Experiment shows AI interactions lower conspiracy beliefs in 955 participants, indicating persuasive potential of AI.
Key Points
Conversations with AI reduced confidence in conspiracy beliefs significantly and effectively, challenging their resistance.
Participants reported lower confidence in both conspiracy and unwarranted beliefs after interacting with AI, underscoring AI’s persuasive capabilities.
Randomized conditions assessed whether the AI's characterization as a tool or human impacted effectiveness but showed no significant differences.
Results support the idea that compelling messages are crucial for persuasion, regardless of whether an AI or human delivers them.