Companies are spending billions on AI, firing workers to pay for it, and getting nothing in return. The research now explains why, and the mechanism is one most executives have ignored.
An Atlanta Federal Reserve study found that roughly 90% of executives believe AI has not yet boosted productivity at their companies. That number alone is striking. What makes it more significant is that the same companies are simultaneously announcing waves of layoffs attributed to AI — and watching the strategy fail.
Research from the University of Pittsburgh, published in The Conversation on August 13, analysed millions of Glassdoor reviews, thousands of corporate financial reports, and hundreds of AI investment and layoff announcements from US public companies over five years. The findings describe a self-defeating cycle: companies invest in AI, cut workers to show short-term savings, and the remaining employees resist the technology so effectively that productivity gains evaporate.
The Pattern Nobody Accounted For
The research identified a clear correlation. As AI investment announcements rise, so do layoff announcements attributed to AI. Some companies began laying off employees before committing capital to AI, using headcount reduction to fund future technology spending.
The stock market has not rewarded this approach. When researchers examined market reactions to AI-linked layoff announcements, the average return was close to zero. For more than half of these events, the reaction was negative. Investors, it turns out, are pricing in costs the announcements do not acknowledge.
The hidden cost is employee sentiment. Researchers analysed millions of Glassdoor reviews, focusing on AI-related comments. Those comments were substantially more negative than overall review tone. Job security fears dominated, ahead of inadequate training, limited reskilling opportunities, and doubts about whether AI genuinely improves work.
When companies announce AI-related layoffs, employee sentiment toward AI drops sharply. Workers who have watched colleagues lose jobs to AI — or who fear they are next — actively resist the tools they are being asked to adopt. And employee sentiment toward AI turned out to be one of the strongest predictors of firm productivity when AI is deployed.
What This Means for Workers
The practical implication is counterintuitive. The companies most aggressively cutting workers for AI are the ones least likely to benefit from it. The companies getting returns are the ones training workers instead of firing them.
This aligns with earlier findings. Gartner’s survey of large enterprises found zero correlation between AI-driven layoffs and actual AI returns, while companies investing in upskilling were the ones seeing ROI. A Robert Half study found that nearly 29% of companies that cut staff for AI reopened the same positions, often at salaries 20% to 35% higher than the roles they eliminated.
The replacement math is brutal. Rebuilding a cut role typically costs 1.5 to 2 times the salary initially saved, according to Lee McCabe, founder of private equity firm Claymore Partners. And the returning employee, as McCabe told InformationWeek, “now knows exactly what your loyalty is worth.”
Gallup’s polling tells the same story from the worker’s side. Half of US workers now use AI at work, but 23% at AI-adopting companies say their employer is letting people go. Productivity and fear are rising in parallel, and the fear is winning.
The Reskilling Alternative
The research points toward a different strategy. Companies that reskill existing workers rather than replacing them are the ones seeing measurable productivity gains. The Lloyds Business Barometer, reported by Bloomberg on August 17, surveyed 1,200 UK businesses and found that 58% plan to raise investment in AI over the next year — most of it aimed at training the staff they already have, not hiring replacements.
Ikea offers a case study in the approach. The retailer retrained call centre staff as remote interior design advisers after automating much of their previous work. The resulting service has been widely reported as a business worth around €1.3 billion.
The difference between Ikea’s approach and the layoff-first strategy is not sentimental. It is structural. Workers who are retrained rather than replaced do not resist the technology. Workers who watch colleagues get fired do. The productivity gap between those two outcomes is what the research is measuring.
The NZ Connection
New Zealand’s government is cutting 8,700 public sector jobs while betting on AI to fill the gap. Finance Minister Nicola Willis told Parliament the cuts would save $2.4 billion over four years. The Digitising Government Minister, Paul Goldsmith, said the public service would use “the best technology available” — which, when pressed, he acknowledged meant US-made tools like Claude and Copilot, since no local AI provider operates at that scale.
Critics, including University of Auckland law and technology professor Alexandra Andhov, have questioned whether the savings will materialise. The research from Pittsburgh suggests a more specific concern: if the remaining public servants resist the AI tools they are being asked to adopt — and the public is far from convinced — the productivity gains may not arrive at all.
❓ FAQ
Does this research mean AI doesn’t improve productivity? No. It means that the specific strategy of cutting workers to fund AI adoption tends to fail because it creates resistance. Companies that train existing workers and deploy AI alongside them, rather than instead of them, are the ones seeing gains.
What percentage of executives see no AI productivity gains? About 90%, according to an Atlanta Federal Reserve study cited in the University of Pittsburgh research.
How many companies rehire workers they cut for AI? Nearly 29%, according to Robert Half, often at salaries 20% to 35% higher than the eliminated roles.
What does this mean for workers worried about AI displacement? The research suggests that companies committed to reskilling rather than wholesale replacement are better long-term bets. Worker sentiment toward AI is a strong predictor of productivity, which means companies that alienate their workforce pay a performance penalty.
🔍 THE BOTTOM LINE
The corporate playbook of “invest in AI, cut workers, show savings” is producing the opposite of what it promises. The research is clear: firing workers for AI destroys the conditions AI needs to succeed. The companies getting returns are the ones that train the people they have. The ones cutting are paying twice — once in severance, once in the productivity they never see.
