As Americans head to the polls, some voters have used AI chatbots to learn about candidates and policies. Pew Research data cited in a WIRED report says around half of surveyed U.S. adults have used an AI chatbot, while 8 percent said they had used one to find election information ahead of the midterms.
But the answers may not be dependable. Research by Forum AI found that almost 40 percent of election-related responses from mainstream chatbots contained at least one factual error. In examples cited by the report, Google’s AI Mode declined to provide the results of the 2020 election, while Gemini gave an incorrect deadline for requesting a mail-in ballot in Pennsylvania. Forum AI also found that Claude confused North Carolina congressman Don Davis with a former school superintendent who had the same name.
The report says chatbot answers can be out of date, fabricated or drawn from partisan sources. NewsGuard research found that mainstream chatbots cited so-called “pink slime” websites nearly half the time when asked about candidates and policies related to the midterms. These sites present themselves as conventional news outlets but are funded by political or ideological groups and individuals.
According to NewsGuard, ChatGPT cited these sites in more than 70 percent of searches in its research, the highest share among the chatbots assessed. Grok had the lowest share, at 30 percent.
Nina Jankowicz, executive director of the American Sunlight Project and a former disinformation official in the Biden administration, told WIRED that many Americans may not be prepared to use consumer AI during an election. She said chatbots depend on the information available online and may make things up when they do not know an answer.
Google and Anthropic did not respond to WIRED’s requests for comment. OpenAI referred the publication to a blog post describing its approach to elections in 2026.
Sources: WIRED.