Neuroscientist‑author Dr. S. Roe argues that generative AI erodes deep thinking by offering fluent answers that feel effortless. In classroom trials, EEG data showed most students’ gamma oscillations, a marker of cognitive effort, collapsed after AI use, while a minority stayed high by challenging the machine.

Roe cites a 2019 Harvard study revealing that students who struggled with problems learned more, yet reported feeling less educated. The EU’s AI Act demands “human oversight” for high‑risk systems, but the law treats the human role as a checkbox, risking automation bias as operators rubber‑stamp AI output.

Roe proposes a Hybrid Intelligence Index to gauge whether human‑AI collaboration sharpens or dulls users. When AI was prompted to ask questions instead of give answers, active engagement doubled. In a labor market leaning on AI products, firms that let employees argue with machines will outpace those that rely on passive fluency.

The study signals that companies must design AI tools with friction to preserve human cognition, or face a workforce increasingly dulled by effortless automation.