Tech CEOs and economists have swapped scripts on whether AI is coming for your job

Attn Spritzler Report staff. Want to maintain your employment? Row harder!

Tech CEOs and economists have swapped scripts on whether AI is coming for your job

Jackie Snow, Quartz Media

July 21, 2026

For the past year, the loudest warnings about AI destroying jobs came from the people building AI.

Anthropic CEO Dario Amodei predicted in May 2025 that AI could eliminate half of all entry-level white-collar jobs within five years and push unemployment as high as 20%. The same month, Mark Zuckerberg said AI could soon do the work of midlevel engineers. In July, Ford's Jim Farley predicted the technology would replace "literally half of all white-collar workers" in the U.S. Amazon told employees to expect a smaller corporate workforce, full stop.

Economists mostly rolled their eyes. The evidence that AI was actually behind layoffs was thin, they pointed out, and executives had every incentive to blame a shiny new technology for cuts they wanted to make anyway.



Tech leaders have spent the past few months walking their doomsaying back. Sam Altman now says the industry got its technological predictions roughly right and its economic ones wrong, and that the companies adopting AI fastest are also hiring the most. Last month, Amodei wrote that he was never trying to be a "prophet of doom," though he still sees lasting job loss as possible.

Zuckerberg, asked about AI-driven displacement a few weeks after laying off 8,000 people, said people treat it as inevitable and "I don't actually think it is."

Economists, meanwhile, have started ringing the alarm bells the CEOs put down. More than 200 economists and researchers, including 16 Nobel laureates, signed a statement this month warning that AI could remake the economy faster and more broadly than the Industrial Revolution did. The signatories include MIT's Daron Acemoglu and Simon Johnson, two Nobel winners who greeted such warnings with public skepticism until recently.

The vibes are off

The data is starting to back the worriers up. New payroll research covering 4.6 million workers found that employment for 22-to-25-year-olds in the jobs most exposed to AI is now shrinking by nearly 4% a year. Through at least April, white-collar payrolls had contracted for more than 30 months, which one former Glassdoor chief economist says hasn’t happened outside a recession.



The shrinking payrolls aren't necessarily all AI's doing. Layoffs have plenty of causes, and unemployment overall has held steady. But the public has soured on AI, and the people selling it have noticed.

Nearly two-thirds of Americans now believe AI will mean fewer jobs over the next two decades with only 5% thinking it will create more, according to Stanford's AI Index. Data centers have become the physical symbol of that resentment.

A Gallup survey found 71% of Americans oppose one being built near them, a worse showing than nuclear power plants. Local opposition delayed or blocked at least 75 U.S. data center projects worth $130 billion in the first quarter alone.

The anger has an economic logic. AI's infrastructure buildout is buying all the chips and pushing up the price of laptops, phones, and game consoles, even as wages stay flat. Electricity rates are up across many places where data centers set up shop. The mood tracks with what some economists have started calling a boomcession, an economy that has a lot of strong markers on paper, but with many Americans reporting feeling worse off than ever before.

It’s also not just an American story. Offshore hubs like India and the Philippines are getting hit first as companies swap outsourced coding and call centers for AI, while unions in South Korea are striking over factory robots. There isn’t necessarily a friendlier labor market to decamp to (besides space for data centers, of course).

Time to make nice

The industry's answer, besides the new talking points, has been a burst of conspicuous helpfulness. Anthropic committed $350 million to easing the economic transition, including a fellowship that pays early-career workers $85,000 to spend a year integrating Claude at nonprofits. OpenAI's foundation pledged $250 million. Both companies, along with Amazon and Microsoft, are founding partners of Raise Us, a new $500 million nonprofit that will test policies like wage insurance at the state level.

Even Bernie Sanders is getting meetings. After the senator proposed giving the government a 50% stake in major AI companies, Altman reportedly told him he supports some version of public ownership. In a separate meeting with the White House, Altman threw out 5% as a nice cut of his company to give to the American people.

Call it hedging, call it reputation management. Several of these programs respond to fears of AI displacement by training workers to use more AI, which tells you something about who benefits either way. The lesson the CEOs seem to have absorbed is simpler. What makes a good pitch to an investor makes a terrible talking point to the public.



—Jackie Snow, Contributing Editor