I heard a CEO get asked what AI would produce for his company. He responded quickly: “20% over five years.” No benchmark, no plan, no doubt. The pressure to deliver on that promise sooner is where the layoffs come from. And the layoffs are not paying off.
As Gartner Distinguished VP Analyst Helen Poitevin told Fortune: “Many CEOs turn to layoffs to demonstrate quick AI returns; however, this disposition is misplaced. Workforce reductions may create budget room, but they do not create return.”
The technology is not failing. The story is. Executives are promising outcomes before the operational conditions exist to produce them — and that has consequences.
Five publications in the last six months have circled the same conclusion. In January, Harvard Business Review, surveying 1,006 global executives, found companies are laying off workers based on what they hope AI will do, not what it is doing. In December, MIT Technology Review Insights surveyed 500 business leaders: 83% believe a culture of psychological safety measurably improves AI initiative success. Most are not building it. In May, a Gartner study found that 80% of companies piloting or deploying AI have reduced their workforce, and that reduction rates were nearly identical at companies reporting strong returns and those reporting modest gains or losses. The cuts did not differentiate the winners from the losers. In between, ADP executive Andrew Hallinson, writing in CIO, named the mechanism: an “aversion tax,” the cash value lost to human friction every time a company deploys a tool the people on the other end do not trust. And in February, HBR returned with 3M’s Jayshree Seth describing how AI tools meant to enhance productivity are creating predictable patterns of dysfunction.
Five publications. Five different methodologies. Five different flavors of the same conclusion.
The layoffs became the deliverable because the actual transformation work is the hard work. The technology is easy. Transformation is slow, unglamorous, and uncomfortable to fund. Cultural infrastructure does not show up in the P&L the same quarter you invest in it. A headcount reduction does.
Psychological safety — which 83% of executives in the MIT Technology Review survey identified as measurably improving AI initiative outcomes — is eroding inside those same initiatives. When workers are unsure whether the technology is meant to augment them or replace them, adoption slows, experimentation stops, and the conditions under which AI actually returns value disappear. Companies that succeed with AI tend not to treat labor reduction as proof of innovation. They are doing the harder thing — building the conditions under which AI can pay off.
Boards did not ask for layoffs. They asked for an AI strategy. The layoffs were what came back.
A director asks the CEO what the company is doing about AI. The CEO, who has been asked the same question by every analyst on the last earnings call, needs an answer that sounds like leadership. So the CEO produces a number. A percentage. A timeline. Something legible. And then has to deliver against it.
The actual work of producing AI returns — redesigning workflows, retraining managers, rebuilding data foundations — does not show up in the quarter you start it. It shows up two or three quarters later, if you are disciplined. None of that fits inside the timeframe the CEO just committed to.
A layoff does. A layoff shows up in operating margin the same quarter you announce it. It is the only AI-adjacent move that produces a number a board can see immediately.
There is also evidence that many layoffs announced as AI-driven were driven by something else. Goldman Sachs analysts reported in late 2025 that companies announcing layoffs carried higher debt, higher capital expenditure, and lower profit growth than peers, suggesting the cuts were responses to financial distress rather than AI efficiency gains. Investors began penalizing layoff announcements rather than rewarding them.
For boards facing this dynamic, the question worth asking is not “what is our AI strategy.” That phrasing produces a legible number and the layoff that follows it. A more useful question: “What would have to be true about our people, our data, and our culture for our AI investment to produce the returns we are forecasting?”
That question cannot be answered with a hiring announcement or a press release. It can only be answered with the kind of work most companies have been treating as optional.
The signal is becoming harder to ignore. The companies seeing AI returns are not the ones with the most aggressive workforce reductions. They are the ones whose people are not afraid to use the tools.
By Julie Averill
Source: forbes.com
The modern hiring funnel is under unprecedented pressure. According to recent research from the recruiting platform Ashby, the number of applications per hire has tripled since 2021. Today, an open role receives an average of more than 300 applications, a massive jump that has forced human resources teams to change how they evaluate talent.
The City of Cleveland, in partnership with the Site Readiness for Good Jobs Fund, today announced The Midline, a 350-plus acre redevelopment initiative that will convert long-vacant industrial land on the city’s near east side into a connected employment district, public greenway and community asset linking major job centers.
We cut middle management. AI is cutting entry-level work. Nobody planned it that way, but the result is the same: fewer people developing others, and fewer people being developed.
