Why do workers hide productivity gains from AI use?

Anthropic built an AI-powered interview tool, Anthropic Interviewer, "powered by Claude," and used it to run 1,250 interviews with professionals: the general workforce (N=1,000), scientists (N=125), and creatives (N=125). Among the general workforce, 86% reported that AI "saves them time" and 65% said they were "satisfied with the role AI plays in their work," yet 69% "mentioned the social stigma that can come with using AI tools at work" — one fact-checker told the interviewer, "A colleague recently said they hate AI and I just said nothing. I don't tell anyone my process because I know how a lot of people feel about AI." Creatives reported similar productivity (97% said AI saved them time, 68% said it increased quality) alongside comparable concealment: 70% "mentioned trying to manage peer judgment around AI use," with a map artist saying "I don't want my brand and my business image to be so heavily tied to AI and the stigma that surrounds it."

The source frames this as a split between productivity and workplace identity. General-workforce interviewees "want to preserve tasks that define their professional identity while delegating routine work to AI," envisioning "futures where routine tasks are automated and their role shifts to overseeing AI systems." Alongside stigma, 55% of the workforce sample "expressed anxiety about AI's impact on their future" (versus 41% who felt "secure" and that human skills are "irreplaceable"); of the anxious group, a quarter said they "set boundaries around AI use" and a quarter said they "adapted their workplace roles," taking on additional or more specialized responsibilities. For creatives the same anxiety centers on economic displacement and creative identity — a creative director says plainly, "I fully understand that my gain is another creative's loss" — while all 125 creative participants said they wanted to "remain in control" of their creative outputs even though several admitted, in the source's words, that this boundary "proved unstable in practice," with AI driving a majority of some decisions.

This differs from Does AI assistance erode the skills needed to oversee it? in population and in the kind of self-protective behavior reported — that note's internal engineers describe a skills-erosion worry around delegation, while this external sample of professionals and creatives reports actively hiding or downplaying AI use itself, independent of how much work is delegated. It also echoes the self-report divergence flagged in Can self-ratings replace objective performance scores for AI competence?: the same interview sample described AI's role as 65% augmentative and 35% automative, a split the source itself notes is "much more even" (47%/49%) in Anthropic's own measured Claude usage data — though the excerpt breaks off before giving its explanation for the gap. The workforce/creative split into augmentation versus automation also parallels, without matching, the copilot/workflow-agent modalities used to sort adoption in How are national lab staff actually using generative AI?.

The sample was recruited through crowdworker platforms rather than Claude's general user base, interviewed by an AI tool whose reliability as an interviewer the source does not independently validate, and the stigma and anxiety figures are self-reported perceptions, not observed workplace behavior — so the finding establishes that these professionals say they conceal and worry about AI use, not how concealment actually affects their work, their colleagues' trust, or their career outcomes. If the self-report gap on augmentation versus automation is any guide, the true extent or effects of concealment could diverge further from what interviewees report.