Is AI the unbiased tool that will solve problems of subjective sport officiating?
Katherine A O’Brien
Senior Lecturer – Queensland University of Technology (Brisbane/Australia)
Artificially Intelligent (AI) tools are often touted for their ability to automate repetitive and mundane tasks, freeing up human workers to focus on more complex and creative endeavours. For example, Forbes Magazine contributor Jia Rizvi (https://www.forbes.com/sites/jiawertz/2024/01/16/4-ways-artificial-intelligence-is-making-companies-more-efficient/?sh=520f3a99545a) suggests that AI can improve efficiency and productivity in various industries by handling routine tasks more quickly and accurately than humans. In e-commerce, entertainment, and healthcare, Gustavo Bodra (https://medium.com/ai-revolution-transforming-the-way-we-live-and/the-role-of-artificial-intelligence-in-industries-such-as-healthcare-finance-and-transportation-e81633185a6a) highlights how AI can enable greater customization and tailoring of products and services to individual needs and preferences. AI is also claimed to possess the potential to revolutionise diagnosis, treatment, and patient care (care https://www.thoughtful.ai/blog/revolutionizing-patient-care-with-ai-driven-diagnostics). In the business domain, the IBM Data and AI Team (https://www.ibm.com/blog/take-advantage-of-ai-and-use-it-to-make-your-business-better/) note that AI tools are increasingly being used to forecast future trends for optimising investment strategies or predicting which products or services customers are more likely to be interested in. At the same time, AI is also transforming the sports sector by enhancing performance, safety, strategic planning, and fan engagement. Performance analytics, player tracking, scouting and recruitment, and sports betting for instance all involve the use of AI for data analysis and tactical planning (https://imaginovation.net/blog/ai-in-sports-industry/). Thus, AI has infiltrated the everyday life of people, meaning that it is now often seen as the panacea or modern solution for helping overcome issues in a range of industries related to fairness, transparency, and adherence to ethical standards, rules, and regulations.
AI, for example, is often viewed as an application that can eliminate human subjectivity or biases in how sport officials perform their role and make decisions during games. Video Assistant Referees (VAR) is already a common feature of football, with FIFA creating a technology department to study the replacement of linesmen with robots, cameras and computers to try and avoid mistakes in ruling offsides and throw-ins (file:///C:/Users/obrienk8/Downloads/117-Article%20Text-230-1-10-20220215%20(1).pdf Sampedro, 2021). Hawk-Eye in tennis purportedly provides accurate and unbiased decisions by tracking the trajectory of the ball. In the future, AI technology could also be judging Olympic gymnastics, with AI being used to review routine scores during the 2023 World Gymnastics Championships (https://www.wcnc.com/article/news/local/connect-the-dots/ai-judge-gymnastics-in-2024-olympics/275-579de210-c83b-486e-9c61-f802b000085b). However, while AI can analyse vast amounts of performance data quickly to identify trends, patterns, and correlations that humans might miss, there are also some uniquely human aspects of sport officiating that AI tools might perceptively miss, such as:
- Complexity of Decision Making
Officiating decisions in sports often involve complex, nuanced judgments that may be difficult for AI systems to replicate or correctly interpret. Officials often need to interpret rules and regulations, assess player actions, and consider situational context in real-time, tasks that usually require human judgment and experience. AI algorithms may struggle to account for the diverse range of factors that influence these decisions, including player intent, timing, and spatial dynamics.
- Real-Time Judgements
Most officiating decisions need to be made quickly, sometimes in a matter of seconds, during fast-paced and dynamic sporting events. While AI systems can process large amounts of data and analyse visual patterns, they may not always be able to provide both timely and accurate assessments in real-time.
- Concerns about Bias and Fairness
There are concerns about the potential for bias and discrimination in AI-powered officiating systems, particularly if algorithms are trained on biased data or fail to account for diverse perspectives and contexts. Ensuring fairness and transparency in AI-driven officiating is essential for gaining trust and acceptance in the viewing population and sport organisations involved. There may also be logistical challenges, such as ensuring compatibility with existing officiating processes, regulations, and media applications.
Therefore, is AI the unbiased tool that will eliminate the vagaries of how sport judgements are made? Can AI replicate how sport officials use human judgment to interpret and apply the rules of the game to specific situations, including evaluating whether an action constitutes a foul or violation based on context and severity? Moreover, will sport officials become overly reliant on AI-based technologies, potentially undermining their authority and decision-making skills or is the viewing public already too accepting of decisions produced by Hawkeye, VAR, and the Bunker in Rugby League? For example, Hawkeye decisions in cricket often come with an “umpire’s call” component, which indicates that the ball-tracking system acknowledges a margin of error. Thus, when the predicted path of the ball is too close to call definitively, the on-field umpire’s original decision stands, which can cause confusion and dissatisfaction among players and fans (https://wisden.com/series-stories/india-v-england-2024/hawkeye-inventor-no-need-for-umpires-call-if-starting-from-scratch). Also, the integration of Hawk-Eye into the Decision Review System (DRS) sometimes leads to decision inconsistencies, with the same type of delivery being judged differently in different matches or even in different parts of the same match (https://wisden.com/series-stories/womens-premier-league-2024/chamari-athapaththu-lbw-leads-drs-debate-hawkeye-projection-accuracy). While VAR was introduced to enhance decision-making accuracy and fairness in football, it has encountered several significant issues such as the interpretation of handball incidents, where subjectivity plays a significant role (e.g., 2019 Champions League quarter-final between Manchester City and Tottenham Hotspur – https://www.abc.net.au/news/2019-04-18/spurs-bounce-man-city-champions-league-crazy-quarter-final-tie/11027874), and the awarding of penalties (e.g., 2020 Premier League season, Manchester United received numerous penalties after VAR reviews – https://www.espn.com.au/football/story/_/id/37587139/how-var-decisions-affected-every-premier-league-club-2020-21). The Rugby League Bunker has also faced significant criticism for disallowing tries based on ball placements measured in millimetres (https://wwos.nine.com.au/nrl/bunker-problems-need-to-be-looked-at-for-2024-season-mark-levy-column/ea71731c-5a55-4974-9075-be689fd6efb2), leading to debates in the media over the precision of technology and whether such fine margins should determine essentially human endeavours. In other words, do AI-driven decision systems automatically “autotune” sport by levelling out the excitement and enjoyment of watching games? Is the pressure of accurate decision-making turning sport officials into less-confident adjudicators who second-guess themselves? Put simply, large volumes of accurate data can clearly help AI-models with identifying and mitigating biases; but, at the same time, inaccurate or biased data, collected over several sporting seasons, can also potentially reinforce existing biases or introduce new ones, causing AI-driven sport technology tools to produce unfair or even incorrect decision outcomes.
Therefore, perhaps at this point in time, AI is not the panacea or unbiased technology tool that solves problems in sport officiating, suggesting that the oft-repeated question of human judgements versus technology needs more time to be answered. Also, given that most AI systems are viewed as “black boxes”, perhaps the process of incorporating AI into sport and sport officiating systems needs to be more transparent by adopting an open communication policy about its purpose and clear explanation of the effects it may have (file:///C:/Users/obrienk8/Downloads/117-Article%20Text-230-1-10-20220215%20(1).pdf Sampedro, 2021). Striving to create a balance between AI’s involvement and preserving the essence of human competition might seem like the perfect goal; however, as highlighted by recent lawsuits (https://www.forbes.com/sites/jemmagreen/2024/03/07/elon-musks-lawsuit-against-openai-sparks-debate-over-ai-transparency/?sh=408475c04332) and OpenAI sites (https://www.mathaware.org/ai-in-sports-upholding-fairness-transparency-and-accountability/), the rapid pace of AI development means that finding this equilibrium might be technically challenging for leveraging AI’s potential and maintaining the spirit of fair play. Moreover, while fans and the media continue to believe that non-human brains are somehow superior to human brains (https://www.forbes.com/sites/forbestechcouncil/2021/11/29/can-tech-really-better-human-decisions-in-sports/?sh=2dc26af4698d), and, hence, make better decisions than sport officials, further work is clearly required to uncover AI’s potential to address subjective decisions in the sport officiating domain.
