Researchers study how AI is changing how voters get election info

In short: More voters are using AI tools for election information, and researchers say it can help people learn, but it can also spread wrong details and influence opinions.

What's going on

AI is increasingly built into everyday tools people already use, like search engines and chatbots. That means voters may see AI-written answers about candidates, issues, or how to vote, sometimes without realizing an AI system made the text.

Election experts warn that this creates a new problem: AI can sound confident while still being wrong. The U.S. Election Assistance Commission has cautioned that AI-generated election guidance can look believable but include inaccurate details, especially for basics like voting dates, hours, and locations.

At the same time, early research suggests AI is not always making people less informed. A UK study from the Alan Turing Institute found that about one in eight voters used AI chatbots to ask election-related questions. In that study, chatbot use was linked to higher political knowledge, not lower.

What to watch

Researchers are now testing whether chatbots truly improve understanding or whether they mainly produce smooth answers that are hard to fact-check (like a fast-talking tour guide who sometimes makes things up). They are also studying how AI can shape political attitudes through how it frames information, and whether it nudges users toward certain candidates or parties.

Another focus is regulation and safety. Policy groups warn that AI can help spread deepfakes (fake audio or video that looks real), fake endorsements, and targeted deception aimed at specific groups. Several U.S. states are already adopting rules that require disclosures or restrict certain AI-made political ads and deepfakes.

Source: NYTimes