What do you call it when a company lays off thousands of workers, blames AI, then quietly rehires humans because the AI couldn't do the job? Klarna found out - after cutting 40% of its workforce and trumpeting AI as the replacement, the company brought back roughly 20 customer support staff when the technology fell short.
Klarna is not an outlier. It's the clearest example of a pattern now supported by hard data: most companies citing AI as the reason for layoffs have not actually deployed AI to replace those roles.
The Numbers Don't Add Up
A Harvard Business Review survey found that only 2% of executives made large workforce reductions based on actual AI implementation. The other 60% who cut jobs did so "in anticipation" of AI efficiencies that have not materialized.
Of 1.2 million job cuts in 2025, only 55,000 - about 4.5% - actually cited AI, according to outplacement firm Challenger, Gray & Christmas.
The most telling data point comes from New York State, which requires companies to check a "technological innovation" box on layoff notices. Of 160 companies filing layoff notices - including Amazon and Goldman Sachs, both of which publicly blamed AI for cuts - zero checked that box. Not one.
AI as the Least Bad Excuse
Peter Cohan, a professor at Babson College, put it bluntly: blaming AI is "the least bad reason companies can use." Citing tariffs or weak sales triggers negative market reactions. Citing AI signals forward-thinking efficiency.
The incentive structure backs this up. AI-related stocks drove 75% of S&P 500 returns. A survey found 60% of hiring managers said they emphasize AI's role in layoffs because "it's viewed more favorably than financial constraints." Companies want the AI narrative attached to their brand.
Meta has cut over 700 jobs recently with plans for up to 20% of its workforce, while investing $600 billion in AI infrastructure. Block, under Jack Dorsey, told shareholders it would shed nearly half its workforce. Salesforce eliminated 4,000 customer support roles. Hewlett-Packard announced up to 6,000 cuts by 2028. All cited AI.
Even OpenAI CEO Sam Altman acknowledged that some companies are "blaming unrelated layoffs on the technology."
What's Actually Happening
The real driver, according to multiple researchers and economists, is a post-pandemic overhiring correction. Manish Jain of Info-Tech Research Group said it plainly: organizations overhired during the pandemic, and now "with AI as the facade, many, if not all, are happy to offload a large number of people."
Goldman Sachs economist Ronnie Walker found "no meaningful relationship between productivity and AI adoption at the economy-wide level." The Yale Budget Lab found no overall employment change in AI-exposed occupations. Goldman estimates only about 2.5% of US employment is genuinely at risk from current AI capabilities.
There are real, limited impacts. Stanford data shows job-finding rates for 22-25 year olds in AI-exposed fields dropped roughly 14% since ChatGPT launched. Individual productivity gains of 10-15% in programming tasks are measurable. But as University of Sydney professor Uri Gal put it: "The workers being let go are not being replaced by AI today; they are subsidising the AI bet their employer is making on the future."
For anyone worried about AI taking their job tomorrow: the data says your bigger risk is a CFO who overhired in 2021 and needs a good story for the earnings call.