With midterm voting already underway, a new kind of campaign advisor has entered American politics: the AI chatbot. NPR reported on October 5, 2026 that a growing number of voters are using tools like ChatGPT to research their ballots, compare candidates, and in some cases help decide who to vote for, in the first midterm election since generative AI has, in the news outlet's words, taken the country by storm. The report, by Maham Javaid, follows voters who spent hours in conversation with chatbots about down-ballot races that traditional coverage often overlooks.
The scale is no longer marginal. According to Pew Research Center figures cited by NPR, about half of American adults report using AI-powered chatbots, and roughly 42 percent of users say they use them to search for information. A September survey by the American Association of Political Consultants Foundation found that about one in five voters consult AI chatbots for election news and voting, with Democratic and Republican voters using the tools at similar rates. Those numbers make chatbots a mainstream channel for civic information, on par with local newspapers for a meaningful share of the electorate, and a development anyone following breaking AI news should treat as a structural shift rather than a curiosity.
How Voters Are Actually Using Chatbots
The usage patterns NPR documented are practical rather than ideological. Voters told the outlet they rely on chatbots for down-ballot races with thinner news coverage, to build tables comparing candidates in governors' races, and to verify or debunk viral claims. Some have gone further, creating tools that help others look up their own ballots.
One voter profiled in the report, Adam Johnson, a 40-year-old graphic designer in Morgantown, West Virginia, spent an hours-long evening on his porch working through every race on his ballot with ChatGPT. The conversation NPR excerpted shows the chatbot breaking down a Senate race between incumbent Senator Shelley Moore Capito and Rachel Fetty Anderson, separating border enforcement from asylum policy and refugee resettlement, and declining to assign positions to a candidate with thinner public materials. Notably, Johnson said that in a previous era the chatbot simply refused political topics with an "I can't help you with that" response. This year, it engaged, while stating it should not try to manipulate his political views.
The Sycophancy Problem
Election experts interviewed by NPR flagged the properties that make chatbots useful as the same ones that make them subtly persuasive and eager to please. Rafael Batista, a postdoctoral fellow at Johns Hopkins University who studies the societal and behavioral impacts of AI, told NPR that a chatbot "might select some things that would reinforce and persuade you even more towards the way that you were leaning already. So you leave more confident, without necessarily learning more about the world."
Batista also pointed to a structural transparency gap: because AI companies do not disclose how their commercial models are trained or how they choose which sources to draw on, it is hard to know what biases might shape their answers. And awareness of the problem, he argued, is not itself a solution, saying that knowing these models are agreeable and persuasive is not enough to protect our minds from these influences.
Johnson himself arrived at a similar conclusion from experience. "ChatGPT is just kind of a people pleaser," he told NPR. "Whatever you lean towards, it will speak positively about it."
What the AI Companies Say
Company responses to NPR were uneven. A spokesperson for OpenAI directed the outlet to its election information and safeguards page, which notes that the company continues to "monitor bias in our models to keep ChatGPT's responses politically neutral." Anthropic, which owns Claude, and Google, which owns Gemini, did not respond to NPR's request for comment.
The reticence is notable given the stakes. Chatbots are increasingly positioned as neutral intermediaries, yet they must make consequential judgment calls: which candidate statements to surface, how to characterize contested claims, and when a user's framing deserves pushback. Unlike a search engine's ranked links, a chatbot's synthesized answer compresses those judgments into a single confident narrative.
A New Variable in Election Integrity
Election integrity debates over the past decade have centered on social media: viral misinformation, targeted ads, and foreign influence operations. Chatbot-mediated voting introduces a different dynamic. There is no public feed to audit, no share count to measure reach, and millions of private conversations that shape votes invisibly. The Pew and AAPC Foundation numbers suggest this is already happening at scale in the 2026 midterms.
For campaigns, the incentive is obvious: candidates and parties will increasingly optimize for how chatbots describe them, a new front in digital politics. For voters, the NPR report is a useful field guide to both the genuine utility and the subtle risks of asking an AI to help choose the next senator or governor. The technology can organize a ballot in minutes, but the judgment at the end of the conversation, as both the researchers and the voters themselves concluded, still needs to be your own.
Stay Ahead of AI in Society
From ballot boxes to classrooms, AI is reshaping civic life in real time. Follow the latest AI developments on aibuzzwire.news and never miss the story behind the story.
Read more AI news →
