Women’s Soccer Coach Let ChatGPT Run Her Defense

Another day, another reminder that nearly everyone in every sector of the world is using generative AI in some capacity. Whether it’s chefs inventing recipes, teachers creating lesson plans, or young men turning to it for therapy and/or companionship, AI is bleeding into all corners of our lives. Women’s soccer is, unfortunately, officially no exception. 

National Women’s Soccer League coach Laura Harvey, who leads Seattle Reign FC, recently said she used ChatGPT to help craft a defensive strategy—and I hate to say it worked. 

“I never played a back-five until this year,” Harvey said on the Soccerish Podcast, hosted by Tampa Bay Sun FC president Christina Unkel and retired U.S. national team player Lori Lindsey, who’d asked her about her evolution as a coach and a player. 

The back-five is a soccer formation in which the team plays with five defenders, three midfielders, and two strikers (or a 3-5-2, a formation that allows the two outside halfbacks to drop back while on defense), which the Reign have played all season. “I’ve never played it, I never coached it…I’d never really done a lot of research on it,” Harvey continued. “I’d never really invested into how it could be played in the women’s game.”

So what made Harvey try a formation she had “only ever seen from afar?” The artificially intelligent chatbot that’s sucking up all our water and whose data centers are taking over rural neighborhoods, of course. 

During this past NWSL off-season, Harvey says she used the chatbot to ask, “What is Seattle Reign’s identity?” The sheer absurdity of a team’s own coach asking a chatbot that question aside, Harvey was somewhat unimpressed with its response, but says she kept asking it broad questions about winning in the league.

“Then I put in ‘what formation should you play to beat NWSL teams?’ And it spurred out every team in the league and what formation should you play,” she said. “And for two teams [Harvey refused to say which ones] it went ‘you should play back-five.’ So I did.”

To give her some credit, Harvey and the rest of the Seattle Reign FC coaching staff researched the formation and figured out how it would work for them. “Went for it, and we liked it. It worked,” she said. “We won the game.” 

Last year, the Reign finished 13th out of 14 teams. This year, they are currently in fifth place. 

Good for them, but if you’re the coach of a professional sports team, you’re in that position because you know the sport. Asking a chatbot to define your team’s identity seems woefully out of touch, and turning to a chatbot instead of your coaching staff seems dismissive of your team’s and staff’s abilities and of your confidence in them.

Because what many still don’t realize—or choose to ignore–is that generative AI programs are huge resource sinks with massive environmental consequences. The gargantuan data centers needed to power these programs (often housed in huge warehouses built in rural neighborhoods, like the one Elon Musk recently built in Memphis for his xAI supercomputer) suck up enormous amounts of power and require vast amounts of water to keep the processors cool. According to a study from the University of California, Riverside, writing a 100-word email using the GPT-4 (the latest model, GPT-5, launched in August) consumes an entire bottle of water and uses enough energy to charge an iPhone Pro Max to max battery seven times. 

“Training GPT-3 at Microsoft’s state-of-the-art US data centers can directly evaporate 700,000 liters of clean freshwater,” the study states. “More critically, the global AI demand is projected to account for 4.2 – 6.6 billion cubic meters of water withdrawal in 2027, which is more than the total annual water withdrawal of….half of the United Kingdom.”

In other “using ChatGPT to help me be a better coach news,” Ryan Huska, head coach of the Calgary Flames, appears to be using ChatGPT in his coaching efforts, and the Flames are ranked dead last in the Pacific Division. 

---

Like what you just read? You’ve got great taste. Subscribe to Jezebel, and for $5 a month or $50 a year, you’ll get access to a bunch of subscriber benefits, including getting to read the next article (and all the ones after that) ad-free. Plus, you’ll be supporting independent journalism—which, can you even imagine not supporting independent journalism in times like these? Yikes.