How AI and 'Digital Twins' Are Rewriting the Rules of Olympic Judging

For over a century, the Olympic Games have relied on the naked eye to measure the limits of human potential. In sports like figure skating and gymnastics, this reliance has birthed legendary performances, but also bitter controversies. Fractions of a point, decided by subjective human panels, can separate lifelong glory from devastating heartbreak.[7]

But as the world turns its attention to the 2026 Milano-Cortina Winter Olympics, a quiet revolution is taking center stage. Dubbed by technologists as the "First Intelligent Olympics," the upcoming Games will feature unprecedented integration of artificial intelligence and computer vision in the judging process.[1][3]

The shift is not about replacing human judges, but rather giving them a high-tech audit tool to manage cognitive overload. In figure skating, judges have mere seconds to determine if a skater fully rotated a triple axel or launched from the correct edge of a four-millimeter blade while traveling at twenty miles per hour.[1][6]

To solve this, the International Skating Union (ISU) and official Olympic timekeeper Omega are deploying advanced AI systems that track a skater's "digital twin." Using an array of high-definition cameras capturing thousands of images per second, the system builds a real-time 3D skeletal model of the athlete on the ice.[1][3]

This computer vision technology can instantly detect the exact angle of a competitor's blade and calculate metrics like jump height, airtime, and rotation with millimeter precision. If a skater is a quarter-rotation short on a quadruple jump, the system flags it immediately, eliminating the need for lengthy, inconclusive slow-motion video reviews.[1][5][6]

The foundation for this winter sports revolution was actually laid on the gymnastics mat. Since 2019, the International Gymnastics Federation (FIG) has partnered with Japanese technology giant Fujitsu to develop the Judging Support System (JSS), which is now capable of scoring all ten gymnastics apparatuses.[2][4][8]

Fujitsu's system originally utilized solid-state Lidar sensors emitting over two million pulses per second to map a gymnast's movements. It has since evolved into an advanced camera-based AI solution that uses proprietary correction algorithms to reduce "jitter"—the estimation error in posture recognition that traditionally plagued deep learning models.[4][8]

By comparing the athlete's real-time skeletal data against a vast database of defined skills, the JSS can recognize thousands of different movements with remarkable accuracy. It provides objective biomechanical data to support the "D-Score" (Difficulty), ensuring that a gymnast is properly credited for the exact skills they perform.[4][8]

Beyond the judges' table, this technology is fundamentally altering how elite athletes train. The Japan Skating Federation recently adopted Fujitsu's skeleton recognition AI to analyze jump performances during national training camps, replacing cumbersome marker-based motion capture suits with markerless video analysis.[3][4]

Athletes and coaches now receive instant, data-driven feedback. New AI-powered applications allow skaters to record their practice sessions on a smartphone and immediately see their jump height and rotation speed. This rapid feedback loop is helping athletes push the boundaries of the sport, with some experts predicting that AI optimization could soon lead to the world's first successfully landed quintuple jump.[3][6]

Furthermore, machine learning is being deployed as a preventative medical tool. By analyzing a skater's movement history and detecting microscopic changes in their form, AI systems can predict fatigue and flag potential injury risks—like stress fractures—before the athlete even feels them.[4][6]

Despite the technological leaps, the integration of AI into subjective sports is not without its skeptics. The primary concern revolves around the "soul" of the sport: while an algorithm can count rotations and measure blade angles, it cannot feel the emotional resonance of a performance or evaluate its creative brilliance.[5][7]

Governing bodies have addressed this by strictly dividing the scoring criteria. In figure skating, the AI assists exclusively with the Technical Element Score (TES). The Program Component Score (PCS)—which evaluates composition, presentation, and skating skills—remains entirely in the hands of human judges.[1][7]

There are also valid concerns regarding algorithmic bias. If the AI models are trained on datasets that lack diversity in body types, heights, or skin tones, the system could inadvertently penalize athletes who do not fit the historical mold. Ensuring the training data is comprehensive and representative is a massive ongoing priority for developers.[6][7]

Ultimately, the goal of AI in Olympic judging is not to achieve algorithmic perfection, but to restore trust. By handling the objective physics of a routine, the technology frees human judges to do what they do best: appreciate the artistry, evaluate the expression, and celebrate the human spirit that makes the Games so captivating.[5][6][7]