Do Androids Dream of Corrective Tweets? Dynamic Patterns in AI Verification of Political Misinformation

• In Event: AI, Misinformation, and Political Beliefs

Brief Overview

An observational study of verifications by the AI chatbot Grok shows that its accuracy is weakest at the onset of misinformation but improves as experts weigh in, with implications for research on LLMs and debates over replacing human fact-checkers.

Author

• Agustin Prinetti, Rice University