AI Chatbots Shift Voter Views in Study

A Cornell-led study in Nature found that a single AI chatbot exchange advocating for a candidate measurably moved voter preferences.

Updated on Nov. 3, 2026 in Artificial Intelligence

Live Poll

Do you trust AI chatbots to give you accurate information about candidates?

Published in Nature in December 2025, a Cornell-led study had more than 2,300 Americans talk with an AI model about two months before the 2024 presidential election. The pro-Harris model moved likely Trump voters 3.9 points toward Harris on a 100-point scale.

Why it matters

The researchers found the AI's effect exceeded that of traditional ads, which are less personalized than chatbot conversations. For voters across the United States, the findings raise questions about how persuasive AI-driven political exchanges may be as the 2026 midterms approach.

The Nature study tested more than 2,300 Americans, and the pro-Harris AI model moved likely Trump voters 3.9 points toward Harris on a 100-point scale, about four times the effect of traditional ads tested in the 2016 and 2020 elections. The stub does not report how long these shifts lasted after the conversations ended.

The players

Cornell University

A private Ivy League research university in Ithaca, New York, where the researchers conducted the persuasion study.

The details

Researchers had participants converse with AI models instructed to advocate for a candidate, and the models produced arguments and evidence tailored to each individual. A separate 2024 Science study found roughly a quarter of conspiracy believers stopped believing after about eight minutes of AI conversation, with the AI's claims accurate around 99% of the time.

December 2025: Nature publishes the study on chatbot persuasiveness.
About two months before the 2024 presidential election: Participants chat with the AI models.
2024: Science publishes the study on AI reducing conspiracy beliefs.
November 3, 2026: Voters head to the polls for the midterm elections.

Roadmap

The study measures AI persuasion against the advertising baseline established in the 2016 and 2020 U.S. presidential elections, where the chatbot effect was about four times larger. This places conversational AI as a potentially stronger political influence than the paid media that has shaped recent campaigns.

Voters who discuss candidates or issues with AI chatbots may be influenced by personalized arguments more than by the static ads they usually encounter. Readers should consider how AI-generated political content tailored to them could shape their views before the midterm elections.

The takeaway

Personalized AI conversations can shift political opinions more than traditional advertising, so readers should weigh chatbot political claims carefully. Checking candidate positions against primary sources helps guard against unintended persuasion.