AI for studying · How Founders Use AI
Does using ChatGPT make you worse at thinking? What does the research actually say?
It depends on how you use it, and the warning signs are real. An MIT Media Lab preprint followed 54 people writing essays: those using ChatGPT showed the weakest brain connectivity and struggled to quote their own essays. A Microsoft survey of 319 knowledge workers found that the more they trusted the AI, the less critical thinking they reported. Anthropic's coding trial found lower mastery when people handed the work over, but not when they asked questions. India's Economic Survey 2025-26 cited the MIT and Microsoft work in warning about over-reliance. These are small or early studies, not proof of lasting harm. The practical reading, as Ethan Mollick puts it: think first, write first, then bring AI in to question your work.
On the AI map, 29 October 2026 AI assistantsResearch and search
Go deeper7 sources
-
Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task
Nataliya Kosmyna and others (MIT Media Lab) · Article · Jun 2025
The EEG study of 54 essay writers, a preprint not yet peer reviewed, which is worth reading past the headlines.
-
The Impact of Generative AI on Critical Thinking: Self-Reported Reductions in Cognitive Effort and Confidence Effects From a Survey of Knowledge Workers
Hao-Ping Lee, Advait Sarkar and others (Microsoft Research and Carnegie Mellon) · Article · Apr 2025
The CHI 2025 survey of 319 knowledge workers linking confidence in AI to less critical thinking.
-
How AI assistance impacts the formation of coding skills
Anthropic · Article · Jan 2026
A randomised trial showing the same tool can help or hurt learning depending on how people asked it.
-
Generative AI Without Guardrails Can Harm Learning: Evidence from High School Mathematics
Hamsa Bastani, Osbert Bastani, Alp Sungu and others (University of Pennsylvania) · Article · May 2025
The field experiment where students did worse once AI was removed, unless the tutor only gave hints.
-
Anthropic Education Report: How university students use Claude
Anthropic · Article · Apr 2025
An analysis of a million student conversations, split between asking for answers and working through problems together.
-
Against "Brain Damage"
Ethan Mollick (One Useful Thing) · Article · Jul 2025
What the MIT study does and does not show, and the think-first habit that follows from the evidence.
-
India is teaching Google how AI in education can scale
TechCrunch · Article · Jan 2026
Reports the Economic Survey's warning, citing MIT and Microsoft, about over-reliance on AI in learning.