The rehire rate tells you everything about who's actually using AI and who's just using it as cover.
The Summary
- At least 17 major companies — including Uber, Snap, Block, and GitLab — have explicitly cited AI as a factor in layoffs, with Challenger, Gray, and Christmas finding AI cited in 8% of 2025 job-cut plans
- A 2025 Robert Half survey found 29% of hiring managers reopened positions they'd eliminated after implementing AI, suggesting many cuts were premature or misguided
- An MIT study found 95% of corporate AI investments have generated "zero return", raising questions about whether AI is replacing workers or just providing cover for standard cost cuts
The Signal
The data points in opposite directions. Companies are cutting jobs and saying it's AI. Then they're quietly rehiring those same roles. The gap between those two facts is where the truth lives.
Uber cut 10% of its customer service staff in July, explicitly citing AI. GitLab eliminated 350 positions in June to prepare for what it called "the agentic era." Block, Snap, Atlassian, Angi — the list runs long. These aren't small players. These are companies with scale, resources, and presumably some understanding of what AI can actually do.
"29% of hiring managers reopened positions they'd eliminated after implementing AI."
But here's the tell: nearly one in three managers brought those jobs back. That's not a rounding error. That's a pattern. Either these companies fundamentally misunderstood what AI could handle, or they used AI as convenient cover for layoffs they wanted to do anyway. Sam Altman said it directly: some companies are blaming AI for cuts that would have happened regardless.
The MIT study adds weight to the second theory. 95% zero return on corporate AI investments. That's not AI replacing humans. That's AI sitting in a Slack channel while humans do the actual work, then getting credit for "efficiency gains" when the headcount drops.
The timeline matters here. We're not talking about 2023 experiments. This is 2025 and 2026 data. Companies have had years to figure out what AI can do. If they're still cutting jobs, realizing it doesn't work, and rehiring, one of two things is true:
- They're running AI deployments without any real testing or measurement
- They're using "AI efficiencies" as a cleaner story than "we need to hit our margin targets"
The Challenger, Gray data showing 8% of job cuts citing AI is either the leading edge of a real shift or a new corporate buzzword for downsizing. The rehire rate suggests it's closer to the latter. Real automation doesn't require do-overs.
What's actually happening in customer service and content moderation roles is more nuanced. AI can handle tier-one support. It can route tickets. It can answer the same question 10,000 times without getting tired. But it can't handle edge cases, angry customers who want a human, or situations where judgment matters more than pattern matching.
The Implication
If you're in a role that might be on an AI layoff list, watch what companies do, not what they say. The 29% rehire rate means there's a decent chance the company gets this wrong the first time. Document what you do that AI can't. Be ready to explain it.
If you're running a company, stop using AI as a euphemism for cost cuts. The MIT data shows most AI deployments aren't generating returns yet. That's fine. Early adoption is expensive. But don't cut people, discover AI can't do their job, and then scramble to rehire. Test first. Measure. Then cut.
The real signal here isn't that AI is replacing humans. It's that companies still don't know what AI can actually do, but they're happy to cite it when they need to explain layoffs to shareholders. That gap won't last. The companies that figure out the real capabilities now will have an edge. The ones still using AI as cover will keep cycling through that 29% rehire loop.