Generative AI was pitched as relief from drudgery. New research says a majority of knowledge workers experience it as the opposite: a second job of checking the first job’s output.
A study from Adaptavist, surveying 2,500 knowledge workers across the UK, US, Canada, Germany, and Spain, finds that 65% regularly feel nostalgic for how work operated before generative AI, and 38% say they would remove the technology entirely if they could. The desire to roll it back skews young: 40% of both Gen Z and millennial respondents would remove GenAI from the world, compared with 32% of Gen X and 29% of boomers, inverting the assumption that younger workers are the most AI enthusiastic cohort.
The mechanism behind the fatigue is specific. Adaptavist calls it a “verification tax”: 42% of workers say they now spend more time checking and correcting AI output than the time the tool saves them. Beyond the extra labor, 36% say they do not understand why they are expected to use AI in their role at all, and half report their performance is being implicitly compared to AI generated work, whether or not that comparison is fair.
For HR leaders, this data reframes adoption metrics that only count logins or prompts sent. A workforce can hit every usage target a rollout sets and still be quietly building resentment toward the tool and the manager who mandated it, echoing what other 2026 research on AI skills gaps has found: the deployment problem and the trust problem are not the same problem, and closing one does not close the other. The fix Adaptavist’s data points toward is not less AI, but clearer answers to why a given task needs it and time budgeted for the verification work AI actually creates.
Source: Adaptavist
