How Brain Drain from Academia Could Impact the AI Talent Pool
When companies poach academics that study and teach artificial intelligence, breaking the chain of knowledge transfer, they could be hurting their future talent pool in the field.
In the emergent war to have the best artificial intelligence capability, academia might have the most casualties. According to the National Science Foundation, 57 percent of new computer-science doctoral graduates in the United States take industry jobs, meaning they leave academia for the private sector. This is compared to 38 percent a decade ago, according to The Wall Street Journal.
Given that academia is the primary breeding ground for skills in emerging fields like AI, what would a constant academic exodus of talent in the field mean for the future development of its talent pool?
One of the biggest concerns is that there will be fewer graduates with a thorough education in AI. “The number of graduating master’s and Ph.D.-level computer scientists may decrease, which is the opposite to what the current market is demanding,” said Peter Morgan, chief AI officer at Ivy Data Science, an AI-as-a-service platform and training company based in New York City. “These companies are in effect eating their own lunch.”
The hot market for AI talent means salaries are generous, luring would-be academics away to big paychecks, said Daniel Tauritz, associate professor and associate chair for undergraduate studies and outreach activities for the department of computer science at Missouri S&T in Rolla. One of Tauritz’s undergraduate students received an offer from Google with a six-figure base salary, much more than most professors make.
“The biggest problem is psychological if the gap gets too large,” Tauritz said. “For instance, with your undergraduate students, when their starting salary out of school is more than you make after you’ve been working for 20 years, then you just feel a bit underappreciated.”
When top researchers go into the industry they’re studying, the short-term impact isn’t great, Tauritz said. However, the long-term impact could be significant, as there would be fewer teachers to cultivate the next generation of AI talent.
The Positives of Poaching
Despite these concerns, the exodus of AI talent in academia isn’t all bad.
Most company research has been kept secret due to competition. In December 2016, Apple moved from that secretive approach to a more open one. “To attract the best and the brightest, including from academia, it was not attractive for a lot of people to go there and do stuff that would be in secret,” Tauritz said. This trend has encouraged companies to allow their staff to publish, which he added helps with the brain drain from academia’s research.
Aside from the knowledge base that AI academics provide companies, they also help create longer-term forecasts than what’s typically popular in business, Tauritz said. Rather than seeing short-term monetary gains as a motivator, they think in long development cycles, which helps with innovation.
Companies also benefit from how AI academics think, according to Andrew Chamberlain, chief economist at employer reviews website Glassdoor. The rigor and depth of theoretical understanding they have is hard to attain in the private sector, so companies can benefit from using those skills and applying them to business problems.
Research from academics also can present itself in those business problems, Chamberlain said. “There’s a lot of really brilliant ideas sitting in journal articles that never get read,” he said. Researchers can therefore act as translators to apply the latest science to a commercial product.
Chamberlain doesn’t see poaching of AI academics as a problem. He sees the main function of universities being in research, and researchers are the ones driving innovation, but they aren’t always those doing the teaching. Even if some of them are being poached away, Chamberlain isn’t worried, because even when academics get poached, most of them continue doing academic research, going to conferences and publishing in journals, he said. Additionally, some top academics who have gone on to large tech companies “continue to work at the universities in a diminished capacity,” according to The Wall Street Journal.
Chamberlain added that many academics aren’t well-suited for a private-sector environment, as it requires business sense and communicating in layman’s terms. This limits the talent pool to poach from.
Investment in AI Talent Futures
For companies concerned about talent poaching of AI academics and the implications it might have on the future of knowledge transfer at universities, there are ways to stem the risks.
For starters, more companies with subject-matter expertise in a given emergent field can partner with universities to teach students. Next IT, an intelligent interface company based in Spokane Valley, Washington, has one lead researcher teaching courses at universities, said Jen Snell, the company’s vice president. Next IT also works with business schools, speaking with students about working in IT and AI. These practices pass knowledge on to students, but they also have benefits for the company. Snell said the company’s interns bring a fresh perspective of how to engage with and apply technology.
Tauritz highlighted other ways that companies can make sure knowledge in emergent fields doesn’t leave out the door as its top talent does:
- Provide corporate grants for universities.
- Allow for employees to act as guest scientists, teaching courses at a university while a professor is on sabbatical.
- Host seminars to bring the company’s research and ideas to the university.
Lauren Dixon is an associate editor at Talent Economy.