Slate Daily Feed What Next: TBD | Ready for Chatbot Canvassers?
Dec 12, 2025
David Rand, a Cornell University professor specializing in misinformation and persuasion, explores the impact of AI chatbots on political beliefs. He reveals that while chatbots can effectively present factual claims, they may risk flooding conversations with inaccuracies when pushed to persuade. Rand discusses how these digital canvassers could outperform human ones in voter outreach, highlighting both the potential benefits and risks to democracy. The conversation dives into the balance between honest communication and persuasive power, calling for transparency and regulation.
AI Snips
Chapters
Transcript
Episode notes
Example Conversation That Changed A Mind
- A recruited participant preferring Kamala Harris described abortion nuances before chatting with a bot.
- After a polite chatbot presented pro-Trump arguments the participant said Trump's economic policies appealed to them.
Chatbots Can Shift Voter Preferences
- Chatbots can convincingly present many factual-sounding claims and sway voter preferences.
- David Rand found measurable vote changes from AI dialogues across the U.S., Canada, and Poland.
Measured Effects Are Small But Meaningful
- The experiments asked people their candidate preference, let them converse with a pro-candidate bot, then rechecked preferences.
- One in 25 U.S. voters and up to one in 10 in Canada/Poland changed toward the bot's candidate.