IBM study: 3 out of 4 employees say AI erodes critical thinking

IBM's report arrives as multiple recent studies raise alarm about AI's effects on human cognition. The University of Bath said earlier this year that as workers use AI for various processes rather than engaging directly, they may lose familiarity and expertise. Cangrade's 2026 report found younger workers had below-average critical thinking, attention to detail, and problem-solving skills. A multinational group of researchers warned in August that managers may experience "epistemic de-skilling" from overreliance on generative AI, leaving them less able to build moral insights and contextual understanding.

The IBM study adds a workplace-specific angle: 80% of CHROs surveyed said AI adoption creates "invisible" work such as validating recommendations and fixing errors. Nearly half of employees agreed AI added to their workloads, yet significantly fewer said managing the tech's outputs would be an important future skill. IBM's recommended fixes—rotations, supervised decision-making, coaching, earlier customer exposure—point to a view that judgment must be deliberately rebuilt.

The stakes may hinge on whether employers treat these findings as a training problem or a workflow design problem. If judgment is embedded in work, 62% of CHROs reported increased employee confidence in AI-enabled decisions; without it, 57% reported declining confidence. For HR leaders, the test is whether early-career programs can offset the erosion employees themselves are reporting. This pattern is likely to influence how companies structure AI rollouts and where they invest in skills development.