Just over half of workers (52%) say they have used AI on a task they believed they could have finished faster without it, according to a survey from JobAssist.
The pressure seems to come from the top. Two in three workers (63%) said their employers encourage or expect AI use for at least some tasks. Separately, 39% said they had highlighted their AI use to managers despite believing it did not help their outcome. Nearly a third (31%) said they kept using AI after deciding it had not improved speed or quality, and 34% said they had avoided giving negative feedback about AI because leadership encourages adoption. The material does not say how those shares relate to the 63%, so we report each as stated.
The risk lands on whoever is counting. "If employees hesitate to report disappointing results, managers may receive an overly favorable account of how well their AI initiatives are working," JobAssist wrote. The company added that firms can measure whether employees use AI, but those measurements cannot show whether the technology improves their work. Overstated adoption, it argued, could send money to the wrong places.
This is an old management problem with a new tool. When a usage count becomes the thing that earns approval, people supply the count. Logins, prompts and activated licenses all reward showing up, not results.
One caveat belongs here: these are self-reported answers. A worker's belief that a task would have gone faster without AI may be right or wrong, so the survey says more about incentives and candor than about hours actually lost. Still, the gap between what workers do and what they say they think is the part a dashboard cannot show. Closing it would take comparing the quality of finished work, which is harder to track than a login.