Don't Ask Whether AI Makes Us Dumber. Ask What Kind of Thinking We Stop Practicing

Don't Ask Whether AI Makes Us Dumber. Ask What Kind of Thinking We Stop Practicing is a research-informed narrative synthesis about generative AI, learning and independent performance. The essay argues that the public question of whether AI makes people "dumber" combines distinct outcomes including memory, effort, persistence, task performance, learning, transfer and creativity. It distinguishes successful performance while assistance is available from the ability to reproduce a process after assistance is withdrawn. The central comparison concerns the structure of help. Ready-made answers can replace the attempt through which a skill develops, while systems that require an attempt, offer graduated hints, provide feedback and gradually withdraw support may preserve or improve learning in defined conditions. The article also examines cognitive offloading, possible deskilling and the trade-off between individual creative ratings and collective diversity. The evidence set contains 22 sources: 15 peer-reviewed publications, five preprints, one official report and one journalistic source used as the starting point for the public debate. The work is not a systematic review, not a meta-analysis, not original empirical research and has not undergone external peer review.