The head of the company that launched ChatGPT has stepped back from one of the more alarming forecasts about artificial intelligence, telling a business audience that a sweeping, near-term collapse of the job market is unlikely. It was a notable reversal of tone from an executive who had earlier warned that entire categories of white-collar work could be automated away, and it landed at a moment when hard data on employment had failed to show the mass displacement many predicted.
What Altman actually said, and where
Speaking virtually in late May 2026 at a Commonwealth Bank of Australia event in Sydney, OpenAI chief executive Sam Altman said he no longer expected the technology to trigger a “jobs apocalypse.” As reported at the time, he described being initially worried about the scale of job losses but said he was “delighted to be wrong about this,” pointing to a smaller-than-feared impact on entry-level office roles. Altman framed the shift as a lesson in humility about predicting consequences: the company had been “roughly right” about the pace of the technology itself, he said, but “pretty wrong” about how quickly and severely it would reshape the social and economic landscape.
The human element he said machines have not replaced
Central to the reassessment was Altman’s argument that the “human part” of most jobs has proven stickier than automation forecasts assumed. He acknowledged using AI tools for routine tasks such as drafting messages and email, but said the interpersonal core of work, judgment, relationships and accountability, was not something he could imagine handing to software any time soon. That distinction, between automating discrete tasks and replacing whole roles, has become a recurring theme among researchers who study how AI is actually adopted inside companies: tools tend to absorb pieces of a job rather than eliminate the position that contains them.
The labor data that undercut the doomsday case
Altman’s softer stance aligned with employment figures that stubbornly refused to show a wave of AI-driven layoffs. Research from the Yale Budget Lab found that AI was likely not the driver behind a general softening in the labor market, and that through early 2026 there had been no meaningful divergence in unemployment for workers in occupations most exposed to the technology compared with everyone else. In other words, the jobs theory most exposed to automation were not disappearing faster than the rest. That absence of a measurable shock is part of why several prominent figures who had sounded alarms about imminent mass unemployment have tempered their language over the course of the year.
How people are actually using the tools
The reassessment also fits a broader picture of how the technology is being used day to day, which looks less like wholesale automation and more like assistance. OpenAI’s own analysis of how people use ChatGPT describes a heavy tilt toward writing help, information seeking, and practical guidance rather than the autonomous replacement of complete workflows. That usage pattern, augmenting a person’s tasks instead of standing in for the person, supports the argument that the near-term effect on employment is more evolutionary than catastrophic, even as the tools spread rapidly across workplaces.
Why caution still runs through the optimism
The revised outlook is not a declaration that AI will leave work untouched. Altman’s phrasing emphasized timing and scale, an apocalypse is unlikely soon, rather than a promise that no jobs will change. Automation has historically reshaped occupations gradually, shifting which tasks humans do rather than erasing employment outright, and analysts caution that a slow, uneven transition can still be painful for specific workers and industries even when aggregate unemployment holds steady. The honest reading of the moment is that the most extreme predictions have not materialized on schedule, that the executives who made them have publicly walked them back, and that the data so far describe a workforce absorbing new tools rather than being displaced by them. Whether that holds as the technology grows more capable remains the open question, and even the optimists frame their revised confidence as provisional rather than final.
This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.
More from Morning Overview
