When They Say Layoffs, They Actually Mean Layoffs: 89% of AI Replacing Employees Is Just a Gamble

The gamble behind "AI layoffs"

A recent Harvard Business Review study led by Thomas Davenport and Laks Srinivasan reportedly found that 89% of firms that have cut staff in the name of AI did so before the technology had demonstrably replaced the work — in short, they were betting on potential, not proven capability. The story has real teeth in China after it has been reported that NetEase (网易) accelerated a planned round of outsourcing cuts because the company was “pleased with efficiency gains from AI,” a move that reportedly pushed thousands of contract workers into immediate risk. Why does this matter? Because when capital markets and investor narratives reward the promise of AI, executives face enormous pressure to show action — even if that action is premature.

The decision chain is clear and blunt. Public companies announce AI-driven cost cuts and markets cheer. CEOs feel compelled to respond, middle managers translate targets into headcount reductions, and the first to go are the most flexible work arrangements: outsourced teams, contractors, junior and support roles. That pattern helps explain why NetEase’s cuts targeted external suppliers first. But there’s an important difference between a tool that raises productivity by 20% and one that can fully substitute a planner, artist or engineer — and that distinction is often ignored in the rush to meet short-term market expectations.

Real-world blowback and wider stakes

There are cautionary precedents. The Commonwealth Bank of Australia replaced dozens of customer-service staff with AI voice bots and later apologized and rehired the majority after service problems and rising call volumes. Careerminds has reportedly found many firms re-hiring a sizable fraction of roles they initially cut. In China the NetEase episode has already rippled across industry discussion boards — from Maimai to Weibo and Zhihu — amplifying anxiety among outsourced game workers and prompting a rapid corporate PR response in which NetEase denied wholesale AI-driven firings while acknowledging some reductions in basic, outsourced roles.

This is more than a management misstep. It is a governance and labor-stability issue in a geopolitical moment when governments, investors and companies are racing to capture AI advantage amid trade frictions and export controls on advanced chips. Who pays when companies “bet” wrong — the laid-off worker, the returning rehired staff, or long-term investor trust? If layoffs are motivated by sentiment rather than evidence, the hidden costs — eroded morale, damaged employer brand, and expensive reversals — may well outweigh short-term savings. Should firms stop signaling AI intent to markets? Or should they be held to higher evidentiary standards before using AI as a reason to change lives? The debate is only beginning.