AI could reshape 410,000 local jobs. Who benefits and who doesn’t?

This story is the second in a three-part series examining findings from the latest Silicon Valley Index. Read the first installment here.

Robin McCarthy watches the images appear on her screen.

Inside her San Jose architecture studio, she types a short prompt into an artificial intelligence program. Within seconds, it produces polished design concepts and photo-realistic renderings.

“It’s exciting and scary at the same time, because you’re trying to figure out, ‘Is this going to affect my job?’” McCarthy said. “I like to think it won’t take away my role, but maybe one day it will.”

That mix of awe and anxiety is spreading across Silicon Valley.

Nearly 410,000 jobs in the region include tasks artificial intelligence can perform, according to the latest Silicon Valley Index, the annual report produced by think tank Joint Venture Silicon Valley. While many of those roles are expected to evolve rather than disappear, others could shrink as companies deploy systems that write code, draft legal documents, design marketing campaigns and analyze data in minutes.

Unlike past waves of automation that displaced factory workers first, this technological shift is poised to hit the professional core of Silicon Valley’s economy.

“What’s different is the exposure — it’s people at the high end of the economy,” Joint Venture CEO Russ Hancock said.

The exposure is not evenly distributed. Households earning more than five times the federal poverty level — about $150,000 for a family of four — account for 19% of AI-aligned jobs, compared with just 5% among households at or near the poverty threshold, about $31,000 for a family of four.

Because the Index looked at language and image tools typically used in office and professional jobs, the risk is concentrated in higher-paying roles rather than spread evenly across all workers.

The Index groups AI’s employment impact into three categories: augmenting human work, restructuring how work gets done and replacing certain tasks or positions outright. The effects are expected to vary widely by occupation and even within the same job.

The report identified dozens of fields with significant AI overlap, including architects, software developers, school psychologists, marketers and lawyers.

At Project 100, a marketing firm with locations in San Jose, Oakland, Hercules and Las Vegas, AI now handles multiple functions once done manually, said founder and marketing director My Nguyen. The technology has accelerated communications with clients and reduced the time needed to build a website mockup from weeks to a few hours.

Project 100 is also testing an AI chatbot version of Nguyen to conduct initial client consultations, he said.

“Our competitive edge is we wanted to make decisions using patterns that a lot of people don’t see,” Nguyen said. “With AI, it’s a lot easier.”

Nguyen sees the potential for AI to replace some marketers and graphic designers, but for now, he is using it to augment his staff’s work. “I had my team go back to school and learn these tools,” he said. Still, he expects “some consolidation” in the industry.

At Costanzo Law Firm in San Jose, AI has saved “a tremendous amount of time” drafting documents, founding partner Lori Costanzo said.

“Boom, it just spits it right out within five or 10 minutes, and that’s something that would take several hours for an attorney to do,” she said, adding that AI outputs still need to be reviewed by people.

While Costanzo does not expect AI to eliminate attorney positions, “I could see it replace a paralegal,” she said.

Not all work is easily automated. San Jose criminal defense lawyer Stephanie Rickard said much of her job depends on human interaction and judgment.

“It’s face-to-face interaction,” Rickard said. “It’s negotiation, it’s knowing what’s important to a particular prosecutor or a particular judge in a particular fact set.”

Silicon Valley’s concentration of software developers makes the region especially exposed. Roughly 140,000 residents work as software developers — more than a third of California’s total and the region’s largest occupational group, according to the Index. As AI tools have grown increasingly capable of writing code, worries have circulated that recent Bay Area tech layoffs may reflect workers being replaced by bots.

“We have no evidence that they are AI layoffs,” Hancock said, noting many companies are still correcting for pandemic-era hiring sprees.

However, Steve Blank, an adjunct professor of management science and engineering at Stanford University, said it appears low-level coders are being replaced by AI, though no reliable count exists.

“This is an industrial revolution with all the social consequences wrapped in,” Blank said. “I think what we’re going to find out is it’s going to create new jobs, but possibly not at the same scale.”

The 2022 release of OpenAI’s ChatGPT — generative AI capable of producing text and later audio and video — ignited a wave of investment and competition among Bay Area firms, including Google in Mountain View, Meta in Menlo Park and Anthropic in San Francisco, as well as global rivals.

More recently, so-called “agentic” AI systems have emerged, allowing companies and individuals to create software agents that carry out multistep tasks with minimal human direction, from writing and debugging code to managing workflows.

“You’re not typing computer code,” former Tesla AI chief and OpenAI founding member Andrej Karpathy wrote Wednesday on social media. “That era is over. You’re spinning up AI agents.”

Those software agents heighten the potential for automation, said Shomit Ghose, a lecturer at UC Berkeley’s engineering school and a partner at Menlo Park venture capital firm Clearvision Ventures.

An AI agent can learn the processes behind many white-collar roles and then “automate all these job functions,” Ghose said. “It never gets sick, never gets a day off, never gets tickets to the Niners game.”

The anxiety is already shaping decisions beyond the workplace. Applications to San Jose State University’s software engineering master’s program fell 40% for the coming school year, and undergraduate applications dropped 16%, said Sheryl Ehrman, dean of the school’s College of Engineering. She said the declines appear tied in part to fears that AI will narrow job opportunities.

“Our students are going to be prepared,” Ehrman said, though she expects the market to be “smaller than it was.”

Oakland payments company Block announced Thursday it would cut 4,000 jobs from its workforce of 10,000. CEO Jack Dorsey said on social media that despite improving profitability and rising customer numbers, “intelligence tools” were helping deliver “a new way of working.” The company did not respond to detailed questions about how artificial intelligence factored into specific job cuts. Investors pushed Block’s stock price up 23% following the announcement. Dorsey later said the company had overhired during the pandemic.

Nationally, job-search firm Challenger, Gray & Christmas reported in December that companies publicly cited AI as a factor in nearly 55,000 layoffs last year.

Some analysts caution that economic realities may slow widespread displacement. A Feb. 20 report by financial services firm Citadel Securities said large-scale replacement of white-collar workers would require vast computing capacity. If the costs of chips, data centers and energy exceed the price of human labor for certain tasks, a “natural economic boundary” could limit how far automation goes.

AI firms are racing to expand that capacity, spending hundreds of billions of dollars annually to build infrastructure in pursuit of what many believe will be a transformative market.

As AI reshapes work at accelerating speed, those in AI-exposed occupations are wrestling with what may be lost.

Architect McCarthy worries widespread adoption could diminish ingenuity and artistic vision. Lawyer Costanzo fears young attorneys may never fully develop foundational skills if drafting becomes automated. Marketing executive Nguyen foresees a flood of AI-generated materials that erode quality if used without human judgment.

The uncertainty now extends beyond individual professions and into the region’s broader economic future.

“How do we balance profitability with the need for gainful employment?” said Ghose, the UC Berkeley lecturer. “Employment is what pays taxes, employment is what drives purchases at the restaurant, and flowers for Mom on Mother’s Day, and new shoes for the kids.”