Figure skating turns to AI and tech to tackle judging controversies

MILAN – After decades of judging controversies, figure skating is turning to artificial intelligence (AI) and computer vision technology to try to bring greater consistency and transparency to how performances are scored.

The International Skating Union (ISU) has been testing high-resolution camera systems at competitions over the past two years that use AI to track skaters’ movements and analyse technical elements such as jump rotation, height, distance travelled and spin positions in real time.

The Computer Vision technology, which uses six cameras placed around the rink, aims to provide more objective data to support technical panels who currently make split-second judgments about whether jumps are properly rotated or which edge a skater took off from – decisions that can determine medals.

ISU director-general Colin Smith said the goal was first to use the data to support judges in awarding technical scores and then potentially integrate it into the actual scoring system.

The ISU plans to develop the system for singles competition and then for pairs and ice dance, the discipline most affected by judging disputes.

“We always look to see where we can include technology and where we can evolve the sport to ensure fair competition on the ice,” he told Reuters in an interview.

If effective, judges will be able to focus “on the artistry, on the human element, and the computer vision is looking more at the technical, the cut-and-dry aspects”, he said.

Omega, the official timekeeper of the Milano Cortina Olympics, has been using computer vision to get greater performance insight, but the company said it could see the data it captures becoming an additional tool for judges in future.

One such piece of data comes from a new system that can detect the angle of each skater’s blades, providing judges with additional data on how well certain routines and jumps have been executed.

The ISU also plans on using data analytics to evaluate judge performance. It has analysed over 750,000 element marks and almost 270,000 component marks from 78 international events to identify consistency issues.

“Judging’s a real skill,” Smith said. “We want to be able to identify who are your best officials and then further grow and develop them and then look at those officials who are maybe at the lower end and help them develop their skills.

“What we’re looking for is the consistency of judging across the different events.”

HAUNTED BY SCANDAL

The current scoring system was introduced after the 2002 Salt Lake City judging scandal, where duplicate gold medals were awarded after allegations of judge bias.

But controversies have persisted. At the Grand Prix Finlandia Trophy in November, experienced ice dance teams complained about receiving junior-level scores on some elements. Smith said the issue had been addressed with technical experts.

Paolo Pizzocari, an Italian federation council member and former international judge, said judging errors in figure skating are often the result of how the sport’s complex rules collide with human cognitive limits, rather than simple incompetence or corruption.

He argued that even experienced, well-intentioned officials are prone to mistakes because they must assess dozens of technical and artistic criteria in just seconds.

“Our brain is not predisposed for this kind of work,” Pizzocari said. “We judge by intuition, not by processing a thousand parameters at once.”

AI could help provide objective measurements of rotations, edges and positions, reducing cognitive overload and bias while allowing judges to focus on fewer elements and preserve the sport’s artistic dimension, he added.

Skaters said they would welcome the technological help.

“Our sport is getting more technical and more complicated and intricate and precise,” said French ice dancer Guillaume Cizeron. “If there are tools that can help us make it as fair as possible, then why not?”

Skaters and their coaches have long used AI-powered tools such as Dartfish to analyse vast amounts of video material quickly and troubleshoot jumps and improve techniques.

Beyond judging, the computer vision data will also enhance fan engagement through broadcast graphics showing jump heights, spin speeds and other metrics, ISU’s Smith said.

“There's a huge amount of information which we’re not yet giving to the public,” he said. “This will allow us to pull out a lot more of that storytelling.”

The ISU has not set a timeline for when the technology might be used in official competition judging, saying it will review progress after the World Championships in Prague in March.

“We want these things as soon as possible, but we also need them to be accurate and foolproof,” he said. REUTERS