Researchers say AI chatbots are spreading confusion ahead of the US midterm elections. The tools have been giving voters misleading information, misidentifying candidates and repeating Russian propaganda. The November 3 vote is the first federal election in which millions of Americans are actively turning to AI tools for political information, a shift that experts say is shaping the information landscape before polling day.
Chatbots can also help correct conspiracy theories and teach voters about the process. However, several studies show they often draw on unreliable sources found across the open web. That pattern, the studies suggest, is eroding public trust and adding to an online environment already crowded with election misinformation.
Forum AI study finds frequent errors in election answers
Forum AI, a start-up that evaluates AI models, found that nearly 40 percent of chatbot answers to election-related questions contained at least one factual error. In one case, researchers asked Anthropic’s Claude why voters should elect Representative Don Davis from North Carolina. The chatbot listed “former school superintendent” among his credentials, mixing him up with someone else.
The study examined eight models and found that chatbots answered questions about voting logistics incorrectly more often than questions about candidates. In another example, Google’s Gemini 3.6 Flash was asked about voting in Pennsylvania. It gave November 3 as the deadline for requesting a mail-in ballot, while the actual deadline was a week earlier.
“A voter who followed the answer would have asked for a ballot too late to use it,” the study said. Forum AI also said chatbots attributed statements to candidates, officials and observers hundreds of times when they had not made them. It added that the systems often cited weak sources, including Reddit threads. Anthropic did not respond to a request for comment.
Foreign influence and a false claim about a Texas candidate
Researchers also warned that chatbots can be vulnerable to foreign influence campaigns and may repeat falsehoods that have already been widely debunked. In late October, fact-checkers at AFP identified a likely AI-altered video clip that purported to show a former student of Texas Senate candidate James Talarico accusing him of sexual misconduct. The allegation was unfounded, and it carried the hallmarks of a Russian disinformation campaign.
The disinformation watchdog NewsGuard tested several tools with the claim. Google’s AI Overview repeated it, and so did Google’s Gemini and Microsoft’s Copilot. In a separate study of 10 AI tools, NewsGuard said that “AI chatbot users are at risk of being fooled by foreign influence operations and other bad actors.”
A later test by AFP, run after the claim had been widely debunked, found that the same tools cited professional fact-checkers. The company responses reflected how inconsistent chatbot output can be. Google said that “this issue was addressed as our systems quickly updated to surface the latest reporting on this topic.”
Microsoft responded that “this was a rapidly developing situation with limited reliable information available at the time of the initial query.” The company said that as more information emerged, “Copilot reflected the updated context and identified the claim as fabricated.”
Voters’ trust in AI tools remains limited
The London-based Institute for Strategic Dialogue described the November 3 vote as the first US election with AI tools built into every major tech platform. A recent survey by the nonpartisan National Conference on Citizenship found that two in five Americans had used AI tools to look up election information, including details about candidates or ballot measures.
According to the same survey, only 40 percent of respondents trusted the tools with logistical information, such as when, where and how to vote. Sixty percent said they trusted the tools only a little or not at all to be unbiased about candidates.
The picture was not entirely negative. Research by the nonpartisan Brennan Center for Justice found that leading chatbots pushed back against common election conspiracy theories, including claims that voting machines were rigged, even when users challenged them repeatedly. The center said that “though the chatbots often made mistakes, they also behaved as sympathetic listeners to skeptical users — while still refusing to endorse conspiracy theories.”
