AI Chatbots Shown to Sway Voters, Raising New Fears about Election Influence

AI chatbots are incredibly good at political persuasion

By Deni Ellis Béchard edited by Claire Cameron

Forget door knocks and phone banks: chatbots could be the future of persuasive political campaigns.

Fears about the influence of artificial intelligence on elections are not new. But two new articles published today in Nature And Science show that robots can successfully change people’s political attitudes, even if what they claim is false.

The findings run counter to the prevailing logic that it is extremely difficult to change people’s minds about politics, says David Rand, lead author of both papers and a professor of information sciences and marketing communications and management at Cornell University who specializes in artificial intelligence.

On supporting science journalism

If you enjoy this article, please consider supporting our award-winning journalism by subscription. By purchasing a subscription, you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

Stephan Lewandowsky, a cognitive scientist at the University of Bristol in England, who was not involved in the new studies, says they raise important questions: “First, how can we guard against – or at least detect – when LLMs [large language models] were designed with a particular ideology in mind that is antithetical to democracy? he asks. “Second, how can we ensure that “rapid engineering” cannot be used on existing models to create undemocratic persuasive agents?

The researchers studied more than 20 AI models, including the most popular versions of ChatGPT, Grok, DeepSeek and Meta’s Llama.

In the experiment described in Nature In the paper, Rand and colleagues recruited more than 2,000 American adults and asked them to rate their candidate preference on a scale of 0 to 100. The team then asked participants to chat with an AI trained to advocate for one of two candidates in the 2024 US presidential election: Kamala Harris or Donald Trump. After the conversation, participants again ranked their candidate preferences.

“This moved people on the order of a few percentage points toward the candidate the model was advocating, which is not a huge effect but is significantly larger than what one would expect from video ads or traditional campaign ads,” says Rand. Even a month later, many participants still felt convinced by the robots, according to the newspaper.

The results were even more striking with around 1,500 participants in Canada and 2,100 in Poland. But interestingly, the biggest change in opinion occurred in the case of 500 people talking to robots about a statewide vote to legalize psychedelics in Massachusetts.

Notably, if the bots did not use evidence to support their arguments, they were less convincing. And while the AI ​​models mostly stuck to the facts, “the models that championed right-wing candidates — and particularly the pro-Trump model — made much more inaccurate claims,” Rand says. This trend persists across countries and AI models, although people least informed about politics in general are most convincing.

THE Science The article addressed the same questions but from a chatbot design perspective. In three studies conducted in the United Kingdom, almost 77,000 participants discussed political issues with chatbots. The size of an AI model and what the robot knew about the participant had only a slight influence on its persuasive power. Rather, the biggest gains come from how the model was trained and tasked with presenting evidence.

“The more factual claims the model contained, the more convincing it was,” says Rand. The problem arises when such a robot lacks specific evidence to support its argument. “He needs to start grasping at straws and making up claims,” he said.

Ethan Porter, co-director of the Institute for Data, Democracy and Politics at George Washington University, describes the findings as “milestones in the literature.”

“Unlike some of the more pessimistic narratives, they make clear that facts and evidence are not dismissed if they are inconsistent with the individual’s prior beliefs. Instead, facts and evidence can form the basis of successful persuasion,” says Porter, who was not involved in the articles.

The finding that people are more effectively convinced by evidence rather than emotions or feelings of group membership is encouraging, says Adina Roskies, a philosopher and cognitive scientist at the University of California, Santa Barbara, who was also not involved in the studies. Yet, she warns, “the bad news is that people are swayed by apparent facts, regardless of their accuracy.”

It’s time to defend science

If you enjoyed this article, I would like to ask for your support. Scientific American has been defending science and industry for 180 years, and we are currently experiencing perhaps the most critical moment in these two centuries of history.

I was a Scientific American subscriber since the age of 12, and it helped shape the way I see the world. SciAm always educates and delights me, and inspires a sense of respect for our vast and beautiful universe. I hope this is the case for you too.

If you subscribe to Scientific Americanyou help ensure our coverage centers on meaningful research and discoveries; that we have the resources to account for decisions that threaten laboratories across the United States; and that we support budding and working scientists at a time when the value of science itself too often goes unrecognized.

In exchange, you receive essential information, captivating podcasts, brilliant infographics, newsletters not to be missed, unmissable videos, stimulating games and the best writings and reports from the scientific world. You can even offer a subscription to someone.

There has never been a more important time for us to stand up and show why science matters. I hope you will support us in this mission.