A long line of job seekers are seen Tuesday, Oct. 13, 2009 at a career fair in Chicago.
(AP Photo/Kiichiro Sato)
Many tech jobs are clustered in a few regions of the US.
A new wave of generative AI jobs could follow the same pattern, per a Brookings Institution report.
Nearly half of the new US generative AI job postings in May came from one of six metro areas.
Experts say that AI like ChatGPT won’t just replace jobs, but create new ones. So far this job creation is following the same pattern as past tech booms — clustering only in a few regions of the country.
That’s according to a new Brookings Institution report that analyzed online job postings data from Lightcast. In May, 25% of new generative AI job postings — which contained terms like “ChatGPT” or “generative AI” — were posted in the Bay Area (San Francisco and San Jose). 47% of the jobs were posted in the Bay Area, New York City, Los Angeles, Boston, or Seattle.
The data includes 978 generative job openings that were posted in May. Over the 11 months leading up to May, there were roughly 2,100 generative AI postings identified, with the bulk of them — 29% and 47% respectively — coming in April and May of this year. Researchers noted that some postings could lead to multiple hires for the same position.
11% of the jobs were posted in nine other metropolitan areas: Washington, DC, San Diego, Austin, Raleigh, Boulder, Lincoln, Santa Cruz, Santa Maria-Santa Barbara, and Santa Fe. In total, the 15 metros accounted for nearly 60% of May’s generative AI openings.
In the below chart, “early adopters” refers to the 13 aforementioned metros, excluding the Bay Area metros. “Potential adoption centers” refers to 87 additional metros that have a limited AI presence. “Others” refers to a multitude of other metros with little or no AI activity. There are 21 “federal research centers” the report examined.
Before the generative AI boom ever began, about two-thirds of AI research, start-up, and job posting activity in the US was concentrated in one of the same 15 metros, per Brookings, each of which boasts major universities and tech companies. Between 2005 and 2017, five cities — San Francisco, San Jose, Boston, San Diego, and Seattle, accounted for over 90% of the tech jobs created in the US.
The report’s findings suggest that the same major metros could be poised to benefit from the newest tech boom — leaving the rest of the country on the outside looking in again, Mark Muro, a senior fellow at the Brookings Institution and a co-author of the report, told Insider.
“AI activity, as an early-stage technology, is if anything even more concentrated in the Bay Area than other technologies,” he said, adding, “These data show that AI, for all of its uniqueness, appears in many respects set to repeat the usual intense clustering of core tech activity in just a few places, like the Bay Area, that has characterized wave after previous wave of digital buildout. While that allows for efficient innovation, it can lock out many places and technology options.”
While generative AI jobs appear to be clustering in a few geographic areas, it’s possible that many of these positions are remote, meaning that workers across the country won’t truly be left out. Muro said this information wasn’t available in the dataset.
A Morning Consult survey of 750 tech workers, for instance, conducted in April and May of last year found that 48% of them were fully remote, up from 22% before the pandemic began.
But over the past year, some big tech companies like Google and Meta have called employees back to the office a few days per week, suggesting that there are fewer fully remote tech positions than there used to be. And there’s reason to believe the AI industry in particular is more averse to remote work than the rest of the tech sector.
ChatGPT creator Sam Altman for one, the CEO of OpenAI, is not a fan of remote work — at least in the tech startup space.
“I think definitely one of the tech industry’s worst mistakes in a long time was that everybody could go full remote forever, and startups didn’t need to be together in person and, you know, there was going to be no loss of creativity,” he said in May. “I would say that the experiment on that is over, and the technology is not yet good enough that people can be full remote forever, particularly on startups.”
“I feel pretty strongly that startups need a lot of in-person time,” he added, “And the more fragile and nuanced and uncertain a set of ideas are, the more time you need together in person.”
Brookings’ Muro said that he expects many generative AI jobs to be in-person some or all of the time.
“We believe that this kind of early-stage work is exactly the kind of work that’s most often being done by teams in the same space,” he said. “This is the work that is drawing developers back to the Bay Area to participate in person in the excitement.”
The Brookings report listed several measures the US could take to diversify the geography of the generative AI industry in the years ahead. This included expanding public sector research, democratizing access to AI computing resources and datasets, training workforces across the country for AI jobs, and providing government funding to new areas to promote development.