Is AI Making Our Brains Weaker?

Researchers in the U.S. and U.K. published a study in April showing that using AI to solve math or reading problems for just 10 minutes weakened participants’ ability to perform the same tasks without assistance. Those who relied on AI not only scored lower than a control group but also abandoned challenging problems faster, suggesting AI dependency may reduce persistence and impair unaided performance. A separate MIT-led study from last year found that students using ChatGPT to draft essays underperformed in follow-up tasks, including recalling their own writing and producing a new essay without AI help. The findings indicate that offloading cognitive work to large language models may hinder memory retention and deeper engagement with tasks. Experts warn that excessive reliance on AI—known as cognitive offloading—could lead to long-term declines in critical thinking and problem-solving skills. MIT researcher Nataliya Kosmyna noted that skipping foundational tasks like structuring arguments weakens essential cognitive abilities that develop through practice. These concerns arise amid broader declines in U.S. math and reading scores, raising questions about AI’s role in education and workplaces. Policymakers are now weighing the trade-offs as AI tools become ubiquitous in classrooms and professional settings. The studies suggest that while AI can assist, overdependence may undermine human capabilities it was designed to support. Researchers emphasize the need for balanced use to preserve cognitive skills.