Twenty-one percent of organizations that cut headcount in anticipation of AI did so based on projected capabilities, not demonstrated ones. That number, from a Harvard Business Review survey published in January, is the single most important data point in the 2026 layoff story, and almost nobody is treating it that way. It means roughly one in five companies fired people because of what they believe AI will eventually do. That is not a restructuring. It is a bet, placed with other people's livelihoods.
My argument is simple: "AI restructuring" has become the most effective reputational shield in corporate America, allowing profitable companies to shed workers without facing the investor skepticism, regulatory attention, or public backlash that ordinary cost-cutting would attract.
The Label That Pays for Itself
Consider the incentive structure. A company that announces layoffs because margins are fine but could be better invites questions about executive compensation, buyback priorities, and whether the board is squeezing labor to juice next quarter's earnings call. A company that announces the same layoffs under the banner of "AI transformation" gets a different reception entirely: forward-thinking, tech-savvy, adapting to the future. Wall Street rewards the second story. Pinterest cut 675 people, cited an "AI-forward strategy," and the stock still dropped, but the press coverage framed it as strategic vision rather than cost discipline. Salesforce's Marc Benioff told Bloomberg that AI now handles 30% to 50% of the work. What does that claim actually mean in measurable output? Which tasks? Verified how? Nobody pressed him, because the AI label functions as its own justification.
Block is the most instructive case. Jack Dorsey said the cuts were "not driven by financial difficulty, but by the growing capability of AI tools." Grant him this: some of those productivity gains are real. Code generation tools have genuinely changed the math on certain engineering tasks. But a 40% workforce reduction at a company that was not in financial distress raises a question that the AI narrative conveniently obscures: if the tools are so effective, why did Block need 10,000 people six months ago? Either management badly misjudged headcount for years, or the AI story is doing some heavy lifting to smooth over a correction that has older, less glamorous roots.
Who Gets to Avoid Scrutiny
The deeper problem is structural. When companies attribute layoffs to AI, they place the cause outside their own decision-making. The technology did it. The market shifted. This framing strips agency from executives and strips recourse from workers. A laid-off engineer at WiseTech Global, told that generative AI made her role "increasingly obsolete," has no grievance to file, no mismanagement to point to, no regulatory body to petition. She was simply overtaken by progress. How convenient for WiseTech's balance sheet.
The 80% of AI-linked layoffs that may reflect genuine productivity shifts deserve honest accounting. I am not claiming every company is lying. But the 20% figure, 9,238 jobs explicitly tied to AI out of 45,363, coexists with the Harvard finding that a significant share of cuts are speculative. And the remaining 36,000 layoffs are happening at companies posting record revenues, in an environment where "AI restructuring" provides rhetorical cover that "cost optimization" never could.
So who should act? The SEC. If a public company attributes a material workforce reduction to AI, it should be required to disclose specific, auditable metrics: which workflows were automated, what measurable productivity change occurred, and over what timeline. Dorsey's claim and Benioff's percentage and WiseTech's assertion about obsolescence are, right now, unverifiable statements made to justify decisions affecting thousands of families. The securities disclosure regime already requires companies to substantiate material claims about their financial condition. Why should claims about AI-driven restructuring, which move stock prices and destroy careers in equal measure, face a lower bar?
The technology is real. The productivity gains, in some cases, are real. But a label that lets profitable companies fire workers while looking visionary, that discourages regulatory scrutiny by wrapping layoffs in the language of innovation, that shifts blame from boardrooms to algorithms: that label is doing more work than the AI.
