Voters are already asking AI about elections.

When a voter asks about a candidate’s biography, policy position, or contrast with an opponent, an assistant has to retrieve and synthesize available information. A fragmented, stale, or poorly sourced public presence leaves more room for omission and error.

A 2025 paper in Nature also found that conversations with AI models produced significant changes in candidate preferences across election experiments. The researchers reported effects larger than those typically found for traditional video advertising, while also finding that some factual claims made by the models were inaccurate.

That combination—real influence and imperfect accuracy—is the case for better evidence, not more campaign spin. A well-structured candidate profile should make it easier for voters and their tools to distinguish a documented position from an unsupported claim.