"I Saw the 0.43-Point Difference" — The Milan Olympics Figure Skating Judging Scandal: Can AI Guarantee 'Fairness'?
Summary
Analyzing the possibilities and limitations of AI judging in figure skating, sparked by the 2026 Milan Winter Olympics ice dance scoring controversy. Examines Omega's 14-camera 8K AI system, ISU Vision 2030's hybrid judging model, and the dilemma between technology and artistry from an AI perspective.
Key Points
The Injustice of 0.43 Points
A single French judge's approximately 8-point bias toward her own country's team determined the gold medal. Despite 5 of 9 judges scoring the American team higher, the final result was reversed.
The Potential of 14 8K Cameras
Omega's AI-powered computer vision system can analyze jump height, rotation count, and landing angle down to the millimeter. However, this data is currently only provided for broadcasts, not shared with judges.
The Hybrid Judging Model Is the Answer
A division of labor where AI handles technical scores (rotation count, edge calls) and humans evaluate artistic scores (musical harmony, emotional expression) is the most rational solution. ISU Vision 2030 is pursuing this direction.
Guarding Against New AI Biases
AI could systematically disadvantage certain body types or styles due to training data bias, and there are concerns about excessive precision — penalizing deviations invisible to the human eye.
Transparency Is the Core Value
The true value of AI judging lies not in correct verdicts but in transparent ones. Simply detecting and publicizing statistical outliers in real time would dramatically reduce bias.
Positive & Negative Analysis
Positive Aspects
- Objective Technical Judging
14 8K cameras and AI can precisely measure jump rotations, landing edges, and hang times down to the millimeter, eliminating human judges' technical scoring biases.
- Enhanced Transparency
AI can detect statistical outliers in judges' scores in real time and make them public, automatically preventing biases like the French judge's 8-point discrepancy.
- Proven Success in Gymnastics
Fujitsu's JSS successfully evaluates 2,000 gymnastic movements at 90% accuracy, proving the technical feasibility of applying similar systems to figure skating.
Concerns
- Fundamental Limits of Artistic Judging
Artistic scores — harmony with music, emotional expression, choreographic creativity — are inherently subjective, and quantifying 'being moved' remains impossible with current technology.
- Concerns About New Forms of Bias
If an AI system is trained on data skewed toward certain body types or styles, it could systematically disadvantage diverse athletes, creating new forms of bias.
- The Excessive Precision Problem
There is debate about whether penalizing athletes for angle deviations of a few degrees — invisible to the naked eye — aligns with the spirit of sport.
Outlook
ISU Vision 2030 is pushing to introduce a hybrid model of AI technical scores + human artistic scores from the 2026-27 season. Omega's AI camera data may be incorporated into the judging process at ISU-sanctioned events later this season, marking the first structural step toward resolving 12 years of recurring judging controversies since Sochi.