BGU Study Warns AI Sports Scouting May Reinforce Bias, Threaten Privacy

Source Domain: themedialine.org

Researchers at Ben-Gurion University of the Negev have identified hidden risks associated with the growing use of artificial intelligence in sports scouting and youth talent identification, warning that some AI systems may reinforce existing inequalities and create new ethical challenges, according to a study published in Big Data and Cognitive Computing.

The study examined AI tools used to evaluate athletic performance and identify promising young athletes. These systems increasingly rely on large datasets, machine learning algorithms, video analysis, and other digital assessment methods to support recruitment and talent selection decisions across the sports industry.

Researchers found that algorithms trained on historical data can reproduce social and economic biases already present in existing datasets. According to the study, AI systems may use indirect indicators, such as place of residence, school background, and other socioeconomic factors, as proxies when evaluating athletes. As a result, opportunities for young players could be influenced by factors unrelated to athletic ability.

The research also highlighted concerns about what it described as “early determinism,” in which AI-driven profiling may label children at a young age and influence future opportunities. Researchers warned that such systems could make it more difficult for late-developing athletes to gain recognition if early assessments become overly influential in talent identification programs.

Privacy concerns were another focus of the study. Researchers said the growing use of large-scale datasets, including information that may extend to social media activity, raises questions about the long-term handling of personal data and the potential use of childhood information outside the sports sector.

The study further noted that AI systems often depend on historical datasets that may contain existing imbalances, potentially amplifying inequalities while failing to account…

Source