Computer science and now AI studies are booming at U.S. universities—in part because of the great job prospects that they offer to grads. But those same prospects are draining away AI and machine learning professors, who are lured by lucrative, challenging jobs in the private sector. That’s happening even at top universities like Stanford, says Jennifer Chayes, a former UCLA mathematics professor who now heads Microsoft Research for New York City and New England. “I don’t know if there are any—or there are very, very few—of the core [machine learning] faculty at Stanford left,” she says. “I just hired one last year.”
She talked about the issue at Bloomberg‘s Spotlight on Artificial Intelligence conference in San Francisco today. “Stanford can’t begin to pay them enough,” says Chayes. A bigger incentive, she thinks, is that working in the private sector is more intellectually appealing. Professors can focus on their research without having to teach hundreds of students. And they have access to better data from customers. “If you are a researcher in AI and ML, you need massive amounts of … interactive data,” she says.
Many private-sector AI researchers teach at universities on the side, but Chayes says that’s not a reliable fix. And while companies do train interns and employees—she says Microsoft has about 1,500 AI interns per year—this doesn’t help with the basic research needed to develop new types of applications. “So I think it’s a really bad trend,” Chayes says. “[A]ll of us need to figure out a way to make it possible for AI and ML faculty to stay at universities and train the next generation.”