In the first eleven months of 2025, companies attributed 55,000 job cuts to artificial intelligence - more than twelve times the number from just two years earlier. The pace hasn’t slowed in 2026. Amazon cut 16,000 positions. Pinterest eliminated 15% of its workforce. Dow axed roughly 4,500 jobs. HP announced plans to shed up to 6,000 employees. In every case, AI was part of the stated rationale.
The message from the C-suite has been consistent: AI is here, it’s replacing jobs, and we’re getting ahead of the curve. CEOs at Ford, Salesforce, Amazon, and JP Morgan Chase have all declared that white-collar roles at their companies will soon disappear.
There’s just one problem. When researchers actually looked at whether AI was doing the work these people used to do, the answer was mostly no.
The 2% Reality
A survey of 1,006 global executives published in Harvard Business Review found that 60% have already made headcount reductions in anticipation of AI. But only 2% of those cuts were based on actual AI implementation results. The rest were based on what AI might do someday.
That’s a gap worth sitting with. Companies are firing real people today because of hypothetical productivity gains tomorrow. And 44% of those same executives admitted that generative AI is the most difficult type of AI to measure economically - meaning they can’t even quantify the thing they’re betting on.
Oxford Economics put it more bluntly: “Firms don’t appear to be replacing workers with AI on a significant scale.” Their analysis found that if companies were genuinely substituting AI for labor, productivity should be accelerating. Instead, productivity growth has decelerated, following normal cyclical patterns that have nothing to do with artificial intelligence.
The AI-attributed layoffs, Oxford concluded, look more like “dressing up layoffs as good news rather than bad.”
Why Companies Do It Anyway
Peter Cappelli, a management professor at the Wharton School, offered a succinct explanation: companies announce layoffs they expect AI will handle, “but it hasn’t happened yet. They’re just hoping. And they’re saying it because that’s what they think investors want to hear.”
The incentive structure is straightforward. Tell the market you’re cutting headcount because of AI and your stock gets a boost from the efficiency narrative. Tell the market you’re cutting headcount because you over-hired during the pandemic or because your business is struggling, and your stock takes a hit. The label matters more than the reality.
Challenger, Gray & Christmas, the outplacement research firm, noted that “companies are increasingly replacing routine tasks with AI-driven solutions” - but also that AI’s actual effect on the broader labor market remains limited.
The term for this has entered the business vocabulary: AI-washing. Just as companies “greenwashed” environmental credentials, they’re now AI-washing workforce decisions. A Forrester investigation found that many organizations invoking AI to justify layoffs lack the mature, vetted applications that would realistically replace the eliminated roles.
The Klarna Lesson
No company illustrates the AI-washing cycle more clearly than Klarna.
The Swedish fintech giant eliminated roughly 700 positions between 2022 and 2024, mostly in customer service, replacing them with an OpenAI-powered assistant. CEO Sebastian Siemiatkowski loudly proclaimed that AI was doing the work of 700 people. At its peak, Klarna claimed AI handled two-thirds to three-quarters of all customer interactions. It was the poster child for AI-driven workforce transformation.
Then Klarna started rehiring.
Siemiatkowski admitted the company “went too far,” acknowledging that the focus on efficiency and cost “ultimately reduced the quality” of the company’s offerings and “eroded trust with customers.” Klarna is now piloting an Uber-style model to bring human agents back into customer service, targeting students, rural workers, and - in a particularly grim irony - dedicated Klarna users.
Klarna is not alone. Research cited by HR Executive found that 55% of organizations that executed AI-driven layoffs now regret it. Many are quietly rehiring for the same roles they eliminated.
The Body Count
Meanwhile, the layoffs are real. The people losing their jobs aren’t hypothetical.
Here’s a partial list of companies that cited AI in recent workforce reductions:
- Amazon: 16,000 jobs. CEO said fewer white-collar roles needed as the company invests in AI agents.
- Pinterest: 15% of workforce. Redirecting resources toward AI systems.
- Dow: ~4,500 jobs. Ramping up AI and automation.
- HP: 4,000–6,000 employees. Expected $1 billion in savings by 2028.
- Workday: ~1,750 jobs. “Aligning resources with customer needs.”
- Indeed/Glassdoor: ~1,300 combined. Must “adapt accordingly” to AI changes.
- CrowdStrike: ~500 positions. “Leaning into AI” amid market changes.
- Chegg: 45% of workforce. Facing “new realities of AI.”
Of the 55,000 AI-attributed job losses in 2025, 51,000 were in tech, concentrated in California and Washington. But these numbers are a fraction of the total: job losses from “market and economic conditions” were four times larger at 245,000. The AI layoffs get the headlines; the ordinary layoffs don’t.
The Real Risk
There’s an irony buried in all of this. AI will eventually transform work. Early evidence shows 10-15% productivity gains in some roles, particularly programming. But the Harvard Business Review researchers note that AI typically automates “specific tasks and not entire jobs.” The gap between automating a task and eliminating a position is enormous.
The article recalls Geoffrey Hinton’s 2016 prediction that AI would outperform radiologists within five years. A decade later, radiologists remain scarce and fully employed. AI didn’t replace them - it became one more tool in their workflow.
By firing workers now based on promises about AI’s future, companies create several problems at once. They lose institutional knowledge that’s expensive to rebuild. They eliminate the entry-level pipeline that develops future senior talent - Forrester found that new graduates and early-career professionals are disproportionately affected. They generate employee cynicism and public distrust of AI that makes genuine adoption harder later. And as Klarna demonstrated, they may end up rehiring the same people at higher cost.
What This Means
The AI layoff wave is real in impact and largely fake in causation. Tens of thousands of people have lost their jobs, but the technology that supposedly justifies those losses either hasn’t been implemented, can’t be measured, or - as 55% of companies have discovered - doesn’t actually work well enough to replace human workers.
What’s actually happening is simpler and older: companies are cutting costs and using the hottest technology narrative available to make it sound like strategy instead of struggle. A decade ago, they might have said “digital transformation.” Two decades ago, “outsourcing.” Today it’s AI. The jobs disappear either way, but the press release sounds better.
The workers getting cut deserve to know the truth. And investors betting on AI-driven efficiency gains should probably ask the 98% of executives who can’t point to actual results: what exactly are you cutting for?