Researchers at the Massachusetts Institute of Technology’s Sports Lab are exploring how artificial intelligence can improve elite figure skating performance and deepen understanding of the sport’s technical and aesthetic demands. In an interview with MIT News, Jerry Lu, creator of an AI-powered optical tracking system called OOFSkate, describes how video analysis helps skaters refine jumps such as quadruple axels and potentially the five-rotation quint. The system uses computer vision to extract physical metrics from standard video and compare an athlete’s execution with data from elite performers. It also includes an automated classifier that estimates how international judges might score a jump, helping athletes identify areas for precise improvement.
Lu, a former MIT Sports Lab researcher now working with NBC Sports during the 2026 Winter Olympics, plans to use these AI insights in broadcast commentary for figure skating, snowboarding, and skiing. His goal is to make the complexity of scoring and the difficulty of technical maneuvers more accessible to audiences.
Professor Anette “Peko” Hosoi, co-founder and faculty director of the Sports Lab, is leading related research into how AI systems might assess aesthetic elements of performance. Unlike the strictly defined technical elements of a jump, artistic impression in figure skating is subjective and harder for machines to gauge. Hosoi and collaborators are investigating whether AI evaluations reflect human aesthetic reasoning or merely mimic patterns, using skating as a controlled environment where numerical scores anchor subjective judgment.
On the possibilities for future achievements, both researchers see milestones ahead. Lu believes that data-driven narratives at the Olympics can highlight just how hard figure skating is, while Hosoi says a quintuple jump is within reach in the near term, though a six-rotation jump likely exceeds human physical limits.