Don’t believe the hype.

Why AI training programmes probably didn’t help anyone win gold in Paris ‘24.

‍

The Paris Olympic games 2024 has been heralded as the most technologically advanced games ever, with AI playing a huge part, from timekeeping to performance mapping, energy usage to tv coverage, to the athletes pre-olympic training.

It’s generated a tonne of headlines and some great PR for the games, which (I’m guessing) was entirely the point. But as one of our missions at VOR is to cut through the constant AI overhype, I feel like we need to take a minute to look at what’s really going on behind those headlines.

Is the buzz around AI use at Paris 2024 a true reflection of the way AI will be integrated into the future of elite sport? Not at all.

Here’s why.
‍

Don’t believe everything you read

Obviously the use of new technologies like AI at a huge global event like the Olympic games is exciting. The stakes are high and everyone is watching. But making it the headline is problematic in a couple of ways.

I can’t (and wouldn’t attempt to) comment on the use of AI for timekeeping or media coverage, but when it comes to AI use in the athlete’s training plans, the over-hype seems to be real.

The headlines are misleading and they encourage people to use AI in ways that are potentially unsafe. Jess, our CTO puts it perfectly:

“I would stress that people should use caution when reading about any technology that is making headlines. The hype bubble always ends up bursting. AI is an effective tool for very specific problems. The media hype means we might start using AI for the sake of using it, rather than for the things it is good at.”
‍

Using AI for the sake of it is, at best, a waste of people’s time and money. But it can also be dangerous - messing up finely tuned training routines and potentially causing injury.

Also, the hype around AI and tech is taking the credit away from the athletes and their coaches. Olympic athletes have to sacrifice a huge amount of time, energy and money to even qualify for the Olympics, let alone get anywhere near the medal podium.

In fact, I saw figures recently that showed we have only progressed finish times for the mens’ 100m by 0.81 seconds in the last one hundred years. Now I know, 0.81 seconds is a massive figure in terms of the 100m, but you’d think with all the technological, nutritional and kit advances we’ve seen since 1924 that the difference would be greater.

It goes to show that winning gold is still mostly about the grit, determination and physiology of the athletes. So credit where credit is due, right?

What do athletes and their coaches think of AI? And how are they really using it in their training? We asked elite level coach Kieran Leggett to give us his run down.
‍

How are athletes really using AI?

Kieran is a health and fitness coach who trains some of Great Britain and Ireland’s top athletes including international Marathon runner, Helen Davis and hopeful Olympic Irish gymnast, Halle Hilton. He’s been around for a long time, and seen the way technology has (and hasn’t) impacted elite training for athletes and more regular gym goers.

He says: “It's tough to tangibly measure the AI input on an athlete’s final performance (at Paris 2024) as it just hasn't had the time to be utilised in this environment and with teams/coaches willing to invest in this area.”
‍

“I haven't seen enough measurable evidence (for the effectiveness of AI on athletes performance at Paris 2024) although I believe this will change quickly with the more open minded coaches and it will be more common at the next Olympics with science driven data to prove its benefits in performance.”

Kieran told us that right now AI is being used by athletes, but for movement testing: “You would essentially have the AI interpret and measure movement testing and scores, and produce an instant report based on its findings.”

That’s great, and you can see how useful it can be to athletes that need to get into the detail of their performance, but it’s not quite what the news outlets are suggesting in their headlines.

You get out what you put in

For Kieran, the effectiveness of AI in elite sport and to the general public, is all about the quality of input: “My view is the coach still needs the qualifying input and can use the AI to cross reference, aid and improve training where human error can occur.”

“Ensuring the information input is high quality is key for a high quality output. It still requires a certain element of foundational knowledge to input the correct prompts and also for the user to have a basic movement quality to execute the exercises.”

Though he is optimistic that AI will play a role in the future of elite sport: “I think there can be automation in terms of delivery of workout such as the software to deliver the session, and reminders or motivational messaging which can be helpful for a coach without it taking more time to reach out to each individual.”
‍

“One thing all coaches must remember is AI can never replace the human interaction side of training. As a team, coach working with tech, I believe it's a really formidable partnership.”

The general message from Kieran seems to be, ‘we’re getting there, we can see the potential, but we’re not there yet.’ Which makes sense. But that’s not the story being shared with the public.

AI for AI’s sake

As Kieran explains, AI isn’t magic. It’s not going to instantly make you a better sprinter or swimmer.

Just using AI is not going to win team GB any medals, despite what the headlines kept suggesting. Like all new technologies, it’s not the technology itself that matters, it’s how you use it - and that’s the bit we’re still figuring out.
‍

Take the headlines around the use of AI in the gym equipment installed in the Olympic Village. Why? I totally understand how AI generated features in gym equipment can help the everyday gym goer - but not elite athletes at the end of an extremely specific and tightly regimented training camp. It just can’t be useful to an athlete using it for the first time in the final days and weeks before their event. It feels like AI for AI’s sake, there to make headlines rather than actually being useful for the athletes.

Then there’s the talk around athlete’s use of Large Language Models like ChatGPT. Some athletes might be using ChatGPT to help with workout plans, but that will be with the full support and knowledge of their coaching and support team. People who are trained to make sure any output from the AI is safe and inline with the athletes existing training programme.

Olympic athletes do not fly solo, they have a team of highly trained coaches, physios and support staff behind them.
‍

AI should be used to solve an existing problem

We believe that the best use cases for AI are to target specific pre-existing problems, in an area you and your team are experts in. You should also be able to understand how (and most importantly why) any automation is being used, so that you can trust that the processes you have created are applicable to the problem that you are solving.

Athletes already have everything they need to get to the podium - their support team: coaches and sports scientists with decades of experience. That might not always be the case, but so far we haven’t seen any athletes thanking AI programmes for their medals.
‍

It’s early days for AI in elite sports

At Tokyo 2020(1) few people knew what ChatGPT and Large Language Models were. The AI revolution was just beginning to appear at the edges of the zeitgeist. In the years since, the technology has exploded, but the way it’s adopted and used is still in its infancy.

There are already people talking about the use of AI training to encourage more people into fringe sports - people that are maybe restricted by money or location etc. A problem waiting for AI to solve it? Time will tell.  The real application of AI in elite sports likely won’t become clear until four years from now, when the hype-cycle starts for LA 2028.
‍

We can see the impact of AI in sport on a more everyday level

Those that will likely benefit first from advancements in AI are those without sponsorships and support teams - the everyday athletes, grassroots sportspeople, and anyone wanting to get fitter.

Kieran Leggett agrees: “The use of AI as a tool can be a really positive thing if utilised properly. One of the most obvious benefits is that the cost is extremely low. Another benefit is that AI can fill in your knowledge gaps in seconds so if a complete novice felt overwhelmed about going to the gym because they didn't know what to do, they could utilise AI to give them 5x beginner exercises that can help them achieve X result in X amount of time. So it's instant actionable information knowing it's moving them in the right direction towards their targets.”

This is the power AI can provide to the sports, fitness and wellness sectors now. It’s going to change the game, and maybe even be one of the first steps towards a technology that can ‘trickle up’ into elite sports too, but it’s unlikely to replace Olympic coaches anytime soon.