Last month, 40 policy and labor organizations took their fight for a federal framework to protect workers from AI disruptions to Capitol Hill. Led by the Economic Policy Institute, the AFL-CIO Tech Institute, and two social justice groups, We Build Progress and Workshop, they reminded members of Congress that majorities of Americans want to establish guardrails that protect jobs and states’ regulatory powers.
Since Congress hasn’t really stepped up, states are doing the heavy lifting. All 50 have introduced AI legislation in areas ranging from education and health care to criminal justice. But specific protections to shield workers from job losses are still in their infancy. Even in California, an AI law pioneer, state lawmakers are hard-pressed to keep up. One pending bill would require employers to provide 90 days’ advance notice of layoffs “before any technological displacement affecting 25 or more workers during any 30-day period.” It also would mandate that employers with more than 100 workers allow the affected individuals to apply for other positions within that organization.
Tech companies are increasingly laying off employees who are slow to adapt to the new technology or never learn to use it at all.
Layoffs are outpacing those kinds of legislative fixes especially in the hard-hit tech sector, where there are deeper problems that need addressing. An entire generation of new graduates sold on the promise of job security in tech can’t even get their collective foot in the door before layoffs hit.
Software engineer Zaul Moayedian began his career at Paystand, a Santa Cruz–based financial tech company handling business-to-business fiat and cryptocurrency-based transactions. He started out in their blockchain department before transferring to an IT role, and then moved to an engineering team. His team began using AI roughly two years ago after company executives realized the technology could help engineers develop and push their products faster. (Disclosure: I worked with Moayedian at Paystand as a college intern with the company.)
Moayedian, who worked on coding for a cross-border payment system, says his manager pushed his team to start using an AI agent. He claims that the results persuaded the manager to call for the entire company to adopt it and that top executives realized they could get much more work done, cut costs, and maximize profits with far fewer people involved. Moayedian was laid off this past October after five months on the engineering team.
Tech companies are also increasingly laying off employees who are slow to adapt to the new technology or never learn to use it at all. At the financial tech company Block (formerly known as Square), in addition to cutting 4,000 employees, Block’s CEO Jack Dorsey enthusiastically endorsed the move. The Oakland, California–based firm was also tracking employees’ AI tool usage. A former employee claimed in a New York Times opinion column that “adoption was not optional” and that “layoffs later became the enforcement mechanism.”
Lorena Gonzalez, president of the California Federation of Labor Unions, says the problem extends beyond letting people go: “We’re seeing, not only large layoffs because of the introduction of artificial intelligence, but we’re seeing a lack of hiring as well.”
Some tech company executives may think that AI is capable enough, so they proceed with laying off a good chunk of their workers anyway. But Gonzalez believes these numbers could be an attempt to pad stock prices rather than a true reflection of AI’s impact on the workforce. “We need to know what jobs are being replaced, who’s being replaced and at what rate, and if those jobs will ever come back,” she says.
Stanford Digital Economy Lab researchers found early-career workers aged 22 to 25 in AI-exposed occupations such as software engineering or customer service experienced a 16 percent relative employment decline, while employment for workers in less-exposed occupations either remained stable or grew.
There’s not much hard data on the number of jobs AI is eliminating. Without more concrete information, California state lawmakers could be flying blind and risk coming up with fixes that may not fully address the problem—especially when it comes to the challenges younger workers face. Gonzalez believes that legislative proposals must also quantify AI’s effects. “We don’t know what’s actually happening and that’s why it’s so important to have legislation [so] we’re not just relying on anecdotal data,” she adds.
Junior engineers were historically responsible for a lot of the menial day-to-day tasks in startups and big tech companies—work that AI is beginning to take over. Stanford’s Digital Economy Lab noted that software engineering tests showed that “AI systems improved from solving 4.4% of coding problems to 71.7% … [and] current systems could match or outperform up to 47 percent of industry professionals on [selected] tasks.” Being merely human, particularly in a hyper-competitive environment like Silicon Valley, can lead executives to believe outsourcing entry-level work like coding to AI is cheaper, more efficient, and just as effective.
When senior software engineers either shift into other roles or retire, who will replace them?
“Businesses want to farm out all of their relatively easy tasks to AI tools, [and have] senior software engineers guide the AI tools … the assumption is they’ll understand the problem,” says Alex Rudnick, a University of California, Santa Cruz, computer science lecturer. Rudnick had a long career as an applied scientist and software engineer at Google and Etsy before moving into academia.
This approach has major drawbacks: Senior software engineers will either shift into other roles or retire. When they move on, who will replace them? “Senior software engineers definitionally have at one point been young,” Rudnick says. “They have to have been junior software engineers. And the question is … how do you grow [from] a junior software engineer into a senior software engineer?”
In more complex technical areas, if junior engineers don’t gain that experience, a company won’t have people capable of managing or, when necessary, reining in AI systems once the senior employees who understand a company’s systems intimately are gone. Instead, having junior engineers continue to learn how to write code and do software maintenance enables them to catch poorly written code that AI often generates.
When companies remove the junior employee from the equation and instead turn to AI, they will eventually run into an even bigger problem: Who will maintain these systems in five to ten years? Training junior engineers on menial tasks reinforces fundamentals learned in college or in internships, allowing them to develop the skills that eventually mold them into seasoned engineers with superior skills.
California lawmakers can work to protect employees from mass layoffs in the short run with stopgap measures, but they also should take into consideration what happens when mass AI adoption comes at the cost of identifying, hiring, and molding the next generation of talent.