Companies Are Firing Workers Because They Expect AI to Replace Them—Not Because It Has

Companies aren't firing workers because AI replaced them. They're firing workers because they expect AI will replace them.

Harvard Business Review surveyed 1,006 executives in December 2025. The finding: AI is behind layoffs, but "these are almost completely in anticipation of AI's impact."

Translation: the jobs aren't being lost to robots. They're being lost to PowerPoint decks about robots.

The Gap Nobody's Measuring

Executives expect big efficiency gains. They haven't realized them yet. But they're cutting staff anyway.

Block (Jack Dorsey's company) just laid off thousands. Dorsey cited AI. Workers pushed back. "You can't really AI that," multiple employees told The Guardian. Current AI tools can't actually do their jobs at this scale.

They got laid off anyway.

Hunt Scanlon Media researched what's actually happening: "Most AI systems are simply not robust enough to replace human workers effectively, making legitimate AI-driven layoffs extremely rare."

But layoffs citing AI? Those are everywhere.

The Numbers Tell the Story

Goldman Sachs predicts unemployment will rise by 0.5 percentage points during the "AI transition period." Half a percent. Manageable, they say.

But transition to what? If AI isn't actually replacing the work yet, what are companies transitioning toward?

Oxford Economics looked at the data. AI-related cuts represent just 4.5% of total job losses. The other 95.5%? Financial pressures, market slowdowns, pandemic hiring adjustments.

AI is the convenient excuse. Not the cause.

The Reality Check

Developers using AI tools took 19% longer to complete tasks in one study. They believed AI had sped them up by 20%.

The gap between perception and reality is massive.

Fortune summarized it: "AI use remains experimental in nature and isn't yet replacing workers on a major scale."

But 37% of companies expect to have replaced jobs with AI by the end of 2026. Nearly a third say they already have.

How do you fire someone for a job AI will do—when AI can't do the job yet?

Who Pays

Goldman expects 11 million jobs displaced over the long term. Not today. Not this year. Eventually.

But "eventually" doesn't pay rent. The workers getting cut now aren't losing their jobs because AI proved it could do the work better. They're losing jobs because someone in a boardroom ran a forecast.

The forecast might be right. It might be wrong. Either way, the person who got laid off is gone before anyone checks.

What This Means

AI job displacement is currently a confidence game. Not a technology story.

The hype is costing real jobs right now. The capability lags years behind.

Companies are betting they can replace workers before proving the replacement works. Some will be right. Many won't. The people who got cut in the meantime won't be there to see which.

That's the gap between anticipation and performance. And it's wider than anyone's admitting.