Research finds AI users scarily willing to "surrender" their cognition to LLMs

Study involved 1,372 participants and over 9,500 individual trials.

Participants accepted faulty AI reasoning 73.2% of the time and overruled it 19.7% of the time.

High trust in AI was linked to being more likely to be misled, while higher fluid IQ was linked to being less likely to be misled.

The fluency and confidence of AI outputs led people to treat those outputs as epistemically authoritative.

Researchers noted that cognitive surrender to AI is not inherently irrational.