Silicon Valley Is Gobbling Up A.I. Experts

A view of people walking in the Facebook main campus in Menlo Park, California, May 15, 2012.  

Robyn Beck/AFP/Getty Images

Artificial intelligence is the latest Silicon Valley darling. Google, Facebook, Amazon, and Apple are using it in their image recognition algorithms, in their voice-based virtual assistants, and in their attempts to curate news and social media posts specifically for your eyes. It’s a hot area for smaller startups, too. Venture capital funding in AI has risen to more than $9.5 billion in the first five months of 2017 from $3.2 billion in all of 2014. It’s been called the biggest investment trend of the year, but it may also be the biggest hiring trend.

While the field dates back to the 1950s, only recently has AI branched out to become a mainstream computer science topic. That being the case, there are only a finite (and relatively small) number of experts in the field—a number that is dwindling in academia due to lucrative offers from tech companies.

In a report from the New York Times this week, Cade Metz details how big tech firms are luring AI specialists into their fold with salaries of up to $500,000 a year. “In the entire world, fewer than 10,000 people have the skills necessary to tackle serious artificial intelligence research, according to Element AI, an independent lab in Montreal,” Metz writes. At Stanford University, four well-known researchers have left or taken leave over the past few years. At another AI hub, the University of Washington, six of the department’s 20 professors are working in the industry on leave or partial leave. If professors and researchers are leaving academia, who’s to teach the next generation of AI specialists?

This isn’t the first time a situation like this has happened. In 2015, Uber poached the majority of Carnegie Mellon University’s leading robotics lab to work on the company’s self-driving car project. After initially luring a couple of developers from the National Robotics Engineering Center to join its ranks, Uber eventually hired 40 former CMU staffers for its project (roughly one-third of the NREC’s total staff). This included a number of upper-level employees, including the organization’s director. “These guys, they took everybody,” one source told the Verge.

As of March 2016, according to the Pittsburgh Post-Gazette, some of the vacancies left by Uber employees remained unfilled (though the NREC told the Post-Gazette at that time that it planned to use $11 million in new research funding to hire 15 to 20 new staffers). But CMU’s School of Computer Science dean Andrew Moore shrugged it off. “This kind of thing happens to us a few times a year,” Moore said to the Pittsburgh Post-Gazette. “We’re focused on ‘what’s next?’ ”

According to Moore, this sort of thing is cyclical. Researchers may start out in academia, spend a few years applying their research commercially, then come back to academia with fresh perspectives and ideas. In a 2016 interview with TechCrunch, Moore said five to 15 staffers usually leave for industry jobs each year, taking leaves of absence up to four years. Some, but not all, end up coming back.

Hopefully, this is what we’re seeing now. And rather than solely relying on poaching researchers from academic institutions, tech firms (other than Uber) are taking things into their own hands in a more constructive way.

Companies such as Google and Facebook are now offering courses to help staff get up to speed on AI. With so few experts to pluck from the field, programs like the Facebook AI Academy aim to make engineers proficient in deep learning. Once trained, they can then apply their new knowledge to other engineering groups at the company. Strategic acquisitions, such as Google’s purchase of DeepMind in 2014, can also help with the AI labor shortage—but only as long as there are AI startups to farm from.

Even so, for companies searching for Ph.D.-level expertise, demand far outweighs supply. If tech companies want to ensure a robust future of AI specialists, they’ll need to strike a balance between employing experts for their own needs and allowing them to continue teaching the next generation