Griffin on Tech: AI’s mercifully light touch on the Olympics

This could have been the first AI-powered Olympics, with athletes consulting their robo-coaches and AI news readers presenting coverage of the games. 

Thankfully the technology, despite Intel’s big AI push as the official sponsor of the Games, has stayed largely on the sidelines in Paris. That’s likely to change by the time the Games hit Los Angeles in 2028.

The Olympics should, after all, be about unadulterated athletic endeavour, finding and rewarding the best in their fields. Technology certainly has a big role to play in improving athletic performance, and machine learning has for many years been used by sports teams to gain insights into performance using computer vision and data analytics. 

But the big improvements in Olympic performance have come from more mundane technological improvements. Take for example swimming, one of the flagship Olympic sports. In 1924, American swimmer Johnny Weissmuller set a new world record for swimming the 100-metre freestyle race in 59 seconds.

China’s Pan Zhanle set a new world record yesterday in the men’s 100m freestyle of 46.40 seconds, winning gold in the process. How are today’s swimmers able to shave 12 seconds off the record set 100 years ago?

Incremental gains

It really comes down to improvements in nutrition, training, and innovation in swimsuits, goggles and caps. A game changer was the switch from nylon to lycra materials in the 1970s, which meant swimsuits absorbed less water and created less drag on swimmers. High-tech swimsuits that cover the swimmer’s entire body are not allowed at the Olympics, which stipulates exactly the type of swimsuit male and female athletes can wear.

The other big factor in swimming is pool design. Swimming teams have complained that the La Défense Arena 50 metre pool at this year’s Olmypics is too slow. The pool was custom-built for the games but is 2 metres deep, much shallower than the 3m pools used in the Rio and Tokyo Olympics. Does a shallower pool make a difference?

It’s complicated, as the Guardian explains. Swimmers create turbulence in the water when they dive in and through the action of their swimming.  

“The slow pool theory says the shallower pool means more waves bounce back to the surface, creating more turbulence and slowing swimmers down,” the Guardian writes.

But the slow pool theory has some “scientific holes”. Anyway, everyone races in the same pool so any slow pool influence in Paris is shared by all competitors.

“It’s also possible swimmers are approaching the limits of human performance – at least until we work out how to break those limits once again,” the Guardian points out.

An AI divide?

That’s what I love about the Olympics - the relentless effort to squeeze out tiny margins in the pursuit of excellence. But it’s also a quest that has led to increasingly sophisticated attempts to use performance-enhancing drugs, and fraught arguments about gender in competitive sports.

AI tech is being used to power intelligent chatbots at the Olympics and to streamline broadcast coverage. AI-powered insights will no doubt help shave milliseconds off the 100m freestyle record in the coming years. 

But will it be available to all competitors or could AI create a divide between the rich and poor nations competing? Will the team with the best algorithms have the tiniest edge in future?  I hope not. That would make for a very dull Olympics.

Photo: Unsplash/Brian Matangelo