This MIT grad built an AI tool to show how hard Olympic figure skating actually is

MIT graduate Jerry Lu has developed an AI-powered tool called OOFSkate that quantitatively analyzes figure skating performances, highlighting the complex athleticism behind the sport at the Olympic level. This optical tracking system uses computer vision and AI to break down standard video footage of skaters’ jumps into detailed metrics such as jump height, rotation speed, airtime, and landing quality. By transforming subjective artistic performances into measurable, data-driven insights, OOFSkate aims to bridge the gap between the sport’s physical challenges and audience understanding.

Lu’s innovation allows coaches and athletes to pinpoint marginal gains, optimizing technical execution by comparing jumps to benchmarks established from elite skaters. The technology also helps commentators and viewers appreciate how difficult elements like quadruple Axels truly are, turning visually impressive yet complex maneuvers into clear, accessible information during broadcasts. This quantifying of skating performance addresses a long-standing lack of data-driven storytelling within artistic sports, similar to advancements seen in other televised athletic events.

In addition to technical analysis, MIT researchers are exploring whether AI can evaluate the artistic side of figure skating, assessing machine judgments in comparison with human perceptions of aesthetic quality. The development of OOFSkate reflects a broader movement to use AI to enhance both athlete performance and audience engagement in Olympic sports, marking a significant innovation in sports technology that blends athletic excellence with cutting-edge artificial intelligence.