AI access makes people almost entirely unwilling to say "I don't know," study finds
44 to 3 percent. Access to AI nearly erased people's willingness to admit uncertainty—even when the AI was almost always wrong.
Why it matters
A behavioral study reveals a critical gap between AI confidence and AI accuracy: users become far less likely to express doubt when AI is available, yet their correctness plummets. This has direct implications for enterprise deployments where human judgment, verification, and appropriate skepticism are safety-critical.
The key facts
8 to knowStudy of 3,000+ participants
Willingness to say 'I don't know' dropped from 44% to 3% with AI access
AI was almost always wrong in the experiment
Users with AI felt more confident but were correct only one-third as often as those without AI
Confidence-accuracy gap widens significantly under AI access
3,000+ participants
Users with AI felt more confident but were correct only one-third as often as those without AI access
Study published in September 2026
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: A study with more than 3,000 participants shows that just having access to AI answers nearly eliminated people's willingness to say "I don't know." In one experiment, it dropped from 44 to 3 percent, even though the AI was almost always wrong. Participants who used AI felt more confident but were…