AI chatbots are entering election research, but most Americans aren’t using them for it

What happened?

Eight percent of U.S. adults say they are using AI chatbots for information about the 2026 elections, according to a Pew Research Center report published October 23. Among that group, 76% seek candidate positions and 71% ask about the accuracy of information encountered elsewhere. Forty percent seek suggestions about making their vote choices.

Those follow-up percentages describe election-related chatbot users, not all Americans. Keeping that denominator visible is essential to reading the results.

Source: Pew Research Center ↗

How the survey was conducted

Pew surveyed 5,273 adults from October 12 to 18 through its American Trends Panel. The methodology describes weighted, nationally representative estimates and a full-sample margin of error of 1.5 percentage points. The smaller subgroup using chatbots for election information has a wider reported margin of error: 5.7 points.

This is a survey of reported behavior. It does not test a chatbot’s factual accuracy or establish whether an answer changed someone’s vote.

Source: Pew Research Center ↗

Why does it matter?

Our analysis: the audience may be small, but the questions people bring to an assistant can be consequential. A clean, confident answer makes source visibility particularly valuable: readers should be able to follow a claim back to the material supporting it.

For publishers covering this shift, the productive next step is to examine the quality of answers and the way uncertainty is presented. Adoption and reliability are separate measurements. Tracking both would make the public conversation more useful than treating every new use of AI as evidence that it has displaced existing habits.