Microsoft paper warns AI use erodes critical thinking at work

A Microsoft and Carnegie Mellon survey of 319 knowledge workers found 62% of them reported engaging in less critical thinking when they used generative AI, particularly on routine or lower-stakes tasks. The paper, covered by 404 Media, is built on 936 first-hand examples participants volunteered about their own AI use.

The headline framing comes straight from the researchers. "A key irony of automation is that by mechanising routine tasks and leaving exception-handling to the human user, you deprive the user of the routine opportunities to practice their judgement and strengthen their cognitive musculature, leaving them atrophied and unprepared when the exceptions do arise," they write.

The examples the paper surfaces are deadpan rather than alarming. A teacher used DALL-E to make images for a presentation about hand washing. A commodities trader asked ChatGPT to "generate recommendations for new resources and strategies to explore to hone my trading skills." A nurse "verified a ChatGPT-generated educational pamphlet for newly diagnosed diabetic patients."

One finding runs the other way. Workers with more confidence in their own expertise were 27% more likely to critically assess AI output rather than accept it at face value. The authors describe a "shift in cognitive effort as knowledge workers increasingly move from task execution to oversight when using GenAI," and warn that AI "may also reduce critical engagement, particularly in routine or lower-stakes tasks in which users simply rely on AI, raising concerns about long-term reliance and diminished independent problem-solving."

The study is self-reported. Participants told the researchers how much thinking they felt they did, not how well they actually caught errors. It is also a single cross-sectional survey, not a longitudinal measurement of any skill decay over time.