Artificial intelligence is revolutionizing sports analytics. It’s capable of sorting and analyzing data at a much faster pace than humans, including straightforward stats and more subtle details like athletes’ form or endurance.
Yet University of Wisconsin-Madison assistant statistics professor Sameer Deshpande told The Daily Cardinal the value of human decision-making in the world of sports analytics hasn’t decreased even as AI has stepped onto the field.
Deshpande, who’s been a part of UW-Madison’s Department of Statistics since 2021, has been researching sports statistics for nearly a decade, meaning he entered the field in close tandem with AI. He earned his PhD in statistics from the University of Pennsylvania, where his research included working in baseball, football and basketball.
He argues AI’s efficiency doesn’t remove humans from the sports analytics equation. In football, for example, Deshpande said human brains are needed to analyze and interpret data.
“A robot can measure how fast a wide receiver runs, but why does that matter? What can we do with that? Technology might help analysts, but it doesn’t know what to do with the information it finds,” Deshpande said.
He finds this to be especially true in the highest levels of sports because of the limited amount of information available.
Since AI must draw from what’s already online, it’s harder for it to be creative when looking at a more narrow problem within the world of sports. In comparison, broader fields of study often have more information available in databases. Because of this, human creativity remains the crucial crutch to make computer data useful.
“Stats is an art. We teach it to be mathematical, but it’s an art, learning what can be done. There aren’t always single right answers,” Deshpande said.
Deshpande’s perspective questions the legitimacy of an AI “revolution” and instead looks at it as simply an additional tool.
At the same time, he admits it’s hard to predict the future.
“If a coach was just asking ChatGPT what to do on fourth down, they would be run out of town. But at the same time, if there was a team out there using it to be successful, they also probably wouldn’t tell anyone,” Deshpande said.
It’s a show of how the cross section of statistics and sports is a cross section of uncertainties. Professional teams are already quick to protect their internal systems and strategies. When that culture is combined with a new technology, it leaves the definitive future impossible to predict.
