A RAND survey of 1,214 young people found that 67 per cent agreed, with a 25-point gap between those who do not use AI and those who use it the most.
The RAND Corporation surveyed 1,214 young people in December 2025; the results were published on August 17, 2026. When asked that intensive use of AI impairs critical thinking, 67% agree or strongly agree. The breakdown by level is surprisingly flat: 68% in middle school, 64% in high school, 69% in higher education.
The interesting gap lies elsewhere. Among those who do not use AI, 78% agree with the statement; among those who use it in three or more courses, only 53%. Either the practice is reassuring, or it makes us less lucid about its own effects — the survey does not allow us to decide.
The same research report recalls that a randomized trial conducted with more than 6,000 middle school students in Tennessee on the learning of fractions produced a gain of about three percentage points for an AI tutor associated with a mastery pedagogy. A real gain, but limited: on less familiar exercises, the group accompanied by the AI did not do better than the others.
Read the full article on the original source
Go to the source →