In the last week of May 2026, three of the most influential figures in artificial intelligence did something tech CEOs rarely do: admit publicly they were wrong. Or, in one case, call it “irresponsible” when companies use AI as an excuse to lay people off.
Sam Altman, CEO of OpenAI, spoke on May 26 at a conference in Sydney. “I was wrong about how fast AI would transform the job market,” he said. “I’m delighted to be wrong.” A year earlier, the same Altman was warning that AI would eliminate “entire classes” of jobs.
Jensen Huang, CEO of Nvidia, was more direct. In an interview with Channel NewsAsia on May 24, he called the narrative linking AI to job loss “lazy.” “AI just got here. How is it possible that you’re already losing jobs?” he asked. He added: “We’re scaring people, and that’s irresponsible.”
Dario Amodei, CEO of Anthropic, rounded out the trio. In May 2025 he was predicting AI would “wipe out” 50% of entry-level office jobs. In January 2026 he published a 20,000-word essay about an “unusually painful” disruption. By May 2026, in a meeting with Jamie Dimon of JPMorgan, he was talking about Jevons’ paradox: the idea that efficiency can expand work rather than eliminate it.
The uncomfortable question is why now. OpenAI and Anthropic are preparing to go public this year, each with valuations approaching one trillion dollars. “Labor apocalypse” is not the message you want to send to retail investors. Fortune, Forbes, and Technology Magazine have all noted the timing.
But there’s more to it than corporate cynicism. The data backs the skeptics. A National Bureau of Economic Research study published in February 2026 surveyed nearly 6,000 executives across four countries. The result: roughly 90% of firms reported that AI had no impact whatsoever on employment or productivity over the past three years. Zero.
Another PwC survey of 4,701 global CEOs found that 56% see no financial return from their AI investments. Only 32% reported revenue growth attributable to the technology.
This does not mean AI won’t transform the labor market. But it does mean the timelines were exaggerated, the warnings were disproportionate, and the available data does not support the narrative of an imminent cataclysm.
The story isn’t that AI doesn’t work. It’s that the same people who sold the fear are now selling the calm — right when they need the public to trust them enough to buy shares.
Main source: Firm Data on AI — NBER Working Paper w34836