As the debate over the true impact of artificial intelligence on the workforce continues, OpenAI CEO Sam Altman said some companies are engaging in an “AI purge” in terms of layoffs, or incorrectly attributing workforce reductions to the technology’s impact.
“I don’t know what the exact percentage is, but there is some AI purge where people blame AI for layoffs, and AI does replace different types of jobs,” Altman told CNBC-TV18 at the India AI Impact Summit on Thursday.
The AI wash has come into focus as the latest data on the technology’s impact on the labor market tells a confusing, uncertain story about how the technology is or will destroy human jobs, or whether it hasn’t yet touched human jobs.
For example, a study released this month by the National Bureau of Economic Research found that nearly 90% of thousands of executives surveyed in the U.S., U.K., Germany and Australia said AI had had no impact on workplace employment in the past three years, since ChatGPT was released in late 2022.
However, prominent tech leaders such as Anthropic CEO Dario Amodei have warned that an artificial intelligence white-collar massacre could wipe out 50% of entry-level office jobs. Klarna Chief Executive Sebastian Siemiatkowski said this week that the buy now, pay later company will reduce its 3,000 headcount by a third by 2030, in part due to the acceleration of artificial intelligence. According to the World Economic Forum’s 2025 Future of Jobs Report, about 40% of workers hope to follow Siemiatkowski’s lead and eliminate workers through artificial intelligence.
Altman clarified that he expects AI to lead to more job losses and the emergence of new roles that complement the technology.
“We will find new jobs, just as we do in every technological revolution,” he said. “But I expect the real impact of AI on jobs will start to be felt in the next few years.”
Data from a recent report by the Yale Budget Lab suggests that Altman and Amodei’s vision of mass replacement of workers by artificial intelligence is uncertain and has yet to be realized. Using data from the U.S. Bureau of Labor Statistics’ Current Population Survey, the study found no significant differences in rates of occupational portfolio change or length of unemployment among individuals working in jobs with greater exposure to AI from the time ChatGPT was released through November 2025. These data suggest there are no major changes to the current workforce related to AI.
“No matter which way you look at the data, at this moment, there doesn’t appear to be a significant macroeconomic impact,” said Martha Gimbel, executive director and co-founder of the Yale Budget Lab. wealth earlier this month.
Kimbell attributed the AI purge to companies blaming shrinking profits and revenue on a failure to effectively navigate wary consumers and geopolitical tensions. WebAI founder and CEO David Stout also wrote in a commentary article wealth Tech founders are under increasing pressure to justify excessive and continued investment in artificial intelligence, which is why many are creating a narrative of artificial intelligence disrupting the workforce and economy by predicting mass worker unemployment.
Torsten Slok, chief economist at Apollo Global Management, said this era of waiting for the impact of artificial intelligence to emerge rhymes with the IT boom of the 1980s. Nearly 40 years ago, economist and Nobel laureate Robert Solow observed that productivity had barely improved in the personal computer era despite huge projected gains in productivity, and Solow sees a similar pattern today.
“AI is everywhere except in the upcoming macroeconomic data,” he wrote in a blog post last week.
Slock also said the lull in AI-driven economic impact may follow a J-curve, where large early spending masks an initial slowdown in performance before productivity and workforce changes take off exponentially.
Erik Brynjolfsson, an economist and director of the Stanford University Digital Economy Laboratory, said in a report financial times Column Recent labor data may tell a new story about how AI is indeed impacting productivity and the workforce. He pointed out that the latest revision of employment data reflects the decoupling of job growth from GDP growth: although GDP increased by 3.7% in the fourth quarter, last week’s jobs report revised employment growth down to just 181,000. Brynjolfsson’s own analysis shows that productivity increased by 2.7% year-on-year last year, which he attributes to the productivity advantages of artificial intelligence starting to show.
Brynjolfsson published a landmark study last year showing a 13% relative decline in employment rates for early-career workers in jobs with higher exposure to artificial intelligence. At the same time, employment levels for most experienced workers have remained stable or increased.
“The latest U.S. data for 2025 suggest that we are now transitioning from the investment phase to the harvest phase,” he wrote in the report. financial times“Those early efforts start to show up as measurable output.”
This story originally appeared on Fortune.com