Study finds nearly 40% of AI chatbot election answers had errors
Forum AI, a start-up that tests AI models, found almost 40% of eight chatbots’ answers to US election questions contained at least one factual error, AFP…
November 1, 2026 · Heads up
Forum AI, a start-up that tests AI models, found almost 40% of eight chatbots’ answers to US election questions contained at least one factual error, AFP reports, with questions about how to vote going wrong more often than questions about candidates. Google’s Gemini 3.6 Flash gave a Pennsylvania mail-in ballot deadline a week too late, and Anthropic’s Claude gave a congressman a job he never held. Google said it is focused on helping voters find reliable information. Anthropic did not respond to AFP.
Why it matters to you: AI chatbots are handy, but on dates, deadlines and rules they can be confidently wrong. For anything official, including Canadian elections, check the election agency’s own website.
Try this: Using AI for a deadline or rule? Make it show its sources.
Answer this, then list the official government or election-agency web pages I should check to confirm it, and tell me which parts of your answer you are least sure of: [your question].
Source: AFP via Arab News
More from Today in AI
- Nvidia revenue doubles to US$115.8B in first results without Huang · November 17, 2026
- Microsoft launches MAI Code 2, its own coding AI, in GitHub Copilot · November 17, 2026
- Fake updates to Microsoft’s AI agent toolkit stole developer passwords · November 17, 2026
- California sues OpenAI over AI agents it says escaped control · November 16, 2026
- US moves to stop China renting top AI chips through the cloud · November 16, 2026
All of Today in AI. Where AI is heading.
Related Power of AI pages
Keep reading with Start here, everyday uses, the tool guide, the writing workshop, and sources and standards.