With artificial intelligence creeping into more and more areas of our sport, James Shrubsall sets out to find out how it’s changing cycling and what we can expect in the future
As you may or may not have noticed, artificial intelligence has already infiltrated cycling. It might feel like barely a day goes by without seeing some eye-opening AI related headline, but fear not, as machines become more involved in decision making, this isn't some dystopian Skynet, doomsday scenario.
AI has its proponents and its detractors – in a similar way that the internet, television and even supermarkets have done before it. But whether you can’t wait for it to become even more powerful or the whole thing makes you a little nervous, AI has well and truly arrived. In fact it may already be making your life in cycling better – or at least different – without you realising.
From training apps to city transport management to WorldTour team nutrition and more, AI and machine learning is reaching its digital tentacles into all corners of cycling, whether for sport, leisure or transport.
One use of AI that has been in the headlines recently is in sports nutrition. It is already used in consumer apps such as CalorieMama, whose photo recognition software uses AI to determine the calories in your meal; Nutrition. ai, which claims it will answer any questions you have on the subject – with the help of AI, of course; and Hexis, which might be most pertinent to readers of this magazine thanks to its ability to calculate your nutrition requirements based on the demands of your riding.
However, one of the highest profile uses of AI for cycling nutrition is the Jumbo-Visma WorldTour team’s use of it at last year’s Tour de France, where its rider Jonas Vingegaard triumphed ahead of Tadej Pogačar.
The Dutch team collaborated with Dr Kristian van Kuijk and colleagues at Maastricht University, who helped the riders plan out their diets with the aid of machine learning.
The researchers looked at how much energy riders had burned in previous races, feeding a huge amount of information into their algorithms, including details about the course and the weather, as well as rider data such as power and vital stats.
Machine learning was able to take the information and use it to predict how much food any particular rider would need to eat for the various stages of the Tour de France.
To test its efficacy, the researchers tested their model against Jumbo’s coaches. All were asked to estimate nutritional requirements from stages in the Tour de France and Giro d’Italia from 2019. The test yielded a score between zero to one, with higher scores being more accurate – the coaches
