Is AI in sport a problem – or is it the humans who use it?

Like every sector AI is becoming more prevalent in sport, but it’s vital we retain human checks, balances and oversight to avoid relinquishing all decisions to chatbots which may not get the game as David Granger explains.

For most of us, AI is a decent research assistant, an occasionally amusing meme-maker or an interactive search engine. For one football coach it was the reason given for him getting the sack. If you ask ChatGPT to handle your teams away game travel arrangements, you need to double check things like sleep times and training session schedules. 

It may be apocryphal, but the story of Spaniard Robert Moreno the (now former) FC Sochi manager became a cautionary tale for AI users beyond the world of soccer and sport. 

According to Sochi’s (also) former sporting director, Andrei Orlov, the Russian club’s now ex-coach Moreno would chat with ChatGPT (the large language model used for everything from planning family holidays to helping students with their homework) to take care of the seemingly mundane, straight forward footballing decisions. Mundane decisions such as the week’s training regimes or, using player data from Wyscout, formulating transfer strategies and making the call on which striker to buy. The results, according to Orlov, were players being told to remain awake for 28 hours at a time and told told to attend a training session at 7am two days before the match.

The problem was not the reliance on artificial intelligence per se, it was the abdication of logic and critical thinking. The problem was relinquishing rationale. Like those stories of lorry drivers getting stuck in fords for relying on the sat-nav and not the flood warning signs, this would become AI legend. 

The trouble is when this is done without context or oversight. Relying solely on a machine means failing to take into account such trivialities as Russian time zones, squad dynamics, or what a footballer’s body needs at altitude. If you do not check and abdicate all rational thought in favour of his chatbot. 

It’s worth emphasising that Moreno vehemently denied this story. In a Spanish newspaper, he said: “I have never used ChatGPT (or any AI) to prepare for matches, decide lineups, or choose players. Like any professional coaching staff, we use analysis tools (video, data, scouting) to organize information, but the sporting and human decisions are always made by the coach and his staff.”

So it may be that the club was looking for a scapegoat and found a premium subscription to blame.

What Moreno says is correct – artificial intelligence is being used more and more in sport – an industry in which data and data analysis is crucial in every sector from player rankings and value to odds for betting companies.

Take Women’s Rugby where AI is being deployed with purpose and a healthy dose of human oversight. 

The Australian Women’s Rugby Sevens used VueMotion, an AI-powered movement analysis tool, with cameras capturing athletes’ movements during training and giving instant data feedback to allow coaches and players to correct form, turning athletic effort into measurable, actionable performance insights.

The key here is that AI informs the coach, it doesn’t replace them. The technology knows what it’s good at: pattern recognition at scale, while we humans know what we’re good at which is the motivation of players and forging relationships. In the Women’s Rugby World Cup AI enhances decision-making under match pressure without removing the human in the loop.

According to WSC Sports, three out of four professional teams now rely on real-time analytics for performance and strategy (something we’ve done in F1 since, well… since the first race) while 89% of sports executives expect AI to significantly impact their business operations within the next three years.

From Red Bull Racing’s partnership with Oracle to implement AI for race strategy simulations to FIFA’s semi-automated offside review, AI excels at narrow, well-defined tasks because its boundaries have been defined and are monitored by humans.

The cautionary tale is not that AI is involved, but when it has no defined remit or checks. For Women’s Rugby, AI is succeeding as it as a clear, purposeful context and task. AI delivers great insight and data analysis, but does not (yet?) have the emotional intelligence and leadership which human coaches possess and which great teams deserve.

David Granger is former global editor-in-chief of redbull.com and director of Arc & Foundry, a UK-based content marketing agency.