THE FACTS
In the past several months, a coordinated chorus has risen from the executive suites of some of tech’s biggest names. Atlassian announced it would cut 10% of its workforce. Meta signaled it may shed 20% of staff. Amazon has joined the parade. The stated reason, delivered with the solemn gravity of a press release written by lawyers: AI.
The narrative is consistent to the point of suspicion. AI is making human labor replaceable. Responsible management demands adjustment. The algorithm has spoken.
Except: the algorithm didn’t speak. A human typed the press release. A human scheduled the all-hands. A human signed off on the severance packages. And according to a recent analysis by University of Sydney Professor Uri Gal, the economic evidence for AI-driven displacement at the scale these companies are claiming is, at best, a stretch.
THE ALIBI
Here is what the data actually shows. A 2025 Goldman Sachs report estimated that if AI were deployed across the entire US economy for everything it currently can do, roughly 2.5% of US employment would be at risk. Not trivial—but nowhere close to the existential reckoning these companies are invoking to explain cuts affecting thousands of named individuals.
More damning: workers in AI-exposed occupations are, right now, no more likely to lose their jobs, see their hours cut, or earn lower wages than workers in any other field. The machines are coming, sure. They are not here yet—certainly not in the numbers these layoff announcements imply.
So what is actually going on? Three things, running simultaneously and rarely acknowledged together: post-pandemic over-hiring that inflated headcounts during the online-services boom; investor pressure to show improved profit margins; and the simple, powerful fact that framing a workforce reduction as an “AI efficiency play” sends a very different signal to Wall Street than admitting you hired too many people when times were good.
THE SMOKING GUN
Meta is the case file that breaks open the whole story. The company is reportedly planning to cut up to 20% of its workforce while simultaneously committing $600 billion to building AI data centers and recruiting top AI researchers. Read that again.
The workers being laid off are not being replaced by AI today. They are being laid off to fund the AI bet. They are not casualties of automation—they are the budget line. The algorithm is the destination, not the driver. The humans making that call are the executives who over-extended during the pandemic years and now need someone—or something—to blame for the correction.
Since ChatGPT launched, AI-related stocks have accounted for approximately 75% of S&P 500 returns. That is a financial incentive so large it distorts corporate communications entirely. A layoff framed as an AI pivot is an investor relations exercise. The actual humans losing their jobs are props in a narrative designed for an audience of fund managers.
THE VERDICT
Atlassian, Meta, and Amazon have not been done wrong by a sentient algorithm that decided human workers were redundant. They have been done wrong by their own executives, who over-hired, overpromised, and are now using a technology trend as a laundry bag for decisions that would look bad on their own.
The AI didn’t lay anyone off. It never does. It doesn’t have the authority, the legal exposure, or the stock options. The humans who made these calls do. They just prefer you don’t think about that.
