With the 2026 FIFA World Cup due to kick off on Thursday, Israeli researchers are warning that a quiet technological shift is reshaping the future of soccer: artificial intelligence systems increasingly used to rank, analyze and predict children’s athletic potential.
An Israeli study examined how AI is moving beyond performance analysis toward shaping long-term judgments about human ability from early childhood. The findings, published in the peer-reviewed journal Big Data and Cognitive Computing, warn that as algorithms increasingly process youth performance data, they may not only evaluate talent more efficiently, but also begin to define it prematurely, potentially embedding bias and narrowing opportunity long before players reach professional levels.
Ahead of the tournament, FIFA has announced a partnership with technology company Lenovo to develop the Football AI Pro platform, an advanced system for tactical analysis and performance evaluation based on millions of data points, video analysis, 3D visualizations and machine learning simulations.
Similar developments are taking place in parallel initiatives such as Intel’s collaboration with the International Olympic Committee on AI-based talent identification systems. While some of these tools are already being tested or deployed in countries including Israel, Japan and El Salvador, others remain in early stages of implementation.
According to the study, led by Professor Ofer Ezer of Ben-Gurion University of the Negev and Dr. Ilya Morgolev of Beersheva’s Kay Academic College, the most significant shift is not simply improved analytics, but the emergence of continuous, data-driven tracking of youth athletes. Automated filming systems, wearable sensors and self-recorded training videos now allow children’s performance to be captured, stored and analyzed over time, creating a permanent digital record from an early age.
In Olympic sports, basketball, baseball and tennis, AI is increasingly being used to identify and evaluate athletic talent from a young age. Olympic development programs use biometric profiling and performance simulations to steer athletes toward suitable disciplines, while basketball and tennis systems rely on tracking data, video analysis and movement metrics to assess long-term potential. Baseball similarly applies data-driven models to evaluate prospects across its development pipelines, drawing on large historical datasets of player performance.
Across these sports, the common pattern is a shift from short-term evaluation of performance to long-term algorithmic prediction of potential based on continuously collected data from increasingly younger athletes.
Fighting digital determinism
The researchers describe this development as “digital determinism,” in which early performance data begins to shape and constrain future opportunities. Once stored, childhood metrics may influence which players receive coaching attention, scouting opportunities or funding, turning early measurements into long-term filters for athletic careers.
A key finding of the study is how bias can enter these systems indirectly. Even when algorithms exclude sensitive attributes such as ethnicity or income, they may still reproduce inequality through proxy variables such as school, geographic location or family structure. These signals, drawn from historical patterns of success, can embed socioeconomic disparities into systems that appear objective.
Scouting platforms already aggregate video footage, match events and performance statistics across multiple age groups and leagues. Machine learning systems then identify patterns associated with elite performance, while analysts apply filters such as age, position and competition level. However, because these systems are trained on historical “success,” they often inherit biases embedded in past coaching and selection decisions, the scientists said.
The study also highlights a reinforcing feedback loop: Athletes identified early by algorithms are more likely to receive superior training and exposure, which improves their performance data, which in turn strengthens the system’s original prediction. Over time, this dynamic may narrow rather than expand access to opportunity.
Beyond performance metrics, the researchers warn that data collection could eventually extend off the pitch. AI systems may incorporate social media activity, news coverage and other publicly available information to build broader profiles of young athletes, raising privacy concerns and questions about the long-term use of childhood data.
“Reality shows that competitive sports sanctify competitive achievement above all else, even at the expense of other values,” says Morgolev. “Competitive sports are a unique, extreme arena that relies on exceptional physical data alongside determination, resilience and motivation. A combination that, at best, characterizes only about 1% of the population.”
Ezer adds, “It is time to think together as a society about the place and degree of autonomy we are willing to give to algorithms in photographing, ranking, predicting and determining the development path of the future generation when competitive sports are irrelevant to the vast majority of them,” saying that children should develop under a supportive human framework rather than being prematurely defined by algorithmic scores.
