Why critical thinking and arguing with AI is essential for meaningful human-AI collaboration

Summary

A theoretical neuroscientist argues that generative AI produces a feeling of competence without actual substance, and that putting humans in the loop is ineffective if they simply rubber-stamp authoritative-sounding AI output. The author describes an experiment where students wore EEG headsets while working with AI on assignments, revealing that while students looked identical from the outside, their brain activity told very different stories about engagement and critical thinking. The piece calls for teaching people how to argue with and critically interrogate AI systems rather than passively accepting their outputs.