The jobs crisis is bigger than AI

• Job seekers are struggling, especially entry-level applicants, but sluggish demand, inflation and restructuring are still bigger factors.

Zoom out: In the 33 months after ChatGPT's launch, researchers found no discernible disruption to the broader labor market, according to a report this week from the Budget Lab at Yale, a nonpartisan policy research center.

• Yale researchers measured "occupational mix" — the distribution of workers across occupations — and found that it looked similar to the birth of the PC in the 1980s and the early internet in the 1990s.

By the numbers: So far this year, job losses due to "technological updates" are far less than other factors, according to a report Thursday from Challenger, Gray & Christmas.

• 20,219 jobs were lost as a result of "automation and possibly AI implementation" in 2025 through September, per the firm, which has tracked layoffs for decades.
• Just 17,375 cuts in 2025 have been explicitly linked to AI.
• In the same period the report attributed 208,227 lost jobs to "economic conditions" and another 293,753 cuts to "DOGE actions."

Between the lines: The numbers reflect what current AI tools can accomplish in comparison to humans.

• An OpenAI report released last month shows that today's leading models are approaching parity with human professionals on many tasks, but the results don't capture the cost of human insight required in a real-world setting.
• AI's productivity gains are likely to come as a complement, not a replacement, for most workers, OpenAI chief economist Ronnie Chatterji told Axios.

What they're saying: Economic leaders at the highest levels also have doubts.

• Federal Reserve chair Jerome Powell expressed "great uncertainty" at a recent press conference that AI is creating a lower labor demand.
• "I think my view — which is also a bit of a guess, but widely shared I think — is that you are seeing some effects, but it's not the main thing driving it," Powell said.

• "It may be that companies or other institutions that have been hiring younger people right out of college are able to use AI more than they had in the past," Powell said. "It's also part of the story, though, that ... job creation more broadly has slowed down. The economy has slowed down."

What we're watching: Deploying AI to replace entry-level workers could put us in a pickle down the road, especially in a world where humans work alongside AI.

• If workers can't get a foot in the door, they'll never make it up the ladder, and eventually that could mean that the bots outsmart us.
• "You cannot supervise an intern if you are less skilled than the intern," Stanford computer science professor Jure Leskovec told Axios.

The bottom line: So far, changes are incremental, not catastrophic.

Go deeper: AI job anxiety: It's real, and coming at the worst time