The Numbers Come First
Nearly a third of American workers — 29% — have admitted to actively sabotaging their company's AI tools, according to an April 2026 survey of 2,400 knowledge workers across the U.S., U.K., and Europe conducted by Writer and Workplace Intelligence, which included 1,200 C-suite executives.
That figure lands with more weight alongside fresh research from Apollo Global Management, where chief economist Torsten Slok and co-author Sania Edlich published findings on July 30 showing that AI's earliest measurable impact on the labor market is not mass layoffs — it is wage compression.
'Analysis of actual Claude usage data shows workers in AI-exposed occupations are experiencing slower wage growth, while employment levels in these occupations remain unchanged, suggesting companies are capturing AI productivity gains through wage compression rather than workforce reduction,' Slok wrote.
What the Data Actually Shows
The Apollo paper is notable for its methodology. Rather than relying on theoretical 'exposure' scores — the dominant tool in AI labor research — Slok and Edlich used observed usage data from Anthropic's Economic Index, drawing on actual Claude interaction logs to measure what workers are doing with AI in practice.
The result reframes the entire debate. Employment in AI-exposed occupations is holding steady. Wages in those same occupations are not.
A separate June 2026 survey of 1,005 employed U.S. workers by Software Finder adds ground-level texture. Half of workers described themselves as actively resisting new AI tools. Forty-five percent cited fear of becoming replaceable as their primary reason. Only 16% believe their company is adopting AI for genuine business value rather than hype or competitive pressure.
The pay gap between adopters and resisters is striking on its face: Software Finder reports that workers who resist AI earn roughly 20% less on average than those who embrace it — $65,645 versus $81,526. Slok's research suggests a more complicated dynamic, however, since that gap likely reflects who tends to adopt early — managers and higher earners with greater job security — rather than proof that adoption itself protects pay.
The two effects, Slok's wage compression across AI-exposed roles and the Software Finder pay gap between adopters and resisters, can coexist.
Other findings from the Software Finder survey reveal the depth of the dysfunction: 13% of workers admitted they have faked AI use, appearing to use a tool while completing the task manually, and only 6% believe their managers accurately understand how often employees actually use the tools that have been rolled out.
The Editorial Read
The market has already voted on what AI means for the median worker, and the verdict is uncomfortable for both the boosters and the doomsayers. The machines are not taking jobs — but they are, by Apollo's reading, allowing employers to capture productivity gains without passing them through to wages. That is a straightforward redistribution of value from labor to capital, and workers appear to sense it even without access to Slok's data.
Free enterprise works when the gains from innovation flow broadly. When productivity growth is systematically captured at the corporate level while the workers generating it see slower pay growth, the sabotage numbers should surprise no one. The policy question worth asking is not how to stop workers from resisting AI — it is whether the regulatory and tax environment gives companies sufficient competitive pressure to share the gains rather than pocket them. Right now, the answer is not obvious, and the 29% figure is the market's way of saying so.
