AI Didn't Take Their Jobs
Jack Dorsey fired 4,000 people in February. That's 40% of Block's entire workforce, gone in a single announcement. His explanation to shareholders was blunt: "intelligence tools" had made their jobs unnecessary. A "significantly smaller team, using the tools we're building, can do more and do it better."
Bold claim. Here's the problem: it's almost certainly not true.
A National Bureau of Economic Research study published the same month surveyed nearly 6,000 executives across the US, UK, Germany, and Australia. The finding? Nine out of ten reported that AI had zero impact on employment at their organizations over the past three years. Not a small impact. Not a mixed impact. Zero.
So either Dorsey's company is decades ahead of everyone else on the planet, or something else is going on.
The numbers don't add up
Let's look at what Block actually did before the layoffs. Between 2019 and 2022, the company more than tripled its headcount, ballooning from 3,835 employees to 12,430. Its stock had dropped 40% since early 2025. The business was bloated and under pressure.
That's not an AI story. That's a correction. And Dorsey isn't alone.
Oracle announced plans to cut 20,000 to 30,000 employees to free up $8 to $10 billion in cash. The company's capital expenditure jumped from $6.9 billion in fiscal 2024 to $21.2 billion in fiscal 2025. It's guiding to $50 billion for the current year. Oracle isn't replacing workers with AI. It's firing workers to pay for AI infrastructure.
Meta reportedly plans to cut 20% of its 79,000-person workforce. Amazon confirmed 16,000 job cuts. In each case, the public narrative centers on AI. In each case, the financial reality tells a different story.
Why AI makes such a convenient scapegoat
Here's where it gets interesting. A Resume.org survey of hiring managers found that 59% of companies emphasize AI when explaining layoffs because, and I'm quoting directly, it "sounds strategic and forward-looking." Nearly six in ten companies are choosing AI as their explanation not because it's accurate. It just plays better with investors and the press than "we hired too many people and now we're cutting costs."
Only 9% of those same companies say AI has fully replaced any roles. Nine percent.
Think about that gap. Sixty percent of companies frame their layoffs around AI. Nine percent have actually replaced anyone with it. That's not a technology trend. That's a marketing strategy.
Amazon's CEO Andy Jassy said the quiet part loud. He initially credited AI for the company's workforce reduction, then walked it back, admitting the cuts were "not really AI-driven, not right now at least." Give him credit for the correction. Most CEOs won't make one.
We've seen this movie before
If you work in privacy, this pattern looks familiar. For years, companies slapped cookie banners on their websites and called themselves "GDPR compliant." They did the minimum visible thing that made them look responsible. The substance didn't match the signal.
AI-washing works the same way. Companies announce AI-driven restructuring. Their stock pops. Analysts write approving notes about "operational efficiency." And nobody asks whether the AI actually does what the fired humans used to do.
The term "AI-washing" has earned its place next to "greenwashing" in the corporate accountability dictionary. Both describe the same behavior: using a socially approved narrative to disguise a financially motivated decision.
Harvard Business Review put it plainly in January: companies are laying off workers because of AI's potential, not its performance. They're making bets on what AI might do someday, then firing people today.
The reversal is already starting
Forrester's 2026 future-of-work report predicts that over half of AI-attributed layoffs will be "quietly reversed" as companies realize they moved too fast. The research firm found that 55% of employers already regret laying off workers due to AI. More investment leaders now expect headcount increases (57%) than decreases (15%) over the next year.
That's not a rounding error. That's a full-blown about-face happening in real time.
And it makes sense. The same NBER study that found zero employment impact also found that executives use AI an average of 1.5 hours per week. That's it. Ninety minutes. You can't credibly claim AI replaced 4,000 jobs when your leadership team barely uses it between Monday and Friday.
What this means for companies making real decisions
Here's what I'd tell any executive watching this unfold.
First, don't follow the lemmings. Dorsey predicted that within a year, most companies will reach the same conclusion and make similar cuts. That prediction is going to age like milk. The companies that slash headcount to chase an AI narrative they can't deliver on will spend the next two years scrambling to rehire. Probably at higher salaries. Definitely with worse institutional knowledge.
Second, invest in AI where it actually works. The NBER data doesn't mean AI is useless. It means most companies haven't figured out how to deploy it yet. The 10% that did see measurable impact are the ones worth studying. What are they doing differently? They're identifying specific workflows, building with purpose, and measuring outcomes. Not writing shareholder letters about "intelligence tools" while gutting their teams.
Third, be honest. With your employees, your investors, and yourself. If you're cutting costs, say you're cutting costs. Blaming AI for a financial decision doesn't just mislead your stakeholders. It poisons the well for the people at your company who are actually trying to build useful AI applications. When the layoff narrative is "AI replaced you," every remaining employee starts treating AI as a threat instead of a tool.
The real AI workforce story is quieter and more interesting
The NBER study found something else that barely made headlines. Despite seeing no current impact, those same executives expect AI to boost productivity by 1.4% and reduce employment by just 0.7% over the next three years.
That's a real prediction from people running real businesses. It's modest. It's measured. And it's completely at odds with the "AI is eating the world" narrative that tech CEOs need to justify their restructuring decisions.
The actual story of AI and work is going to be gradual, uneven, and specific to individual workflows. It's not going to look like a CEO firing 40% of his company in one memo. It's going to look like a financial analyst spending two fewer hours on a report. A developer shipping a feature a day faster. A compliance team automating a review that used to take a week.
That story doesn't make headlines. But it's the one that matters.
The companies that get this right won't be the ones who fired first and figured it out later. They'll be the ones who stayed honest about what AI can actually do today, while building toward what it'll do tomorrow.