The Blubber

The CEO of a major artificial intelligence firm is sounding the alarm about the potential for AI to cause a sharp rise in unemployment in the near future—warning that both policymakers and business leaders are unprepared for the scale of disruption that may lie ahead.

“AI is beginning to outperform humans in nearly all cognitive tasks, and as a society, we’re going to have to confront that,” said Dario Amodei, CEO of Anthropic, during an interview with CNN’s Anderson Cooper on Thursday. “AI is advancing across the board, including in areas where I work and where other executives operate.”

Amodei told Axios on Wednesday that he believes AI systems being developed by Anthropic and other tech firms could displace up to 50% of entry-level white-collar roles, potentially driving unemployment rates as high as 20% within one to five years. If that prediction holds, the U.S. could see joblessness soar to levels not experienced since the peak of the COVID-19 pandemic.

Warnings about AI’s economic impact aren’t new. Scholars and economists have for years raised concerns that automation could eliminate specific jobs or job functions, though projections vary. A World Economic Forum report released earlier this year found that 41% of employers surveyed expect to cut staff due to AI-driven automation by the end of the decade.

Amodei’s outlook stands out both for the breadth of disruption he foresees and because it comes directly from a leader deeply involved in building this technology. His comments coincide with the release of Anthropic’s latest AI model, which the company says can perform tasks for almost the entire duration of a standard workday.

Historically, technological advances have mostly affected lower-skill, lower-wage jobs, with displaced workers often retrained for more skilled and better-paying roles. But Amodei warns that this time could be different: AI might eliminate professional roles that require years of training and education—jobs that are not easily replaced or transitioned from.

He also floated the idea that governments may need to consider taxing AI companies if the technology generates vast amounts of wealth that fail to benefit the broader population.

“If AI leads to the creation of enormous wealth, most of that will naturally go to the companies developing the technology, not everyday people,” Amodei said. “That’s not necessarily in my own financial interest to say, but I think it's worth serious discussion—and it shouldn’t be a partisan issue.”