I use ChatGPT every day. Am I getting dumber? (ioplus.nl)

In a What if... column, Elcke Vels argues that widespread daily use of generative assistants like ChatGPT, Grok and Claude may be eroding our cognitive skills and proposes a social experiment: one ChatGPT‑free day per week. Drawing on the neurological “use it or lose it” principle and recent studies showing diminished brain activity, language and reasoning in students who over‑rely on AI, Vels frames the problem as both personal (writer’s block, duller sentence craft) and systemic. She notes the scale—about 800 million weekly ChatGPT users—and shares a cautionary anecdote where a colleague blindly followed incorrect AI instructions during a server configuration and caused a crash, illustrating real safety and reliability risks. For the AI/ML community this raises clear design and policy implications: automation complacency and skill atrophy need mitigation. Technical responses could include better uncertainty quantification and provenance, UI nudges that require user verification or chain‑of‑thought engagement, adaptive modes that train rather than replace skills, and research into long‑term cognitive impacts. Vels’ proposed AI‑free day is a behavioral intervention that highlights the tradeoff between productivity gains and human expertise preservation—an invitation for engineers, educators and researchers to build systems that augment without outsourcing cognition.