⌘K

Change language Switch ThemeSign In

Narrow Mode

The Cognitive Surrender: AI's Dunning-Kruger Effect

The Cognitive Surrender: AI's Dunning-Kruger Effect

@GaryMarcus

AI 时代邓宁-克鲁格效应的恐怖新形态

!Image 3: Lukasz Olejnik

#### Lukasz Olejnik

@lukOlejnik · 11h ago

Does AI make us collectively dumber? Is "grok wat do u think" the new "I've outsourced my common sense to a chatbot and I feel great about it"? A study secretly made an AI give wrong answers half the time while people used it to solve logic puzzles. People followed the wrong AI 80% of the time. Their confidence went up regardless. This is the "cognitive surrender". The moment people stop asking "is this true?" and start asking "what does the model say? I don't care, too". The habit of deference is sticky.Show More

!Image 4: 媒体 1

!Image 5: 媒体 2

!Image 6: 媒体 3

7

21

74

9,935

Mar 20, 2026, 9:55 PM View on X

8 Replies

21 Retweets

87 Likes

9,158 Views  Gary Marcus @GaryMarcus

One Sentence Summary

Gary Marcus shares a concerning study showing people trust AI wrong answers 80% of the time, calling it 'cognitive surrender' - a terrifying AI twist on the Dunning-Kruger effect.

Summary

This tweet references a study about AI's impact on human cognition. The quoted research found that when an AI gave wrong answers half the time while people used it to solve logic puzzles, people followed the wrong AI advice 80% of the time, and their confidence increased regardless of accuracy. Marcus coins this phenomenon 'cognitive surrender' - the moment people stop asking 'is this true?' and instead defer entirely to what the model says. This represents a terrifying evolution of the Dunning-Kruger effect, where AI users become overconfident while actually becoming less capable of critical thinking.

AI Score

78

Influence Score 19

Published At Yesterday

Language

English

Tags

AI Psychology

Dunning-Kruger Effect

Cognitive Surrender

AI Trust

Human Cognition HomeArticlesPodcastsVideosTweets

The Cognitive Surrender: AI's Dunning-Kruger Effect | Bes...