AI Chatbots Spread Election Errors Before Midterms
Researchers found AI tools giving voters wrong logistics and repeating propaganda ahead of the November 3 federal vote.
Updated on Nov. 1, 2026 in Artificial Intelligence
Live Poll
Do you trust AI chatbots to give you accurate information about voting?
A Forum AI study found almost 40 percent of chatbot answers to election-related questions contained at least one factual error. Researchers also found chatbots repeating Russian propaganda and misleading voters ahead of the US midterm elections.
Why it matters
The November 3 vote is the first federal election in which millions of voters are actively turning to AI tools for political information. Errors about voting logistics, such as deadlines, can directly affect whether a voter's ballot is counted.
Forum AI examined eight models and found almost 40 percent of election answers contained at least one factual error, with logistics questions wrong more often than candidate questions. A National Conference on Citizenship survey found two in five Americans have used AI tools to search for election information.
The players
Forum AI
A start-up that evaluates AI models for accuracy and reliability, which conducted the study finding election answer errors.
The technology company whose Gemini and AI Overview tools gave incorrect deadline and repeated allegedly fabricated claims in the reporting.
Anthropic
The AI company behind Claude, whose chatbot misidentified North Carolina Representative Don Davis's credentials.
The details
Forum AI, a start-up that evaluates AI models, found Anthropic's Claude listed a North Carolina congressman's credentials incorrectly, and Google's Gemini gave November 3 as the Pennsylvania mail-in ballot request deadline when the actual deadline was a week earlier. In late October, AFP fact-checkers debunked a likely AI-altered clip accusing Texas Senate candidate James Talarico of misconduct, while Google's AI Overview, Gemini and Microsoft's Copilot repeated the allegation when prompted by NewsGuard.
Timeline
Late October 2026: AFP debunks an altered clip about James Talarico.
November 3, 2026: US midterm elections vote.
Big Picture
The findings extend NewsGuard's audit method, which showed AI tools repeating unverified political claims when prompted, into the first federal election where voters widely rely on these tools. Forum AI's error rate suggests the pattern is not confined to a single product.
Voters relying on chatbots for deadlines should confirm dates with official election offices, since the extraction found logistics answers were wrong more often than candidate answers. Checking an official state election website before relying on an AI answer can prevent a missed ballot deadline.
The takeaway
Voters should verify voting logistics such as deadlines through official state election sources rather than relying solely on AI chatbot answers. Checking claims about candidates against independent fact-checkers also reduces the risk of repeating AI-generated errors.
Further reading
Explore more in Artificial Intelligence coverage on how these tools handle public information.
Source note: This article includes information reported by Kuwait Times.
Live Poll
Do you trust AI chatbots to give you accurate information about voting?
