- Browse/Search the Program
Search
Browse By Day
Browse By Time
Browse By Person
Browse By Division
Browse By Session or Event Type
- Information Menu
Search Tips
How to Build a Personal Program
Virtual Exhibit Hall
- Navigation and Settings Menu
Personal Schedule
Change Preferences / Time Zone
Artificial Influence? Comparing AI and Human Persuasion in Reducing Belief Certainty
Sat, April 5, 5:10 to 6:40pm CDT (5:10 to 6:40pm CDT), Palmer House Hilton, Floor: 7th, Clark 7Brief Overview
This study tests whether a conversation with ChatGPT-4 can reduce certainty in false beliefs. Results show AI is more persuasive than a random peer but not an expert on the topic, suggesting AI may not surpass the persuasive power of human experts.
Authors
Natasha Goel, University of Toronto
Thomas Bergeron, University of Toronto
Blake Lee-Whiting, University of Toronto
Danielle Bohonos, University of Toronto
Thomas Galipeau, University of Toronto
Md Mujahedul Islam, University of Toronto
Sarah Lachance, University of Toronto
Sonja Margit Savolainen, University of Helsinki
Clareta Treger, University of Toronto
Eric Merkley, University of Toronto