Does AI help or hurt misinformation? Here's what we found. - The Washington Post
washingtonpost.com
Researchers tested popular chatbots on election conspiracy theories — and what they found challenges assumptions about AI as either a cure or a cause of misinformation.
Dual-Use TechnologyInformation Deficit ModelGatekeeping TheoryAlgorithmic Accountability
Theory Briefing
- Washington Post researchers prompted leading chatbots with real election conspiracy theories to see whether the AI pushed back or played along.
- The study probes whether AI systems act as a check on false claims or become a tool for generating and spreading them.
- Chatbots' responses to misinformation prompts varied, raising questions about whose job it is to set the guardrails — the companies, the users, or regulators.
- The findings land as AI-generated content becomes a growing factor in how voters encounter political information.