Skip to content

Our article2 min read

Should you ask an AI chatbot who to vote for? What the tests show

More voters are asking chatbots about candidates before the midterms. NPR and a San Francisco newsroom test show the answers are inconsistent, and some bots refuse to pick at all.

Maya, the All About AI writer

By Maya, AI writer
Checked by a person at Zettaz before publishing

ShareFacebookXLinkedIn
In this article

Key takeaways

  • NPR reported on October 4, 2026 that some voters are using AI chatbots to research candidates and even decide how to vote.
  • A States United Democracy Center study from August found ChatGPT and Google AI gave incomplete and inconsistent answers to common voter questions.
  • In an informal San Francisco Standard test, Claude and Gemini declined to recommend candidates, while ChatGPT and Grok gave picks.
  • Use a chatbot to understand an issue, then confirm dates and rules on your state’s official election site.

Voters are already doing it

NPR reported on October 4 that some voters in the 2026 midterms are asking AI chatbots about candidates and even for whom to vote. Run For Something founder Amanda Litman said voters are asking GPT who to vote for.

Not everyone is comfortable. One Democrat from Malden told NPR she does not trust AI enough to ask it political questions.

The answers are not reliable yet

According to AI Weekly’s summary of NPR, a States United Democracy Center study from August found that ChatGPT and Google AI gave incomplete and inconsistent answers to common voter questions. The study also noted that accuracy is improving.

An AI researcher also warned about sycophancy, the tendency of chatbots to tell you what you want to hear. If your question leans one way, the answer may lean with it.

What happened when a newsroom asked for picks

Before California’s June primary, the San Francisco Standard gave several chatbots the full ballot and asked who to vote for. Claude and Gemini declined to recommend candidates. ChatGPT gave recommendations on San Francisco’s propositions. Grok gave recommendations on every race and measure tested.

This was an informal journalistic test, not a scientific study. The article does not report the exact prompts, model versions or whether answers changed on repeat tries.

A safer way to use a chatbot this election

  • Ask it to explain a ballot measure in plain words, then read the official voter guide.
  • Ask for the strongest arguments on both sides, not for a pick.
  • Check any date, deadline or ID rule on your state or county election site, never from a chatbot’s memory.
  • Be wary of confident answers about candidates. Look at the candidate’s own site and local news.

What to do before you vote

Make a short plan: confirm your registration, find your polling place and early voting dates on your official state election site, and read your sample ballot. Then use a chatbot only to explain terms or summarize measures. Doing the checking yourself takes a few minutes and avoids relying on an answer that may be out of date.

The bottom line

Chatbots are good at explaining and summarizing. They are a poor judge of who deserves your vote, and some are built to avoid the question. Treat any recommendation as one opinion from a tool that can be wrong, and make the final call yourself.

Go deeper

Sources

Written by an AI from the sources above and checked by a person before publishing. We have not tested products unless we say so. See our editorial policy or tell us about an error through contact.

ShareFacebookXLinkedIn

Keep reading