What Is AI Washing
"Washing" originally comes from "greenwashing" in the environmental world - presenting something as eco-friendly when it isn't. AI washing follows the same pattern. It originally meant exaggerating or faking AI capability a company didn't actually have. More recently, it has come to mean something else: hiding the real reason behind a layoff and blaming it on AI instead.
OpenAI CEO Sam Altman himself has confirmed this isn't just an outside accusation. He said, "AI washing is real," pointing out that companies are using AI as cover to disguise layoffs. He also added that a day when technology genuinely does replace jobs is still coming.
The Layoff Numbers Don't Add Up
According to Challenger, Gray & Christmas, a consulting firm that has tracked US layoff statistics for more than 30 years, layoffs that executives explicitly attributed to "AI" topped 100,000 in the first half of this year alone - nearly double all of last year's total, and the first time AI has ranked as the number one cited reason for layoffs since the firm started tracking it as a separate category in 2023.
Around the same time, researchers at Duke University and the Federal Reserve anonymously surveyed 700 CFOs at major US companies, and got the opposite answer. 90% said AI has had essentially no impact on their company's hiring over the past year, and 89% said it hasn't meaningfully affected productivity yet either. AI is the number one stated reason for layoffs, even as the executives making those calls say its real impact has been minimal.
The financials of the companies doing the cutting don't line up either. Oracle's most recent quarterly revenue grew more than 20% year-over-year, with its order backlog up 300%. Microsoft and Meta posted record-level results while announcing large layoffs at the same time. Mass layoffs are usually a response to weak earnings - but the big tech companies leading this round are posting some of their best results ever.
Why Companies Attach AI to Layoffs
The first reason is money. Big tech needs enormous capital to build AI chips and data centers. Meta alone has said it will spend more on AI this year than it spent in the previous two years combined. Amazon, Microsoft, Google, and Meta together have said their AI infrastructure spending this year will be nearly double last year's. With that kind of capital to raise or conserve, payroll is the fastest lever executives can pull.
The second is the hiring bubble left over from the pandemic. When online business boomed during lockdowns, many companies hired aggressively, offering developers high salaries to compete for talent. Now that bloat needs to be trimmed - and the AI boom arrived at just the right moment to give that overdue trimming a convincing cover story.
The third is investor pressure. The tech industry has a benchmark called the "Rule of 40," created by a Silicon Valley venture capitalist in 2015: a company's revenue growth rate plus its profit margin should add up to 40 or more to be considered a strong business. If revenue grows 40% a year, zero profit is forgiven; but if growth slows to 20%, margin has to climb to 20% to still hit 40. For publicly traded big tech companies that need to keep raising massive sums for AI, the pressure to show margin improvement to investors is intense - and the fastest card management can play is payroll.
Companies That Replaced People With AI, Then Called Them Back - the AI Boomerang Layoff
Some companies that cut too aggressively are already paying the price. In 2024, the Swedish fintech Klarna adopted OpenAI's technology for customer service and proudly announced that its AI chatbot was doing the work of 700 support agents. In practice, the AI couldn't calm angry customers, and it kept hallucinating - confidently making up answers on important topics like fees and refund policy. As customer satisfaction dropped, Klarna started rehiring people within a year.
Ford had a similar experience. It had steadily let go of quality-inspection engineers and handed the work to AI, but the AI kept missing defects, and vehicle recalls spiked as a result. The veteran engineers' feel for subtle quality differences couldn't be written into a manual - there simply wasn't training data to teach AI that instinct in the first place. Ford ended up calling back 350 skilled engineers, and only then did recall costs come down.
Rehiring isn't free. It brings back recruiting and training costs, and companies often have to offer a higher salary than before to lure someone back after letting them go. Cutting people to save money can end up costing more in the end.
Is Layoff Washing Actually Legal Trouble?
The US operates on the "employment-at-will" doctrine, meaning a company can end employment without a specific reason. Discrimination based on race or gender, and retaliation against whistleblowers, are strictly prohibited - but citing "AI" as the reason for a layoff isn't, by itself, a labor law violation. Companies with 100 or more employees do have procedural obligations, like giving at least 60 days' written notice before a mass layoff.
Even if it's not a labor law problem, there's another risk. The US Securities and Exchange Commission has previously charged asset managers with securities fraud for falsely claiming they used AI-driven investment strategies to attract investors, when they weren't actually using AI at all. Legal experts say "layoff washing" carries a similar risk: if a company misrepresents the reason for a layoff in a way that misleads investors and moves its stock price, that could also amount to securities fraud. No one has been prosecuted on this basis yet, but warnings are growing that cases with a significant stock impact could become legal targets.
So Is AI Ever Actually Replacing People?
Yes. Stanford University's annual 400-page AI Index Report draws on academic research to put concrete numbers on productivity gains. Accounting work saw a 55% productivity boost after adopting AI, marketing saw 50%, and content creation saw gains of up to 200%. Customer service got somewhat faster too. On the other hand, the same report found that productivity for experienced developers actually slowed down.
Another chart in the same report shows the real point of this story. For both software developers and customer-service roles like call centers, employment across all age groups rose together up until 2023. But starting in 2023, the employment curve for 22-to-25-year-old entry-level workers alone dropped sharply. Employment for 26-to-30-year-olds with less experience also stalled. Meanwhile, employment for workers in their 30s, 40s, and beyond kept rising.
In the End, It's Not AI Replacing People - It's People Who Use AI Well
Put those two charts together, and the real shape of what's happening becomes clear. AI isn't taking over jobs wholesale. A small subset of experienced people who are good with AI tools have started doing, alone, the work that used to take several people in accounting, marketing, and content - and the entry-level and junior positions that would have been needed for that extra work are the ones disappearing. "AI replaces people" isn't quite the right sentence for what's happening. "People who use AI well replace people who don't" is closer to the truth.
Left unaddressed, this doesn't stop at today's hiring crunch for young workers. If the first rung of the career ladder - where new graduates learn the job - disappears, then in a few years there won't be a pool of mid-career professionals to fill the roles above it, and the pipeline for future experts dries up with it. Whether a company is using AI as a genuine reason for layoffs or as a convenient excuse, the conclusion for individuals is the same: what determines who survives inside a shrinking headcount is judgment AI can't replace, combined with the actual ability to use AI as a tool.