concerns?

What are possible ways forward to improve ML and AI development and use over their full life-cycle?

We will firstly examine the production of relevant guidelines in the fields along with academic secondary literature. These aspects will then be discussed in the context of two applied cases: (i) recidivism-risk assessment in the criminal justice system, and (ii) autonomous vehicles.

Guidelines and secondary literature on AI ethics, its dimensions and stakes

The production of dedicated documents has been skyrocketing from 2016 (Jobin et al., 2019 ). We here report on the most prominent international initiatives. A suggested reading on national and international AI strategies providing a comprehensive list of documents (Future of Earth Institute, 2020 ).

The France’s Digital Republic Act gives the right to an explanation as regards decisions on an individual made through the use of administrative algorithms (Edwards and Veale, 2018 ). This law touches upon several aspects including:

how and to what extent the algorithmic processing contributed to the decision-making;

which data was processed and its source;

how parameters were treated and weighted;

which operations were carried out in the treatment.

Sensitive governmental areas, such as national security and defence, and the private sector (the largest user and producer of ML algorithms by far) are excluded from this document.

An international European initiative is the multi-stakeholder European Union High-Level Expert Group on Artificial Intelligence , which is composed by 52 experts from academia, civil society, and industry. The group produced a deliverable on the required criteria for AI trustworthiness (Daly, 2019 ). Even articles 21 and 22 of the recent European Union General Data Protection Regulation include passages functional to AI governance, although further action has been recently demanded from the European Parliament (De Sutter, 2019 ). In this context, China has also been allocating efforts on privacy and data protection (Roberts, 2019 ).

As regards secondary literature, Floridi and Cowls ( 2019 ) examined a list of statements/declarations elaborated since 2016 from multi-stakeholder organisations. A set of 47 principles has been identified, which mapped onto five overarching dimensions (Floridi and Cowls, 2019 ): beneficence, non-maleficence, autonomy, justice and, explicability . The latter is a new dimension specifically acknowledged in the case of AI, while the others were already identified in the controversial domain of bioethics .

Jobin et al. ( 2019 ) reviewed 84 documents, which were produced by several actors of the field, almost half of which from private companies or governmental agencies. The classification proposed by Jobin et al. ( 2019 ) is around a slightly different set of values: transparency, justice and fairness, non-maleficience, responsibility and privacy . Other potentially relevant dimensions, such as accountability and responsibility, were rarely defined in the studies reviewed by these authors.

Seven of the most prominent value statements from the AI/ML fields were examined in Greene et al. ( 2019 ): The Partnership on AI to Benefit People and Society ; The Montreal Declaration for a Responsible Development of Artificial Intelligence ; The Toronto Declaration Protecting the rights to equality and non-discrimination in machine-learning systems ; OpenAI ; The Centre for Humane Technology ; Fairness, Accountability and Transparency in Machine Learning ; Axon’s AI Ethics Board for Public Safety . Greene et al. ( 2019 ) found seven common core elements across these documents: (i) design’s moral background (universal concerns, objectively measured); (ii) expert oversight; (iii) values-driven determinism; (iv) design as locus of ethical scrutiny; (v) better building; (vi) stakeholder-driven legitimacy; and, (vii) machine translation.

Mittelstadt ( 2019 ) critically analysed the current debate and actions in the field of AI ethics and noted that the dimensions addressed in AI ethics are converging towards those of medical ethics. However, this process appears problematic due to four main differences between medicine and the medical professionals on one side, and AI and its developers on the other. Firstly, the medical professional rests on common aims and fiduciary duties, which AI developers lack. Secondly, a formal profession with a set of clearly defined and governed good-behaviour practices exists in medicine. This is not the case for AI, which also lacks a full understanding of the consequences of the actions enacted by algorithms (Wallach and Allen, 2008 ). Thirdly, AI faces the difficulty of translating overarching principle into practices. Even its current setting of seeking maximum speed, efficiency and profit clashes with the resource and time requirements of an ethical assessment and/or counselling. Finally, the accountability of professionals or institutions is at this stage mainly theoretical, having the vast majority of these guidelines been defined on a merely voluntary basis and hence with the total lack of a sanctionary scheme for non-compliance.

Points of friction between ethical dimensions

Higher transparency is a common refrain when discussing ethics of algorithms, in relation to dimensions such as how an algorithmic decision is arrived at, based on what assumptions, and how this could be corrected to incorporate feedback from the involved parties. Rudin ( 2019 ) argued that the community of algorithm developers should go beyond explaining black-box models by developing interpretable models in the first place.

On a larger scale, the use of open-source software in the context of ML applications has already been advocated for over a decade (Thimbleby, 2003 ) with an indirect call for tools to execute more interpretable and reproducible programming such as Jupyter Notebooks , available from 2015 onwards. However, publishing scripts expose their developers to the public scrutiny of professional programmers, who may find shortcomings in the development of the code (Sonnenburg, 2007 ).

Ananny and Crawford ( 2018 ) comment that resorting to full algorithmic transparency may not be an adequate means to address their ethical dimensions; opening up the black-box would not suffice to disclose their modus operandi . Moreover, developers of algorithm may not be capable of explaining in plain language how a given tool works and what functional elements it is based on. A more social relevant understanding would encompass the human/non-human interface (i.e., looking across the system rather than merely inside ). Algorithmic complexity and all its implications unravel at this level, in terms of relationships rather than as mere self-standing properties.

Other authors pointed to possible points of friction between transparency and other relevant ethical dimensions. de Laat ( 2018 ) argues that transparency and accountability may even be at odds in the case of algorithms. Hence, he argues against full transparency along four main lines of reasoning: (i) leaking of privacy sensitive data into the open; (ii) backfiring into an implicit invitation to game the system; (iii) harming of the company property rights with negative consequences on their competitiveness (and on the developers reputation as discussed above); (iv) inherent opacity of algorithms, whose interpretability may be even hard for experts (see the example below about the code adopted in some models of autonomous vehicles). All these arguments suggest limitations to full disclosure of algorithms, be it that the normative implications behind these objections should be carefully scrutinised.

Raji et al. ( 2020 ) suggest that a process of algorithmic auditing within the software-development company could help in tackling some of the ethical issues raised. Larger interpretability could be in principle achieved by using simpler algorithms, although this may come at the expenses of accuracy. To this end, Watson and Floridi ( 2019 ) defined a formal framework for interpretable ML, where explanatory accuracy can be assessed against algorithmic simplicity and relevance.

Loss in accuracy may be produced by the exclusion of politically critical features (such as gender, race, age, etc.) from the pool of training predictive variables. For instance, Amazon scrapped a gender-biased recruitment algorithm once it realised that despite excluding gender, the algorithm was resorting to surrogate gender variables to implement its decisions (Dastin, 2018 ). This aspect points again to possible political issues of a trade-off between fairness, demanded by society, and algorithmic accuracy, demanded by, e.g., a private actor.

Fairness may be further hampered by reinforcement effects. This is the case of algorithms attributing credit scores, that have a reinforcement effect proportional to people wealth that de facto rules out credit access for people in a more socially difficult condition (O’Neil, 2016 ).

According to Floridi and Cowls ( 2019 ) a prominent role is also played by the autonomy dimension; the possibility of refraining from ceding decision power to AI for overriding reasons (e.g., the gain of efficacy is not deemed fit to justify the loss of control over decision-making). In other words, machines autonomy could be reduced in favour of human autonomy according to this meta-autonomy dimension.

Contrasting dimensions in terms of the theoretical framing of the issue also emerged from the review of Jobin et al. ( 2019 ), as regards interpretation of ethical principles, reasons for their importance, ownership and responsibility of their implementation. This also applies to different ethical principles, resulting in the trade-offs previously discussed, difficulties in setting prioritisation strategies, operationalisation and actual compliance with the guidelines. For instance, while private actors demand and try to cultivate trust from their users, this runs counter to the need for society to scrutinise the operation of algorithms in order to maintain developer accountability (Cowls, 2019 ). Attributing responsibilities in complicated projects where many parties and developers may be involved, an issue known as the problem of many hands (Nissenbaum, 1996 ), may indeed be very difficult.

Conflicts may also emerge between the requirements to overcome potential algorithm deficits in accuracy associated with large data bases and the individual rights to privacy and autonomy of decision. Such conflicts may exacerbate tensions, further complicating agreeing on standards and practices.

In the following two sections, the issues and points of friction raised are examined in two practical case studies, criminal justice and autonomous vehicles. These examples have been selected due to their prominence in the public debate on the ethical aspects of AI and ML algorithms.

Machine-learning algorithms in the field of criminal justice

ML algorithms have been largely used to assist juridical deliberation in many states of the USA (Angwin and Larson, 2016 ). This country faces the issue of the world’s highest incarcerated population, both in absolute and per-capita terms (Brief, 2020 ). The COMPAS algorithm, developed by the private company Northpointe , attributes a 2-year recidivism-risk score to arrested people. It also evaluates the risk of violent recidivism as a score.

The fairness of the algorithm has been questioned in an investigative report, that examined a pool of cases where a recidivism score was attributed to >18,000 criminal defendants in Broward County, Florida and flagged up a potential racial bias in the application of the algorithm (Angwin and Larson, 2016 ). According to the authors of the report, the recidivism-risk was systematically overestimated for black people: the decile distribution of white defendants was skewed towards the lower end. Conversely, the decile distribution of black defendants was only slightly decreasing towards the higher end. The risk of violent recidivism within 2 years followed a similar trend. This analysis was debunked by the company, which, however, refused to disclose the full details of its proprietary code. While the total number of variables amounts to about 140, only the core variables were disclosed (Northpointe, 2012 ). The race of the subject was not one of those.

Here, a crucial point is how this fairness is to be attained: whether it is more important a fair treatment across groups of individuals or within the same group. For instance, let us take the case of gender, where men are overrepresented in prison in comparison with women. As to account for this aspect, the algorithm may discount violent priors for men in order to reduce their recidivism-risk score. However, attaining this sort of algorithmic fairness would imply inequality of treatment across genders (Berk et al., 2018 ).

Fairness could be further hampered by the combined use of this algorithm with others driving decisions on neighbourhood police patrolling. The fact these algorithms may be prone to drive further patrolling in poor neighbourhoods may result from a training bias as crimes occurring in public tend to be more frequently reported (Karppi, 2018 ). One can easily understand how these algorithms may jointly produce a vicious cycle—more patrolling would lead to more arrests that would worsen the neighbourhood average recidivism-risk score , which would in turn trigger more patrolling. All this would result in exacerbated inequalities, likewise the case of credit scores previously discussed (O’Neil, 2016 ).

A potential point of friction may also emerge between the algorithm dimensions of fairness and accuracy. The latter may be theoretically defined as the classification error in terms of rate of false positive (individuals labelled at risk of recidivism, that did not re-offend within 2 years) and false negative (individuals labelled at low risk of recidivism, that did re-offend within the same timeframe) (Loi and Christen, 2019 ). Different classification accuracy (the fraction of observed outcomes in disagreement with the predictions) and forecasting accuracy (the fraction of predictions in disagreement with the observed outcomes) may exist across different classes of individuals (e.g., black or white defendants). Seeking equal rates of false positive and false negative across these two pools would imply a different forecasting error (and accuracy) given the different characteristics of the two different training pools available for the algorithm. Conversely, having the same forecasting accuracy would come at the expense of different classification errors between these two pools (Corbett-Davies et al., 2016 ). Hence, a trade-off exists between these two different shades of fairness, which derives from the very statistical properties of the data population distributions the algorithm has been trained on. However, the decision-making rests again on the assumptions the algorithm developers have adopted, e.g., on the relative importance of false positive and false negative (i.e., the weights attributed to the different typologies of errors, and the accuracy sought (Berk, 2019 )). When it comes to this point, an algorithm developer may decide (or be instructed) to train his/her algorithm to attribute, e.g., a five/ten/twenty times higher weight for a false negative (re-offender, low recidivism-risk score) in comparison with a false positive (non re-offender, high recidivism-risk score).

As with all ML, an issue of transparency exists as no one knows what type of inference is drawn on the variables out of which the recidivism-risk score is estimated. Reverse-engineering exercises have been run so as to understand what are the key drivers on the observed scores. Rudin ( 2019 ) found that the algorithm seemed to behave differently from the intentions of their creators (Northpointe, 2012 ) with a non-linear dependence on age and a weak correlation with one’s criminal history. These exercises (Rudin, 2019 ; Angelino et al., 2018 ) showed that it is possible to implement interpretable classification algorithms that lead to a similar accuracy as COMPAS. Dressel and Farid ( 2018 ) achieved this result by using a linear predictor-logistic regressor that made use of only two variables (age and total number of previous convictions of the subject).

Machine-learning algorithms in the field of autonomous vehicles

The case of autonomous vehicles, also known as self-driving vehicles, poses different challenges as a continuity of decisions is to be enacted while the vehicle is moving. It is not a one-off decision as in the case of the assessment of recidivism risk.

An exercise to appreciate the value-ladenness of these decisions is the moral-machine experiment (Massachussets Institute of Technology 2019 )—a serious game where users are requested to fulfil the function of an autonomous-vehicle decision-making algorithm in a situation of danger. This experiment entails performing choices that would prioritise the safety of some categories of users over others. For instance, choosing over the death of car occupants, pedestrians, or occupants of other vehicles, et cetera. While such extreme situations may be a simplification of reality, one cannot exclude that the algorithms driving an autonomous-vehicle may find themselves in circumstances where their decisions may result in harming some of the involved parties (Bonnefon et al., 2019 ).

In practice, the issue would be framed by the algorithm in terms of a statistical trolley dilemma in the words of Bonnefon et al. ( 2019 ), whereby the risk of harm for some road users will be increased. This corresponds to a risk management situation by all means, with a number of nuances and inherent complexity (Goodall, 2016 ).

Hence, autonomous vehicles are not bound to play the role of silver bullets, solving once and forever the vexing issue of traffic fatalities (Smith, 2018 ). Furthermore, the way decisions enacted could backfire in complex contexts to which the algorithms had no extrapolative power, is an unpredictable issue one has to deal with (Wallach and Allen, 2008 ; Yurtsever et al., 2020 ).

Coding algorithms that assure fairness in autonomous vehicles can be a very challenging issue. Contrasting and incommensurable dimensions are likely to emerge (Goodall, 2014 ) when designing an algorithm to reduce the harm of a given crash. For instance, in terms of material damage against human harm. Odds may emerge between the interest of the vehicle owner and passengers, on one side, and the collective interest of minimising the overall harm, on the other. Minimising the overall physical harm may be achieved by implementing an algorithm that, in the circumstance of an unavoidable collision, would target the vehicles with the highest safety standards. However, one may want to question the fairness of targeting those who have invested more in their own and others’ safety. The algorithm may also face a dilemma between low probability of a serious harm and higher probability of a mild harm. Unavoidable normative rules will need to be included in the decision-making algorithms to tackle these types of situations.

Accuracy in the context of self-autonomous vehicles rests on their capacity to correctly simulate the course of the events. While this is based on physics and can be informed by the numerous sensors these vehicles are equipped with, unforeseen events can still play a prominent role, and profoundly affect the vehicles behaviour and reactions (Yurtsever et al., 2020 ). For instance, fatalities due to autonomous-vehicle malfunctioning were reported as caused by the following failures: (i) the incapability of perceiving a pedestrian as such (National Transport Safety Board 2018 ); (ii) the acceleration of the vehicle in a situation when braking was required due to contrasting instructions from different algorithms the vehicle was hinged upon (Smith, 2018 ). In this latter case, the complexity of autonomous-vehicle algorithms was witnessed by the millions lines of code composing their scripts, a universe no one fully understands in the words of The Guardian (Smith, 2018 ), so that the causality of the decisions made was practically impossible to scrutinise. Hence, no corrective action in the algorithm code may be possible at this stage, with no room for improvement in accuracy.

One should also not forget that these algorithms are learning by direct experience and they may still end up conflicting with the initial set of ethical rules around which they have been conceived. Learning may occur through algorithms interaction taking place at a higher hierarchical level than the one imagined in the first place (Smith, 2018 ). This aspect would represent a further open issue to be taken into account in their development (Markham et al., 2018 ). It also poses further tension between the accuracy a vehicle manufacturer seeks and the capability to keep up the agreed fairness standards upstream from the algorithm development process.

Discussion and conclusions

In this contribution, we have examined the ethical dimensions affected by the application of algorithm-driven decision-making. These are entailed both ex-ante, in terms of the assumptions underpinning the algorithm development, and ex-post as regards the consequences upon society and social actors on whom the elaborated decisions are to be enforced.

Decision-making-based algorithms rest inevitably on assumptions, even silent ones, such as the quality of data the algorithm is trained on (Saltelli and Funtowicz, 2014 ), or the actual modelling relations adopted (Hoerl, 2019 ), with all the implied consequences (Saltelli, 2019 ).

A decision-making algorithm will always be based on a formal system, which is a representation of a real system (Rosen, 2005 ). As such, it will always be based on a restricted set of relevant relations, causes, and effects. It does not matter how complicated the algorithm may be (how many relations may be factored in), it will always represent one-specific vision of the system being modelled (Laplace, 1902 ).

Eventually, the set of decision rules underpinning the AI algorithm derives from human-made assumptions, such as, where to define the boundary between action and no action, between different possible choices. This can only take place at the human/non-human interface: the response of the algorithm is driven by these human-made assumptions and selection rules. Even the data on which an algorithm is trained on are not an objective truth, they are dependent upon the context in which they have been produced (Neff et al., 2017 ).

Tools for technically scrutinising the potential behaviour of an algorithm and its uncertainty already exist and could be included in the workflow of algorithm development. For instance, global sensitivity analysis (Saltelli, 2008 ) may help in exploring how the uncertainty in the input parameters and modelling assumptions would affect the output. Additionally, a modelling of the modelling process would assist in the model transparency and in addressing questions such as: Are the results from a particular model more sensitive to changes in the model and the methods used to estimate its parameters, or to changes in the data? (Majone, 1989 ).

Tools of post-normal-science inspiration for knowledge and modelling quality assessment could be adapted to the analysis of algorithms, such as the NUSAP (Numeral Unit Spread Assessment Pedigree) notation system for the management and communication of uncertainty (Funtowicz and Ravetz, 1990 ; Van Der Sluijs, 2005 ) and sensitivity auditing (Saltelli and Funtowicz, 2014 ), respectively. Ultimately, developers should acknowledge the limits of AI, and what its ultimate function should be in the equivalent of an Hippocratic Oath for ML developers (O’Neil, 2016 ). An example comes from the field of financial modelling, with a manifesto elaborated in the aftermath of the 2008 financial crisis (Derman and Wilmott, 2009 ).

As to address these dimensions, value statements and guidelines have been elaborated by political and multi-stakeholder organisations. For instance, The Alan Turing Institute released a guide for responsible design and implementation of AI (Leslie, 2019 ) that covers the whole life-cycle of design, use, and monitoring. However, the field of AI ethics is just at its infancy and it is still to be conceptualised how AI developments that encompass ethical dimensions could be attained. Some authors are pessimistic, such as Supiot ( 2017 ) who speaks of governance by numbers , where quantification is replacing the traditional decision-making system and profoundly affecting the pillar of equality of judgement. Trying to revert the current state of affairs may expose the first movers in the AI field to a competitive disadvantage (Morley et al., 2019 ). One should also not forget that points of friction across ethical dimensions may emerge, e.g., between transparency and accountability, or accuracy and fairness as highlighted in the case studies. Hence, the development process of the algorithm cannot be perfect in this setting, one has to be open to negotiation and unavoidably work with imperfections and clumsiness (Ravetz, 1987 ).

The development of decision-making algorithms remains quite obscure in spite of the concerns raised and the intentions manifested to address them. Attempts to expose to public scrutiny the algorithms developed are yet scant. As are the attempt to make the process more inclusive, with a higher participation from all the stakeholders. Identifying a relevant pool of social actors may require an important effort in terms of stakeholders’ mapping so as to assure a complete, but also effective, governance in terms of number of participants and simplicity of working procedures. The post-normal-science concept of extended peer communities could assist also in this endeavour (Funtowicz and Ravetz, 1997 ). Example-based explanations (Molnar, 2020 ) may also contribute to an effective engagement of all the parties by helping in bridging technical divides across developers, experts in other fields, and lay-people.

An overarching meta-framework for the governance of AI in experimental technologies (i.e., robot use) has also been proposed (Rego de Almeida et al., 2020 ). This initiative stems from the attempt to include all the forms of governance put forth and would rest on an integrated set of feedback and interactions across dimensions and actors. An interesting proposal comes from Berk ( 2019 ), who asked for the intervention of super partes authorities to define standards of transparency, accuracy and fairness for algorithm developers in line with the role of the Food and Drug administration in the US and other regulation bodies. A shared regulation could help in tackling the potential competitive disadvantage a first mover may suffer. The development pace of new algorithms would be necessarily reduced so as to comply with the standards defined and the required clearance processes. In this setting, seeking algorithm transparency would not be harmful for their developers as scrutiny would be delegated to entrusted intermediate parties, to take place behind closed doors (de Laat, 2018 ).

As noted by a perceptive reviewer, ML systems that keep learning are dangerous and hard to understand because they can quickly change. Thus, could a ML system with real world consequences be “locked down” to increase transparency? If yes, the algorithm could become defective. If not, transparency today may not be helpful in understanding what the system does tomorrow. This issue could be tackled by hard-coding the set of rules on the behaviour of the algorithm, once these are agreed upon among the involved stakeholders. This would prevent the algorithm-learning process from conflicting with the standards agreed. Making mandatory to deposit these algorithms in a database owned and operated by this entrusted super-partes body could ease the development of this overall process.

Ananny M, Crawford K (2018) Seeing without knowing: Limitations of the transparency ideal and its application to algorithmic accountability. New Media & Society 20:973–989

Article   Google Scholar  

Angelino E, Larus-Stone N, Alabi D, Seltzer M, Rudin C (2018) Learning certifiably optimal rule lists for categorical data. http://arxiv.org/abs/1704.01701

Angwin J, Larson J (2016) Machine bias. https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing

Bahrammirzaee A (2010) A comparative survey of artificial intelligence applications in finance: artificial neural networks, expert system and hybrid intelligent systems. Neural Comput Appl 19:1165–1195

Beam AL, Kohane IS (2018) Big data and machine learning in health care. JAMA 319:1317

Berk R (2019) Machine learning risk assessments in criminal justice settings. Springer International Publishing, Cham

Berk R, Heidari H, Jabbari S, Kearns M, Roth A (2018) Fairness in criminal justice risk assessments: the state of the art. Soc Methods Res 004912411878253

Board NTS (2018) Vehicle automation report. Tech. Rep. HWY18MH010, Office of Highway Safety, Washington, D.C.

Bonnefon J-F, Shariff A, Rahwan I (2019) The trolley, the bull bar, and why engineers should care about the ethics of autonomous cars [point of view]. Proc IEEE 107:502–504

Brief WP (2020) World prison brief- an online database comprising information on prisons and the use of imprisonment around the world. https://www.prisonstudies.org/

Cheng J (2009) Virtual composer makes beautiful music and stirs controversy. https://arstechnica.com/science/news/2009/09/virtual-composer-makes-beautiful-musicand-stirs-controversy.ars

Chin J (2019) The death of data scientists. https://towardsdatascience.com/the-death-of-data-scientists-c243ae167701

Corbett-Davies S, Pierson E, Feller A, Goel S (2016) A computer program used for bail and sentencing decisions was labeled biased against blacks. It’s actually not that clear. Washington Post. https://www.washingtonpost.com/news/monkey-cage/wp/2016/10/17/can-an-algorithm-be-racist-our-analysis-is-more-cautious-than-propublicas/

Cowls J (2020) Deciding how to decide: six key questions for reducing AI’s democratic deficit. In: Burr C, Milano S (eds) The 2019 Yearbook of the Digital Ethics Lab, Digital ethics lab yearbook. Springer International Publishing, Cham. pp. 101–116. https://doi.org/10.1007/978-3-030-29145-7_7

Daly A et al. (2019) Artificial intelligence, governance and ethics: global perspectives. SSRN Electron J. https://www.ssrn.com/abstract=3414805

Dastin J (2018) Amazon scraps secret AI recruiting tool that showed bias against women. Reuters. https://www.reuters.com/article/us-amazon-com-jobs-automation-insight-idUSKCN1MK08G

De Sutter P (2020) Automated decision-making processes: ensuring consumer protection, and free movement of goods and services. https://www.europarl.europa.eu/meetdocs/2014_2019/plmrep/COMMITTEES/IMCO/DV/2020/01-22/Draft_OQ_Automated_decision-making_EN.pdf

Derman E, Wilmott P (2009) The financial modelers’ manifesto. SSRN Electron J. http://www.ssrn.com/abstract=1324878 .

Dragičević T, Wheeler P, Blaabjerg F (2019) Artificial intelligence aided automated design for reliability of power electronic systems. IEEE Trans Power Electron 34:7161–7171

Article   ADS   Google Scholar  

Dressel J, Farid H (2018) The accuracy, fairness, and limits of predicting recidivism. Sci Adv 4:eaao5580

Edwards L, Veale M (2018) Enslaving the algorithm: from A -right to an explanation- to A -right to better decisions-? IEEE Security, Priv 16:46–54

Floridi L, Cowls J (2019) A unified framework of five principles for AI in society. Harvard Data Science Review. https://hdsr.mitpress.mit.edu/pub/l0jsh9d1

Funtowicz SO, Ravetz JR (1990) Uncertainty and quality in science for policy. Springer Science, Business Media, Berlin, Heidelberg

Book   Google Scholar  

Funtowicz S, Ravetz J (1997) Environmental problems, post-normal science, and extended peer communities. Études et Recherches sur les Systémes Agraires et le Développement. INRA Editions. pp. 169–175

Future of Earth Institute (2020) National and International AI Strategies. https://futureoflife.org/national-international-ai-strategies/

Gallagher S (2016) AI bests Air Force combat tactics experts in simulated dogfights. https://arstechnica.com/information-technology/2016/06/ai-bests-air-force-combat-tactics-experts-in-simulated-dogfights/

Goodall NJ (2014) Ethical decision making during automated vehicle crashes. Transportation Res Rec: J Transportation Res Board 2424:58–65

Goodall NJ (2016) Away from trolley problems and toward risk management. Appl Artif Intell 30:810–821

Greene D, Hoffmann AL, Stark L (2019) Better, nicer, clearer, fairer: a critical assessment of the movement for ethical artificial intelligence and machine learning. In Proceedings of the 52nd Hawaii International Conference on System Sciences

Hmoud B, Laszlo V (2019) Will artificial intelligence take over human-resources recruitment and selection? Netw Intell Stud VII:21–30

Hoerl RW (2019) The integration of big data analytics into a more holistic approach-JMP. Tech. Rep., SAS Institute. https://www.jmp.com/en_us/whitepapers/jmp/integration-of-big-data-analytics-holistic-approach.html

Jobi A, Ienca M, Vayena E (2019) Artificial intelligence: the global landscape of ethics guidelines. Nat Mach Intell 1:389–399

Karppi T (2018) The computer said so-: on the ethics, effectiveness, and cultural techniques of predictive policing. Soc Media + Soc 4:205630511876829

Kongthon A, Sangkeettrakarn C, Kongyoung S, Haruechaiyasak C (2009) Implementing an online help desk system based on conversational agent. In: Proceedings of the International Conference on Management of Emergent Digital EcoSystems, MEDES ’09, vol. 69. ACM, New York, NY, USA. pp. 450–69:451. Event-place: France. https://doi.org/10.1145/1643823.1643908

de Laat PB (2018) Algorithmic decision-making based on machine learning from big data: can transparency restore accountability? Philos Technol 31:525–541

Laplace PS (1902) A philosophical essay on probabilities. J. Wiley, New York; Chapman, Hall, London. http://archive.org/details/philosophicaless00lapliala

Leslie D (2019) Understanding artificial intelligence ethics and safety. http://arxiv.org/abs/1906.05684

Loi M, Christen M (2019) How to include ethics in machine learning research. https://ercim-news.ercim.eu/en116/r-s/how-to-include-ethics-in-machine-learning-research

Majone G (1989) Evidence, argument, and persuasion in the policy process. Yale University Press, Yale

Google Scholar  

Markham AN, Tiidenberg K, Herman A (2018) Ethics as methods: doing ethics in the era of big data research-introduction. Soc Media + Soc 4:205630511878450

Massachussets Institute of Technology (2019) Moral machine. Massachussets Institute of Technology. http://moralmachine.mit.edu

McCarthy J, Minsky ML, Rochester N, Shannon CE (2006) A proposal for the dartmouth summer research project on artificial intelligence, August 31, 1955. AI Mag 27:12–12

Mittelstadt B (2019) Principles alone cannot guarantee ethical AI. Nat Mach Intell 1:501–507

Molnar C (2020) Interpretable machine learning (2020). https://christophm.github.io/interpretable-ml-book/

Morley J, Floridi L, Kinsey K, Elhalal A (2019) From what to how: an initial review of publicly available AI ethics tools, methods and research to translate principles into practices. Tech Rep. https://arxiv.org/abs/1905.06876

Neff G, Tanweer A, Fiore-Gartland B, Osburn L (2017) Critique and contribute: a practice-based framework for improving critical data studies and data science. Big Data 5:85–97

Nissenbaum H (1996) Accountability in a computerized society. Sci Eng Ethics 2:25–42

Northpointe (2012) Practitioner’s guide to COMPAS. northpointeinc.com/files/technical_documents/FieldGuide2_081412.pdf

O’Neil C (2016) Weapons of math destruction: how big data increases inequality and threatens democracy, 1st edn. Crown, New York

MATH   Google Scholar  

Rader E, Cotter K, Cho J (2018) Explanations as mechanisms for supporting algorithmic transparency. In: Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems - CHI ’1 8 . ACM Press, Montreal QC, Canada. pp. 1–13. http://dl.acm.org/citation.cfm?doid=3173574.3173677

Raji ID et al. Closing the AI accountability gap: defining an end-to-end framework for internal algorithmic auditing. In Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency pp 33–44 (Association for Computing Machinery, 2020). https://doi.org/10.1145/3351095.3372873

Ravetz JR (1987) Usable knowledge, usable ignorance: incomplete science with policy implications. Knowledge 9:87–116

Rêgo de Almeida PG, Denner dos Santos C, Silva Farias J (2020) Artificial intelligence regulation: a meta-framework for formulation and governance. In: Proceedings of the 53rd Hawaii International Conference on System Sciences (2020). http://hdl.handle.net/10125/64389

Roberts H et al. (2019) The Chinese approach to artificial intelligence: an analysis of policy and regulation. SSRN Electron J. https://www.ssrn.com/abstract=3469784

Rosen R (2005) Life itself: a comprehensive inquiry into the nature, origin, and fabrication of life. Columbia University Press, New York

Rudin C (2019) Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. http://arxiv.org/abs/1811.10154

Russell SJ (2010) Artificial intelligence : a modern approach. Prentice Hall, Upper Saddle River, NJ

Saltelli A et al. (2008) Global sensitivity analysis: the primer. Wiley, Hoboken, NJ

Saltelli A (2019) A short comment on statistical versus mathematical modelling. Nat Commun 10:3870

Saltelli A (2020) Ethics of quantification or quantification of ethics? Futures 116:102509

Saltelli A, Funtowicz S (2014) When all models are wrong. Issues Sci Technol 30:79–85

Samuel AL (1959) Some studies in machine learning using the game of checkers. IBM J Res Dev 3:210–229

Article   MathSciNet   Google Scholar  

Sareen S, Saltelli A, Rommetveit K (2020) Ethics of quantification: illumination, obfuscation and performative legitimation. Palgrave Commun 6:1–5

Sears (2018) The role of artificial intelligence in the classroom. https://elearningindustry.com/artificial-intelligence-in-the-classroom-role

Sennaar K (2019) AI in agriculture-present applications and impact. https://emerj.com/ai-sector-overviews/ai-agriculture-present-applications-impact/

Van Der Sluijs JP et al. (2005) Combining quantitative and qualitative measures of uncertainty in model-based environmental assessment: The NUSAP system. Risk Anal 25:481–492

Smith A (2018) Franken-algorithms: the deadly consequences of unpredictable code. The Guardian. https://www.theguardian.com/technology/2018/aug/29/coding-algorithms-frankenalgos-program-danger

Sonnenburg S et al. (2007) The need for open source software in machine learning. J Mach Learn Res 8:2443–2466

Supiot A (2017) Governance by numbers: the making of a legal model of allegiance. Hart Publishing, Oxford; Portland, Oregon

Taleb NN (2007) The Black Swan: the impact of the highly improbable. Random House Publishing Group, New York, NY

Thimbleby H (2003) Explaining code for publication. Softw: Pract Experience 33:975–1001

Wallach W, Allen C (2008) Moral machines: teaching robots right from wrong. Oxford University Press, Oxford, USA

Watson D, Floridi L (2019) The explanation game: A formal framework for interpretable machine learning. https://papers.ssrn.com/abstract=3509737

Wiener N (1988) The human use of human beings: cybernetics and society. Da Capo Press, New York, N.Y, new edition

Wong YH et al. (2020). Deterrence in the age of thinking machines: product page. RAND Corporation. https://www.rand.org/pubs/research_reports/RR2797.html

Ye H et al. (2018) Machine learning for vehicular networks: recent advances and application examples. IEEE Vehicular Technol Mag 13:94–101

Yu H et al. (2018) Building ethics into artificial intelligence. http://arxiv.org/abs/1812.02953

Yurtsever E, Capito L, Redmill K, Ozguner U (2020) Integrating deep reinforcement learning with model-based path planners for automated driving. http://arxiv.org/abs/2002.00434

Download references

Acknowledgements

I would like to thank Kjetil Rommetveit, Andrea Saltelli and Siddarth Sareen for the organisation of the Workshop Ethics of Quantification , and the Centre for the Study of Sciences and the Humanities of the University of Bergen for the travel grant, at which a previous version of this paper was presented. Thomas Hodgson, Jill Walter Rettberg, Elizabeth Chatterjee, Ragnar Fjelland and Marta Kuc-Czarnecka for their useful comments in this venue. Finally, Stefn Thor Smith and Andrea Saltelli for their suggestions and constructive criticism on a draft version of the present manuscript.

Author information

Authors and affiliations.

School of the Built Environment, University of Reading, Reading, UK

Samuele Lo Piano

Open Evidence, Universitat Oberta de Catalunya, Barcelona, Catalonia, Spain

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Samuele Lo Piano .

Ethics declarations

Competing interests.

The author declares no competing interests.

Additional information

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

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 license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license 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 license, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Cite this article.

Lo Piano, S. Ethical principles in machine learning and artificial intelligence: cases from the field and possible ways forward. Humanit Soc Sci Commun 7 , 9 (2020). https://doi.org/10.1057/s41599-020-0501-9

Download citation

Received : 29 January 2020

Accepted : 12 May 2020

Published : 17 June 2020

DOI : https://doi.org/10.1057/s41599-020-0501-9

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

This article is cited by

A survey on interpretable reinforcement learning.

• Claire Glanois

Machine Learning (2024)

Reframing data ethics in research methods education: a pathway to critical data literacy

• Javiera Atenas
• Leo Havemann
• Cristian Timmermann

International Journal of Educational Technology in Higher Education (2023)

AI ethics: from principles to practice

• Jianlong Zhou

AI & SOCIETY (2023)

The Challenge of Quantification: An Interdisciplinary Reading

• Monica Di Fiore
• Marta Kuc-Czarnecka
• Andrea Saltelli

Minerva (2023)

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies