Despite reports that voters are souring on AI, many will still use AI platforms to inform themselves about elections and politics as we head into midterms. But how will these platforms play a role in influencing voters?
David Rand, who researches how AI chatbots can effectively sway voters and is a professor of psychology, information science and management and co-creator of DebunkBot, says the question isn’t how the platforms will steer voters, but rather “who is steering and in what direction?”
Rand says: “Conversations with AI chatbots are remarkably persuasive. In our experiments across multiple countries, short conversations with AI shifted people's views on candidates and policy issues, substantially more than traditional political ads do.
“The AI mostly persuades by presenting lots of facts and evidence—rather than sophisticated psychological manipulation—but those ‘facts’ aren’t always accurate, and inaccurate facts are just as persuasive. So as campaigns and other actors deploy AI this cycle, voters could be swayed by information that is one-sided or just plain wrong.
“But the same persuasive power can also be used for good. In our work, AI conversations durably reduced belief in conspiracy theories and corrected misperception, for example about climate change, vaccines, and election integrity. And millions of people are already asking AI bots on X to fact-check political claims.
“The question isn’t whether AI will shape how voters think. It's who's steering it, and toward what.”
