How AI Talks People Out of Conspiracy Theories—and What We Can Learn From That - WSJ
wsj.com
AI chatbots are cracking open one of persuasion science's hardest nuts — changing conspiracy believers' minds — and the method reveals surprising lessons about how belief actually works.
Inoculation TheoryBackfire EffectMotivated ReasoningPersuasion Theory
Theory Briefing
- Researchers found that LLMs like ChatGPT and Claude, when instructed to debunk conspiracies, were measurably effective at shifting believers' views.
- AI's success challenges the backfire effect theory, suggesting that patient, tailored counter-arguments can overcome entrenched misinformation.
- The findings point to inoculation theory and motivational interviewing as the hidden mechanics behind why AI-led debunking actually sticks.