As the International Olympic Committee (IOC) introduces AI-powered judging, this technology guarantees greater consistency and improved transparency. But research suggests that trust, legitimacy and cultural values may be just as necessary as technical accuracy.
The Olympic AI agenda
In 2024, the IOC presented its Olympic AI agendaand positions artificial intelligence as a central pillar of future Olympic Games. This vision was affirmed from the start Olympic AI Forumheld November 2025where athletes, federations, technology partners and policymakers discussed how AI could support assessment, athlete preparation and the fan experience.
On 2026 Winter Olympics in Milan-CortinaThe The IOC is considering using AI to help in judging in figure skating (men and girls in singles and pairs) and helps judges accurately determine the variety of spins performed during a jump. The deployment may even extend to disciplines corresponding to big air, halfpipe and ski jumping (ski and snowboard events during which athletes mix jumps and aerial tricks), where automated systems could measure jump height and take-off angle. As these systems move from experimentation to operational use, it becomes necessary to look at what is likely to be going right or what is likely to be going incorrect.
Sports and human error are judged
In Olympic sports corresponding to gymnastics and figure skating, which depend on panels of human judges, AI is increasingly being presented as a kind of sport by international federations and sports federations Solution to problems with bias, inconsistency and lack of transparency. Judges must evaluate complex movements performed in split seconds, often from limited angles, over several hours in a row. Post-competition reviews show that unintentional errors and disagreements between judges will not be exceptional.
This was felt again in 2024 when there was an error in judgment involving a US gymnast Jordan Chiles on the Olympic Games in Paris caused great controversy. In the ground final, Chiles initially received a rating that earned her fourth place. Her coach then filed an investigation, arguing that a technical element had not been properly accounted for in the issue rating. After review, her rating was increased by 0.1 points, temporarily placing her within the bronze medal spot. However, the Romanian delegation objected to the choice, arguing that the US investigation was filed too late and the one-minute window was exceeded by 4 seconds. The episode highlighted how complex the principles are, how difficult it could be for the general public to follow the logic of judging decisions, and the way fragile trust in human panels of judges is.
In addition, cheating was also observed: many still remember the rankings in figure skating scandal on the 2002 Winter Olympics in Salt Lake City. After the pairs event, allegations emerged that a judge had favored a duo in return for promised support in one other competition – exposing vote-trading practices throughout the jury. It is precisely in response to such incidents that AI systems have been developed, particularly by Fujitsu in collaboration with the International Gymnastics Federation.
What AI can (and can’t) fix in assessment.
Our research Research into AI-based scoring in artistic gymnastics shows that it's not nearly whether algorithms are more accurate than humans. Errors in judgment are sometimes on account of the bounds of human perception and the speed and complexity of excellence – which makes AI attractive. However, our study of judges, gymnasts, coaches, federations, technology providers and fans highlights quite a lot of tensions.
AI may be overly precise, evaluating routines with a precision beyond what human bodies can realistically perform. For example, if a human judge visually assesses whether a position is being held accurately, an AI system can detect that the angle of a leg or arm is just just a few degrees off the best position and penalize an athlete for an imperfection invisible to the naked eye.
While AI is usually portrayed as objective, the design and implementation of those systems can create recent biases. For example, an algorithm trained totally on male performances or dominant styles may inadvertently penalize certain body types.
In addition, the AI has difficulty accounting for artistic expression and emotions – elements considered central in sports corresponding to gymnastics and gymnastics figure skating. Finally, while AI guarantees greater consistency, maintaining it requires ongoing human oversight to adapt rules and systems as disciplines evolve.
Action sports follow a distinct logic
Our research shows that these concerns are much more pronounced in motion sports like snowboarding and freestyle skiing. Many of those disciplines were included within the Olympic program to modernize the games and attract a younger audience. Still Researcher warn that Olympic inclusion may speed up commercialization and standardization, on the expense of the creativity and identity of those sports.
A pivotal moment dates back to 2006, when the U.S Snowboarder Lindsey Jacobellis lost Olympic gold after performing an acrobatic move during a jump – she grabbed her board within the air – while leading the race Snowboardcross finals. The gesture, celebrated inside their sporting culture, ultimately costing them the gold medal on the Olympics. The episode highlights the strain between motion sports' expressive ethos and institutionalized evaluation.
AI special stages on the X Games
AI-powered assessment adds recent layers to this excitement. Previous research about halfpipe snowboarding had already shown how evaluation criteria can subtly change performance style over time. Unlike other rated sports, motion sports place particular emphasis on style, flow and risk-taking – elements which might be particularly difficult to formalize algorithmically.
AI has already been tested X Games 2025particularly in the course of the snowboard SuperPipe competitions – a bigger version of the halfpipe with higher partitions that allow for larger and more technical jumps. Video cameras tracked each athlete's movements while AI analyzed the footage to create an independent performance assessment. This system was tested alongside human judging, with judges continuing to award official results and medals. However, the trial had no impact on the official results, and no public comparison has been published on the extent to which the AI results agree with those of human judges.
Nonetheless, Reactions There were major differences of opinion: some welcomed more consistency and transparency, while others warned that AI systems wouldn't know what to do when an athlete introduces a brand new trick – something often greatly appreciated by human judges and the audience.
Not to be judged: training, performance and the fan experience
The impact of AI goes far beyond self-assessment. In TrainingMovement tracking and performance evaluation are increasingly shaping technology development and injury prevention and influencing the way in which athletes prepare for competition. At the identical time, AI is changing that Fan experience through improved repetitions, biomechanical overlays and real-time explanations of performances. These tools promise more transparency, but in addition frame the way in which performance is known – and thus contribute much more “Storytelling” “ about what may be measured, visualized and compared.
At what price?
The Olympic AI agenda The aim is to make sport fairer, more transparent and more appealing. But as AI is integrated into assessment, training and the fan experience, it also plays a quiet but necessary role in defining what counts as excellence. If elite judges are regularly replaced or sidelined, it could have a negative impact – changing the way in which lower-level judges are trained, how athletes develop and the way the game evolves over time. The challenge for Olympic sport is due to this fact not only technological; It is institutional and cultural: How can we prevent AI from eroding the values that give every sport its meaning?
