OpenAI’s Sam Altman admits he was ‘pretty wrong’ about AI job disruption

Sam Altman thought AI would have eaten more jobs by now. It hasn’t, and he’s surprisingly upfront about the miss.

Speaking at the Commonwealth Bank of Australia’s Accelerate AI conference on May 26, the OpenAI CEO acknowledged that his earlier predictions about AI’s impact on the workforce were off the mark. Entry-level white-collar positions, the roles most commonly flagged as vulnerable after ChatGPT launched in 2022, have proven more resilient than he anticipated.

“I’m delighted to be wrong about this. I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened.”

The ‘jobs apocalypse’ that didn’t arrive

Altman described himself as “pretty wrong” about the social impacts of AI. Previous projections suggested that as many as 40% of jobs could be automated by 2030. The reality, at least so far, looks considerably less dramatic.

The key reason, according to Altman, is the human element. Jobs requiring direct interaction have proven harder to automate than expected. He pointed to major firms including HSBC, Amazon, Standard Chartered, and CBA itself as examples of companies that have already linked job changes to AI integration.

Customer support remains the sector Altman has consistently identified as most vulnerable to AI replacement. Altman said he doesn’t think “we’re going to have the kind of jobs apocalypse that some companies talk about.”

What investors should actually watch

If the CEO of the world’s most prominent AI company significantly overestimated the pace of disruption in his own product’s primary use case, that’s worth internalizing. It suggests that the broader market may also be miscalibrating timelines for AI-driven economic change.

One thing hasn’t changed. Altman still believes AI will reshape the economy in profound ways. He’s just recalibrated his expectations about the speed.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.