The AI Jobs Apocalypse Was Supposed to Already Be Here. So Why Are the People Who Predicted It Going Silent?

The AI job apocalypse was supposed to be here by now. It has not happened on the scale many executives and investors warned about.

That shift matters because some of the same people who forecast sweeping labor disruption are now speaking more cautiously, even as AI investment, hiring pressure, and workplace anxiety keep rising across the US economy.

Big predictions are meeting a slower reality

Since ChatGPT’s release in late 2022, warnings about mass job loss have become a constant feature of the AI boom. Nvidia CEO Jensen Huang said in 2025 that “every job will be affected, and immediately,” while Anthropic CEO Dario Amodei said in January that AI should be viewed as “a general labor substitute for humans.” Those comments helped shape a public sense that white-collar cuts were imminent.

But by May 2026, OpenAI CEO Sam Altman publicly conceded the impact had arrived more slowly than he expected. “I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened,” Altman said, stepping back from earlier, more sweeping claims.

The timing matters because AI has become one of the dominant business stories in the country. According to reporting cited by The Guardian, AI accounted for nearly 60% of US economic growth in the last quarter of 2025, underscoring just how central the sector has become to investor expectations.

Why the layoffs story is more complicated

Job losses in tech are real. More than half a million workers in the tech industry have lost their jobs since late 2022, according to The Guardian’s reporting, and companies including Amazon, Meta, and Block have all gone through major cuts.

Still, economists say it is hard to prove AI alone caused those layoffs. Martin Beraja, a professor at UC Berkeley Haas School of Business, said studies tying ChatGPT directly to weaker entry-level software hiring are “problematic” because the industry was already correcting after pandemic-era overexpansion.

Beraja said tech companies staffed up heavily when digital demand surged, then found themselves overbuilt when consumer habits shifted back toward in-person activity. Even venture capitalist Marc Andreessen, one of Silicon Valley’s strongest AI boosters, said in March that some overstaffed firms were using AI as a “silver-bullet excuse” to clean house.

Economists say hype has its own purpose

Some labor and economics experts argue that the most extreme AI job forecasts have always served a business purpose. Suresh Naidu, an economics professor at Columbia University, said companies chasing huge valuations need investors to believe their tools could eventually “eat all the work on the planet.”

That does not mean AI is unimportant. Naidu called the technology transformative and said he uses it in his own work, but he also argued that broad claims about total labor replacement often outrun the actual evidence.

Anil Dash, former CEO of Glitch, made a similar point. He said AI is clearly a leap forward, especially in coding, where output can be tested more directly. But in many other jobs, he said, the “domains of applicability” are still hard to pin down, making immediate, economy-wide replacement far less certain than the rhetoric suggests.

The next phase may be control, not collapse

For many workers, the more immediate AI risk may not be total job extinction but tighter management. Labor experts increasingly warn that employers can use AI tools to monitor performance, measure output, and push people to work faster, even when full automation is not possible.

That pattern is already familiar in gig work, where drivers and delivery workers have long been managed by algorithms. Critics say the same kind of surveillance and micromanagement could spread into office jobs, warehouses, customer service, and logistics before AI directly replaces large numbers of workers.

Experts also say the outcome is not fixed. Beraja argues the strongest evidence so far shows AI can help workers learn faster instead of simply replacing them, while Dash says more responsible, smaller-scale AI systems could offer alternatives to the winner-take-all model. For now, the silence around earlier doomsday predictions says as much as the forecasts once did.