AI and Jobs in California Show Little Displacement So Far

AI and Jobs in California Show Little Displacement So Far

Estimated reading time: 7 minutes · Last updated:

Current evidence on AI and jobs in California does not show widespread displacement, according to PPIC researchers Sarah Bohn, Julien Lafortune and Daniel Payares-Montoya. Workers whose tasks overlap most with AI have seen little change in employment. Their unemployment claims have not jumped since public AI tools arrived in 2022. The anxiety is real, though. Nearly half of employed Californians expect AI to reduce the number of available jobs. About one in three fear it could cost them their job, their hours or their salary. The three researchers wrote in the 12th installment of a PPIC series on issues facing California's next governor. They argue the state's task is to spread AI's gains as well as to cushion its losses.

Key takeaways

  • No broad displacement yet: Californians in the jobs most exposed to AI have seen little employment change and no surge in unemployment claims since 2022, PPIC researchers find.
  • Workers are worried: PPIC's November 2025 survey found nearly half of employed Californians expect AI to cut available jobs, and about one in three fear losing work, hours or pay.
  • Exposure crosses education lines: Of California workers in the most AI-exposed jobs, 55% hold at least a bachelor's degree and 45% do not.
  • Adoption is lopsided: AI use reaches 37% of large U.S. firms but less than 20% of small businesses, and 40% in information and professional services.
  • California's AI industry is hiring: California hosts 33 of Forbes' top 50 AI companies, which added almost 30,000 jobs from 2022 to 2026 according to PitchBook data.

Why Aggregate Data Shows No AI Job Shock Yet

The researchers began with the workers expected to feel AI first. These are people whose daily tasks overlap most with what the technology can do. Employment for that group has been broadly flat. Their unemployment claims have also stayed level since the first public AI tools reached users in 2022. A sudden wave of replacement would have produced the opposite pattern.

Bohn and her colleagues caution that statewide totals can hide trouble in narrower pockets. Soon after ChatGPT was released, claims rose among college-educated workers in highly exposed jobs. The same rise appeared among Bay Area workers in those occupations. A flat average can sit on top of a group that is losing ground.

Losing a job is also not the only way AI can hurt a worker. Some people may keep their positions while their hours shrink, their pay shifts or their duties are rewritten. None of that registers in a claims count. The evidence supports a narrower conclusion: the large shock has not arrived, and the smaller ones are hard to see.

AI Exposure Reaches Far Beyond College Graduates

Office and analytical work lends itself to AI more readily than jobs that need hands-on attention. Exposure measures how closely a job's tasks overlap with AI's capabilities. By that measure, the at-risk group is broader than the usual focus on graduates suggests. Among California workers in the most exposed jobs, 55% hold a bachelor's degree or more and 45% do not, as first reported by the Public Policy Institute of California.

Data scientists and software developers make the list. So do secretaries, office clerks, administrative assistants and customer service representatives. Those clerical roles often employ people with less formal education. By the researchers' count, women, white and Asian workers, and high-income workers are more exposed than others, though no group is exempt.

The occupations least exposed to AI are mostly low-wage. Of the people working in them, 68% have a high school diploma or less. That cuts against a simple story in which AI threatens only the well-paid. Many of the exposed roles are clerical jobs held by workers without degrees.

How Uneven AI Adoption Could Pick Winners

Whether AI replaces work or reshapes it depends heavily on how employers deploy it. Tools that make a worker more productive could also lift that worker's pay. Firms might use AI to improve products, add services or cut costs, and any of these could drive growth and hiring. The same tools could also weaken demand for some workers or make an occupation harder to break into.

Access is uneven among workers. PPIC's November 2025 statewide survey found that AI use on the job climbs with education and income. It is also more widespread among Bay Area workers than elsewhere in California.

The gap among businesses is larger. National data cited by the researchers shows adoption accelerating, but 37% of large U.S. firms use AI against less than 20% of small businesses. Usage reaches 40% in information and professional services, while manufacturing and retail sit below 20%. The researchers say those differences raise a direct question about which firms and workers stand to benefit as use spreads.

Where California's AI Job Gains Are Showing Up

California hosts 33 of the 50 companies on Forbes' list of top AI firms. PitchBook data shows those companies added almost 30,000 jobs between 2022 and 2026. Some of those positions may be based in other states.

AI-related work looks much like the rest of the tech sector. It includes technical roles such as computer hardware engineers and computer and information scientists. It also includes product and commercial jobs such as web design and sales. Since 2022, employment in these kinds of occupations in California has grown 8.3%, compared with 0.5% growth statewide. The researchers note that some of this hiring may come from businesses with no AI connection.

Investment is the second channel. Spending on data centers, computing equipment and software has accounted for a large part of recent GDP growth. Research the authors cite suggests the national buildout may support over 300,000 manufacturing jobs supplying data center construction. California has so far shown little sign of a matching rise in its own construction and manufacturing employment.

What the Next Governor Can Do About AI and Work

The researchers argue that the wide range of possible outcomes calls for broad workforce policy rather than a bet on one scenario. They want planning that expects some new jobs and businesses to appear while others decline.

Education and training systems may need to prepare young people for modernized careers. Working adults may need ways to refresh their skills. State programs that support workers may have to serve a wider group of Californians than before. Beyond retraining, the authors point to a sturdy unemployment insurance system. They also suggest considering alternatives that support workers through job changes, citing wage insurance as one example.

Measurement is the third piece. Tracking hiring, earnings and career entry alongside layoffs can expose pain points, but official statistics can lag. The researchers recommend adding real-time input from business leaders, workers and training providers so that policy can adjust as the labor market shifts.

AI adoption by firm size and industry, U.S. data cited by PPIC researchers
Group Category AI adoption
Large firms Firm size 37%
Small businesses Firm size Less than 20%
Information and professional services Industry 40%
Manufacturing or retail Industry Less than 20%

How AI could play out for California workers

The case for

  • California's 33 of Forbes' top 50 AI companies added almost 30,000 jobs between 2022 and 2026, per PitchBook.
  • Employment in AI-type occupations in California has grown 8.3% since 2022, against 0.5% statewide.
  • AI tools that raise productivity could lift pay, and firms using AI to expand services could add hiring.

The case against

  • Unemployment claims rose for college-educated and Bay Area workers in exposed jobs soon after ChatGPT's release.
  • Adoption below 20% among small businesses, manufacturing and retail could leave many workers without the tools to benefit.
  • The national data center buildout has produced little visible gain in California construction and manufacturing jobs.

What to be careful about

  • Statewide figures may hide job losses concentrated in particular occupations, regions such as the Bay Area, or demographic groups.
  • Clerical workers without degrees, including office clerks and administrative assistants, sit among the most exposed and may have less access to AI training.
  • Official labor data may detect cuts to hours, pay or entry-level hiring too slowly for policy to respond in time.

The bottom line

The evidence gathered by Bohn, Lafortune and Payares-Montoya shows a state that is benefiting from AI's growth. The job losses many workers fear have not yet appeared in aggregate data. That is not the same as safety. Exposure runs through clerical jobs held by workers without degrees. Adoption favors large firms and better-paid employees. The data center boom has not brought visible construction or manufacturing hiring to California. The next governor inherits an anxious workforce, a concentrated industry and statistics that may move too slowly to flag trouble early. The researchers make the case for preparing now: training, a dependable safety net and faster labor market signals, built before disruption grows large enough to measure.

What to watch

  • Watch for PPIC's further monitoring of AI's effects on California business and the workforce, including education and water; no publication date has been set.
  • Watch unemployment claims among college-educated and Bay Area workers in AI-exposed jobs for any sustained rise beyond the early post-ChatGPT increase; no release schedule is given.
  • Watch whether California's next governor proposes changes to unemployment insurance, retraining or wage insurance in response to AI; no date has been set.

Frequently asked questions

Is AI already causing job losses in California?

Not on a broad scale, according to PPIC researchers Sarah Bohn, Julien Lafortune and Daniel Payares-Montoya. Employment and unemployment claims among the most exposed workers have changed little since 2022. Claims did rise for college-educated and Bay Area workers in exposed jobs shortly after ChatGPT was released.

Which California jobs are least exposed to AI?

The least exposed occupations are mostly low-wage jobs that need hands-on attention rather than office or analytical work. Of the workers in them, 68% have a high school diploma or less.

Has the AI data center boom created jobs in California?

Not visibly so far. Research cited by the PPIC team suggests data center construction may support over 300,000 manufacturing jobs nationally. The researchers found little indication of a rise in California construction and manufacturing employment.



Share:

Categories

Newest course every month

Advertise your offline course to a wider audience with our landing page.

You May Also Like

PPIC researchers find AI and jobs in California show no broad displacement yet, but exposure spans education levels while business...
A Federal Reserve Board paper estimates AI and large language models lifted the US natural unemployment rate about 0.2 point...
US data centre emissions reached 25.7 million metric tons of CO2 in a year, more than Denmark, says a Harvard...