Future of Work

AI Layoffs Are Not Paying Off

Are AI layoffs paying off? For most companies, no. The logic seemed sound: cut headcount, let artificial intelligence absorb

By Conor Healy

May 6, 2026

Are AI layoffs paying off? For most companies, no. The logic seemed sound: cut headcount, let artificial intelligence absorb the work, pocket the savings. But the productivity gains have not arrived. A February 2026 study from the National Bureau of Economic Research surveyed nearly 6,000 executives across the US, UK, Germany, and Australia. More than 80% reported no impact from AI on productivity. The technology is everywhere. The results are not.

Oxford Economics put it simply in their January 2026 report: “If AI were already replacing labour at scale, productivity growth should be accelerating. Generally, it isn’t.”

Companies made the cut. The machines did not pick up the slack.

55,000 Jobs, No Payoff

In the first eleven months of 2025, firms cited AI as the reason for nearly 55,000 US job losses. That is real people losing real jobs on the promise that automation would handle it. But when researchers at Deloitte surveyed organisations with AI initiatives, 74% said they wanted the technology to grow revenue. Only 20% had actually seen it happen. A separate PwC survey of 4,500 business leaders found more than half had seen neither increased revenue nor decreased costs from AI adoption.

The UK government ran a controlled trial of Microsoft 365 Copilot across one department. The results, published in September 2025, showed no net productivity gain. Some tasks sped up. Others slowed down. The overall effect was zero.

These are not companies failing to adopt AI. These are companies that adopted it, cut staff expecting gains, and found themselves short-handed with nothing to show for it.

Why It Is Not Working

The NBER study offers a clue. Among executives who use AI, average time spent is 1.5 hours per week. A quarter of respondents do not use it at all. The tools exist. Leadership is not using them. If the people making decisions about AI-driven restructuring barely touch the technology themselves, the gap between expectation and reality makes more sense.

ManpowerGroup’s 2026 Global Talent Barometer found that worker confidence in AI’s usefulness dropped 18% over the past year, even as adoption increased 13%. People are using it more and trusting it less. That is not a recipe for productivity transformation.

Some companies have reversed course. Oxford Economics noted cases where firms brought workers back after service quality collapsed. The AI handled volume but not nuance. Customers noticed. The savings evaporated.

The Solow Warning

This pattern has happened before. In 1987, Nobel laureate Robert Solow observed: “You can see the computer age everywhere but in the productivity statistics.” Businesses had invested heavily in personal computers. Productivity growth dropped from 2.9% to 1.1%. The gains did not arrive until the late 1990s, over a decade later.

Bank of America’s head of US equity strategy, Savita Subramanian, noted that productivity measures “haven’t really improved all that much since 2001.” The lesson is not that technology never delivers. It is that the delivery takes far longer than the hype suggests. Companies that restructure based on projected gains rather than realised ones often find themselves understaffed and underperforming.

Goldman Sachs senior economist Ronnie Walker put it directly: “We still do not find a meaningful relationship between productivity and AI adoption at the economy-wide level.”

Were These Even AI Layoffs?

Oxford Economics argues that many AI layoffs are not really about AI at all. Between 2020 and 2022, tech companies added headcount at historic rates. When growth slowed, they needed to cut. Framing the reduction as AI-driven transformation plays better with investors than admitting over-hiring.

The report states: “We suspect some firms are trying to dress up layoffs as a good news story rather than bad news, such as past over-hiring.”

This creates a problem for companies that took the narrative seriously. If your competitor announced AI layoffs as a strategic move and you followed, you may have cut real capacity based on someone else’s spin. The technology did not replace those workers at your competitor either. They just had better PR.

The Bet That Did Not Pay

The core assumption behind AI layoffs is that the technology can absorb human work at lower cost. For routine, repetitive tasks, there is evidence this works. Data entry, basic customer service queries, document processing. But most jobs are not purely routine. They involve judgment, context, and the ability to handle exceptions. AI struggles with exceptions.

Companies that cut across the board, expecting AI to handle whatever remained, discovered this the hard way. The tasks AI handles well were already the cheapest tasks. The expensive work, the work that requires experience and decision-making, stayed expensive. Now there are fewer people to do it.

For business owners considering AI-driven headcount reductions, the data suggests caution. The productivity gains are not materialising at scale. The companies that cut first are not outperforming. If you are planning layoffs on the assumption that AI will compensate, the evidence says that bet is not paying off.