Wharton Study: AI Makes You Dumber — But More Confident
People followed wrong AI answers 80% of the time, scoring worse than solo — yet feeling more certain than ever.
A new Wharton study across 1,372 participants and ~10,000 trials has a name for what's happening to us: cognitive surrender. When AI gave deliberately wrong answers, people adopted them over 80% of the time — performing worse than if they'd worked alone.
Researchers Gideon Nave and Steven Shaw extend Kahneman's classic System 1/2 framework with a System 3: artificial cognition. The problem isn't just that people trust AI — it's that they borrow its confidence wholesale. Participants felt more certain holding AI-assisted wrong answers than they did with their own unassisted correct ones.
The study used Cognitive Reflection Test problems — puzzles with a seductive wrong answer and a deliberate right one. AI short-circuited both intuition and reasoning.
Why it matters: as AI becomes the default thinking partner, the risk isn't losing skills gradually — it's outsourcing judgment in real time without knowing you did.
Sources
Independent coverage
- Thinking—Fast, Slow, and Artificial (SSRN paper) ssrn.com
- Wharton: How AI Is Reshaping Human Intuition and Reasoning knowledge.wharton.upenn.edu
Written by an AI pipeline from the sources above. Methodology · Report an error
Feed, daily deep-dive and bytes — readable offline, with push alerts for the topics you follow.