Teacher Giving Computer Science Lecture to Diverse Multiethnic Group of Female and Male Students in Dark College Room. Projecting Slideshow with Programming Code. Explaining Information Technology
An arrow pointing leftHome

How UW is addressing the AI faculty shortage

  • Hope Reese

Universities are trying out letting professors work for tech companies, too.

Across the country, many students eager to master AI have found themselves without the tools to do so. A July report from the Center for Security and Emerging Technology highlights what it calls “the lack of AI-teaching capacity at U.S. universities,” which it suggests is due, in large part, from the fact that universities are not doing the best job filling these faculty positions. We covered the phenomenon in relation to how it may affect the students in the Pacific Northwest: Many students want to earn degrees in computer science, yet local universities have not been able to keep up in providing enough faculty and training programs to meet the demand. The report cautions that this gap will have a negative impact on the “production of AI talent.”

The region, home as it is to Microsoft, Amazon, and many big tech companies, may have more at stake if this is true. The so-called “production of AI talent” was named as a particular worry by The Seattle Times editorial board. In an op-ed response, the editors claimed that computer-related degrees are not keeping up with demand––the state’s tech companies are creating twice as many new jobs as could be filled by recent graduates.

According to Ed Lazowska, a professor at the Paul G. Allen School of Computer Science & Engineering at the University of Washington, universities are not meeting demand in computer science—for both students and employers. The demand in CS at the graduate level is “huge compared to other STEM fields,” he said, and this is when CS becomes specialized into areas such as language processing, robotics, and machine learning. Lazowska is a proponent of growing UW’s CS offerings.

UW noticed this gap in its own student enrollment at the Allen School. From July 2012 to July 2018, when it expanded from offering 160 to 450 bachelor’s degrees a year, “growth in student demand vastly exceeded even this substantial growth in capacity,” according to an internal report. Noticing this demand, the Allen School put forward a proposal for a “a new multi-year initiative to grow from 620 to 1,020 degrees per year, including growing bachelor’s degrees from 450 to 770 per year, with a particular focus on ensuring that our undergraduate student body reflects the many dimensions of diversity in Washington State.”

And it is true that local students often end up in local tech jobs, he said. CS students in the Paul G. Allen School at UW are mostly from Washington (last year, that included 79% of undergrads) and often land tech jobs in-state. The most recent data shows that 85-90% of CS graduates from UW, who were employed, were hired in Washington.

On top of all the tech companies with HQ in Washington, it’s also home to the engineering offices of 150 companies, from Facebook to Google, “each of which employs roughly 7,000 engineers in the Seattle area,” Lazowski said.

“There really is no reason to go elsewhere,” he said. “Pretty much every tech company you might want to work for either is headquartered here or has a significant presence here.”

One reason universities may have trouble keeping AI faculty, Lazowska said, is because industry positions often come with the opportunity to access vast amounts of data, which is essential for training machine learning systems on, as well as compute resources. Lazowska says that UW is tackling this issue by allowing faculty to be “dual-hatted” –– able to both teach and work for companies. (The traditional cap is 20% for consulting work, and places like Stanford, for instance, have lost top AI faculty “by being rigid,” Lazowska said). UW faculty, as a result, are simultaneously working for the Allen Institute for AI, Amazon, Apple, Facebook, Google, Microsoft, NVIDIA, and other tech companies.

Not everyone, however, sees filling local tech jobs with local students as realistic. Lazowska pointed to the fact that these global tech companies are “hugely selective,” and are looking at top CS graduates across the country. “I have heard that a typical leading tech company may hire only 1% of those who inquire about employment,” he said.

Additionally, many companies are notorious for “poaching from other companies,” rather than hiring grads directly. “This is particularly true in AI currently,” said Lazowska.

Michael Quinn, who spent fourteen years as the Dean of the College of Science and Engineering at Seattle University, retiring last month, noted the gap in supply-demand. Yet, he said, “it is absurd to lament that Washington’s tech companies are adding new positions faster than universities in the state of Washington are awarding computer-related degrees.” It’s not an apt connection to make, Quinn argues, since these global companies “attract talent from around the world,” and “would not want to fill all their open positions with graduates from Washington’s institutions of higher education, even if these institutions were producing sufficient graduates.”

The region, as home to “the nation’s most vibrant tech industry, will always be a net importer of talent,” said Lazowska.

“The key thing is that every Washington kid who has the inclination and the ability should be able to get the sort of education in this state that prepares them for a first-rank position in our innovation economy,” he continued. “That is not currently the case.”