AI tools in schools raise fresh alarms over critical thinking loss

This cluster of studies doesn’t change what frontier models can do, but it does probe a deeper question: what happens to human cognition when AI becomes the default interface to knowledge and work? If students lean on ChatGPT‑style systems for drafting, explaining and even thinking through problems, we risk training a generation of operators rather than originators. The research highlighted here suggests that when AI handles the hard parts, the brain simply works less—people retain less of their own writing and invest less effort in reasoning.([caliber.az](https://caliber.az/en/post/ai-use-raises-concerns-over-learning-critical-thinking-studies-find))

For the race to AGI, that’s a double‑edged sword. On one hand, widespread adoption of AI tutors and writing assistants creates huge data streams and commercial pull for ever‑smarter models. On the other, if education systems over‑automate cognition, the pool of people capable of pushing the frontier—designing architectures, interpreting failures, questioning assumptions—could shrink. That would concentrate real understanding inside a small elite of labs and researchers, while the broader workforce becomes increasingly dependent on opaque systems. Long term, that dynamic could accelerate technical progress while eroding the very human expertise we need to keep AGI aligned with our values.

Designing AI that scaffolds thinking instead of replacing it may become just as important a research agenda as squeezing out a few more benchmark points.