There is a painful irony playing out across thousands of businesses right now. Companies are cutting workers in the name of AI productivity, and in doing so, are actively destroying the conditions that make AI productive.

That is the central finding of new research from Mark Ma at the University of Pittsburgh, published this week via The Conversation. After analysing more than 3,200 firms across six dimensions — AI talent share, hiring patterns, retention, salary premiums, and both employee and executive sentiment — the conclusion is stark: AI-linked layoffs correlate with lower worker AI sentiment, and worker AI sentiment turns out to be the strongest predictor of firm-level productivity gains from AI. Management optimism shows no significant productivity relationship at all.

In other words, it does not matter how bullish your executive team is about AI. What matters is whether the people doing the work actually trust it. And nothing corrodes that trust faster than watching their colleagues get replaced by it.

The Self-Defeating Cycle

The research team drew on data from Glassdoor job postings, corporate earnings calls, layoff announcements, and market data. When they examined stock 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 or essentially flat — a signal that investors themselves are sceptical of the strategy.

The mechanism Ma’s team identified is straightforward: when workers see layoffs attributed to AI investment, job-insecurity concerns spike. That insecurity damages their sentiment toward AI tools. And companies with positive worker AI sentiment consistently show higher productivity than those where employees view AI with suspicion or fear.

This is not a soft management problem. It is a measurable financial one.

More than 122,000 technology workers were cut in 2025, and another 126,000 have been let go in 2026. Many of those cuts were framed publicly as AI-driven efficiency moves. The research suggests a significant portion of those companies are not getting the returns they expected — and the reason lies in the workforce they are left with.

The Gap Between Executive Confidence and Results

A parallel study from the Federal Reserve Bank of Atlanta found that roughly 90% of executives believe AI has not yet improved productivity at their companies. That number should be alarming. Businesses have poured enormous capital into AI tooling while simultaneously telling their workforces that AI is replacing them. Then they wonder why adoption stalls.

Ma’s research offers an explanation. When managers announce AI investment alongside headcount reductions, employees do not interpret the combination as a signal of efficiency. They interpret it as a threat. And threatened workers do not enthusiastically adopt the tools that threaten them.

What This Means for Business

This research matters for any business leader currently thinking about AI-driven workforce restructuring. A few principles worth taking seriously:

Worker sentiment is a leading indicator. Before cutting headcount, it is worth measuring how your team actually feels about the AI tools you have deployed. If sentiment is low, adding more AI capability and removing people will not improve the outcome.

Framing changes everything. The same AI investment, framed as augmentation versus replacement, produces different workforce responses. Companies that position AI as a tool that frees people from tedious work get different adoption curves than those whose messaging implies headcount reduction is the goal.

Sequencing matters. Productivity gains from AI tend to compound over time as workers develop skill and confidence with tools. Layoffs disrupt institutional knowledge and reset the learning curve — often in ways that are invisible to finance teams until the damage is done.

Rehiring is expensive. Separate data from talent firm Robert Half found that nearly 29% of companies that cut staff for AI have reopened the same roles, often at salaries 20 to 35% higher than the eliminated positions paid. The “savings” from the original layoff frequently disappear.

The Smarter Path

Enterprise DNA’s view on this is consistent with the research: AI should be deployed to increase what your existing team can achieve, not to thin the team first and figure out capability later. The companies seeing genuine AI ROI are the ones that upskill their workforce alongside the tooling — not the ones that treat AI as a headcount reduction mechanism.

That is a harder sell to a board looking for quick cost reduction. But it is the one that holds up when you actually look at the data.

If you are a business leader evaluating how to deploy AI without destroying the conditions that make it work, our Omni Advisory service is built precisely for this — helping leadership teams build AI strategies grounded in how your actual workforce can adopt and benefit from AI tools.