Voting is already underway in the 2026 midterm elections, and a growing number of Americans are experimenting with a novel approach to preparing for their ballots: consulting artificial intelligence chatbots. Rather than relying solely on traditional sources such as voter guides, news coverage, or conversations with friends and family, some voters are asking large language models to help them research races, compare candidates, and even decide how to cast their votes.
One such voter is Adam Johnson, a 40-year-old graphic designer from Morgantown, West Virginia, who is studying to become an illustrator. On a September evening, Johnson sat on his porch and spent hours reviewing his ballot with ChatGPT, the chatbot developed by OpenAI. He walked through each race and each candidate, using the AI tool to help him determine who best aligned with his values.
Johnson's reliance on the chatbot is not an isolated experiment. He described ChatGPT as a significant part of his daily life, using it for tasks ranging from meal planning to job hunting to various household needs. That familiarity made it a natural companion when he turned his attention to the upcoming election.
His process began by asking the chatbot to compile a list of every race appearing on his ballot this year. From there, he dug deeper into individual contests, prompting the AI to provide information about the candidates and their positions. The interaction was lengthy, unfolding over an extended conversation rather than a single query.
Johnson noted that his earlier attempts to discuss politics with the chatbot had been met with resistance. He recalled that on a previous occasion, when he raised political topics, the AI seemed unwilling to engage, offering only a simple refusal. That cautious posture reflects the guardrails that many AI companies have built into their systems to avoid controversial or potentially biased political output.
Despite those earlier limitations, Johnson persisted, and this time the conversation went further. His experience illustrates a broader shift in how voters are incorporating emerging technology into civic life. The appeal is understandable: AI chatbots can process large amounts of information quickly, summarize candidate positions, and respond to follow-up questions in real time, all in a conversational format that feels accessible.
For voters facing crowded ballots and a torrent of campaign messaging, the promise of an on-demand research assistant is attractive. Midterm elections often feature dozens of down-ballot races, from congressional seats to state legislative and local offices, and many voters find it difficult to gather reliable information about every contest. AI tools offer a way to fill those gaps without spending hours sifting through disparate sources.
However, the trend also raises concerns. Large language models are known to occasionally generate inaccurate or fabricated information, a phenomenon often called hallucination. When applied to elections, such errors could mislead voters about a candidate's record, party affiliation, or policy stances. The stakes are particularly high because voting decisions are difficult to reverse once a ballot is cast.
There are also questions about the political neutrality of AI systems. Companies like OpenAI have worked to limit their chatbots' willingness to endorse candidates or take partisan positions. Yet the training data and design choices behind these models can subtly shape the information they provide. Users may not always recognize those influences, especially when responses are delivered in a confident, authoritative tone.
The rise of AI-assisted voting preparation fits into a longer history of technology reshaping how citizens engage with elections. In previous cycles, voters turned to social media, search engines, and online voter guides to inform themselves. Each new medium has prompted debates about accuracy, bias, and the risk of misinformation, and AI chatbots are now the latest front in that ongoing conversation.
For voters like Johnson, the appeal lies in the personalization. A chatbot can tailor its explanations to a user's stated values and priorities, breaking down complex races into digestible summaries. It can also answer follow-up questions immediately, something static voter guides cannot do. That interactivity may make political research feel less daunting for people who might otherwise skip down-ballot races entirely.
Still, election experts generally encourage voters to cross-check AI-generated information against official sources, such as state and county election offices, candidate websites, and nonpartisan voter guides. Treating a chatbot as a starting point rather than a final authority can help mitigate the risk of acting on inaccurate claims.
The phenomenon also highlights how quickly AI has moved from a niche technology into everyday routines. Johnson's use of ChatGPT for meal planning and job searching reflects a broader normalization of AI assistants in American households. As those tools become more embedded in daily life, their use in civic contexts may continue to grow, whether or not companies actively promote it.
With voting already underway in the midterms, the coming weeks will serve as an informal test of how much influence AI chatbots wield over voter decisions. Some voters will find them helpful for organizing information and clarifying their own priorities. Others may encounter limitations, refusals, or errors that undermine their trust in the technology.
What is clear is that the relationship between voters and information is changing. The tradition of studying a ballot with a newspaper, a pamphlet, or a trusted friend now exists alongside a new option: an hours-long conversation with a machine. How that shift affects turnout, ballot completion, and the quality of democratic decision-making is a question that will not be fully answered until the votes are counted.
(0)