When IBM announced in 2023 that it would pause hiring for roughly 7,800 roles it believed AI could eventually handle, the move was treated as a bellwether. Dropbox, Chegg, and Duolingo followed with their own AI-linked cuts. The implicit promise was straightforward: fewer workers plus smarter software equals better results.
Three years later, the data tells a different story. A working paper from the Stanford Institute for Economic Policy Research, led by economist Nicholas Bloom, found that roughly 80% of firms reported no meaningful impact on employment or productivity from their AI deployments over the past three years. A separate Gartner enterprise survey found no statistical correlation between AI-related workforce reductions and improved financial returns. Together, the two findings undercut the central justification companies have used to tie layoffs to automation.
What the Stanford data actually shows
The Stanford working paper, titled “Firm Data on AI,” draws on firm-reported AI impacts and expectations across a broad sample of U.S. companies. Its headline number is blunt: about four in five firms said their AI tools had not moved the needle on headcount or output in any measurable way. That figure is striking not because AI is useless, but because it directly contradicts the boardroom assumption that deploying the technology rapidly displaces workers or generates efficiency gains large enough to register in company-level results.
The paper does forecast a shift. Firms expect AI to drive modest employment reductions in the years ahead. But those reductions have not materialized at scale, creating a gap between corporate planning and corporate experience. Companies are restructuring around outcomes their own data does not yet support.
That pattern is not new. Technology forecasts have a long history of overshooting near-term impact. Gartner built an entire analytical framework around it (the Hype Cycle), and the economist Roy Amara observed decades ago that people tend to overestimate a technology’s short-term effects while underestimating its long-term ones. What makes the current AI wave different is the speed at which workforce decisions are outpacing the evidence.
The Gartner finding and its limits
A Gartner enterprise survey, drawn from the firm’s ongoing research program and widely cited in technology strategy circles, adds a second dimension. Its analysts found no statistical link between AI-related layoffs and improved financial performance. The finding has been consistent across secondary reporting. However, the specific survey name, date, sample size, and full methodology have not been made publicly available, so the conclusion should be treated as a credible but not independently verifiable signal from a respected research firm rather than a result outside analysts can replicate today.
Both studies share a definitional challenge. “AI deployment” covers an enormous range of activity, from a narrow customer-service chatbot handling password resets to a full-scale generative AI integration across product development and supply chain management. A logistics company automating route optimization faces a fundamentally different calculus than a media company using large language models to draft articles. Aggregate data can blur those distinctions, and neither study offers the granular, company-level case studies that would let readers trace a specific firm’s AI timeline from adoption to outcome.
The remaining 20% of firms that did report changes also deserve scrutiny. That group could include companies that saw genuine productivity gains, companies that experienced disruption without clear benefits, or both. The Stanford paper does not break out that subset in detail, which means the most interesting question for practitioners (what separates the firms where AI worked from the firms where it didn’t?) remains largely unanswered.
What boards and executives should be asking in mid-2026
For business leaders weighing AI investments between May and June 2026, the practical takeaway is pointed. If four out of five firms saw no employment or productivity shift from AI over three years, then aggressive headcount cuts justified solely by automation look more like a bet than a data-backed strategy. The burden of proof should sit with the executives proposing cuts, not with the workers absorbing them.
Investors and board members can use these findings to sharpen their questions. In any AI-and-workforce discussion, the specifics matter: Has the company measured task-level time savings from its AI tools? How do those savings translate into revenue or margin improvement? If productivity effects are still small or uncertain, what is the business case for cutting jobs now rather than after results materialize? Without clear answers, AI-linked layoffs may signal short-term cost discipline more than long-term value creation.
There is countervailing evidence worth weighing. Studies of specific tools, such as GitHub’s research on Copilot, have shown measurable productivity gains for individual tasks like code completion. McKinsey’s Global Institute has projected that generative AI could eventually add trillions in economic value across industries. The Stanford and Gartner data do not refute those projections. They show that, for most firms, the gains have not arrived yet at a scale that justifies the workforce decisions already being made.
Workers and policymakers cannot afford to wait for clearer data
For workers, the current evidence offers a complicated kind of reassurance. The data does not support a narrative of sudden, widespread job destruction from AI. But it does show that companies are acting on expectations of future displacement, even when their own track record with the technology is underwhelming. The layoffs are real even if the productivity rationale behind them is not.
Policymakers face a version of the same problem. Waiting for definitive long-term studies before investing in retraining programs, workforce transition support, or transparency requirements means falling behind decisions that companies are making today. The most useful interventions between May and June 2026 are structural: funding skills programs that are not tied to a single technology, building internal mobility pathways within large employers, and requiring companies to disclose how AI adoption is reshaping specific roles rather than issuing vague statements about “efficiency.”
The companies that navigate this period well will not necessarily be the ones that cut fastest. They will be the ones that tracked what AI actually changed in their operations, measured it honestly, and refused to use a technology’s potential as a blanket justification for decisions its track record does not yet support. As of mid-2026, that track record is clear: for most firms, the transformation is still a forecast, not a fact.
David M. Keller is a finance writer based in Columbus, Ohio, covering personal finance and consumer-focused economic topics. He earned his degree in journalism from Ohio University and began his career reporting on local business and economic trends for a regional media outlet. Since then, he has contributed to a variety of online publications, focusing on clear, practical coverage of topics such as cost of living, debt, and everyday financial decision-making.
