As early voting gets underway across the country for this year's midterm elections, a notable shift is taking place in how some Americans are preparing to cast their ballots. Rather than relying solely on traditional sources such as news outlets, candidate debates, or conversations with friends and family, a growing number of voters are turning to artificial intelligence chatbots to help them research races and weigh candidate positions. The trend represents a new frontier in the intersection of technology and civic participation, raising both practical questions and broader concerns about how voters form their decisions.
One illustrative example comes from Morgantown, West Virginia, where Adam Johnson, a 40-year-old graphic designer who is studying to become an illustrator, spent a September evening on his porch reviewing his ballot. Instead of discussing the various races and candidates with another person, Johnson engaged in an hours-long conversation with ChatGPT, the chatbot developed by OpenAI. He said the tool has become a regular part of his daily life, assisting him with tasks ranging from meal planning to job hunting and household chores.
Johnson's approach was methodical. He first asked the chatbot to compile a list of all the races that would appear on his ballot. From there, he dug deeper into individual contests, using the back-and-forth dialogue to explore where candidates stood on issues he cared about. The goal, he said, was to determine which candidates best aligned with his personal values. His experience reflects a broader pattern in which voters are using AI not just for isolated fact-checks but as a kind of interactive research assistant throughout the election season.
The appeal of AI chatbots for this purpose is understandable. They are available around the clock, can process large volumes of information quickly, and allow users to ask follow-up questions in plain language. For voters facing lengthy ballots filled with down-ballot races that receive little media coverage, the technology can seem like a convenient way to fill informational gaps. Supporters argue that this kind of access could potentially increase engagement, particularly among voters who might otherwise feel overwhelmed by the sheer number of decisions they are asked to make.
However, the practice also comes with significant caveats. Large language models are known to sometimes generate inaccurate or outdated information, a phenomenon often referred to as hallucination. When applied to something as consequential as voting, even small errors could mislead users about a candidate's record or policy positions. The source material notes that Johnson himself encountered limitations in his earlier interactions with the chatbot, recalling that on a previous occasion when he raised political topics, the AI seemed reluctant to engage and offered a simple refusal rather than substantive discussion.
That hesitation is not accidental. Many AI companies have designed their chatbots to be cautious about political content, often declining to endorse candidates or provide definitive voting recommendations. These guardrails are intended to avoid undue influence and to sidestep the risk of spreading misinformation. Yet as the NPR reporting illustrates, users are finding ways to work around or through those boundaries, treating the chatbot as a brainstorming partner rather than an oracle.
The trend also raises questions about the broader information ecosystem. If voters increasingly rely on AI-generated summaries rather than original reporting, candidate statements, or official voter guides, the quality of their decisions may depend heavily on the underlying data the models were trained on. Those datasets can carry their own biases, gaps, and time lags. Moreover, the conversational nature of chatbots can create a false sense of confidence, making users feel they have done thorough research when they may have only skimmed the surface.
Election officials and technology watchdogs have begun paying closer attention to these developments. While there is no evidence in the source material that AI chatbots are systematically distorting the midterm elections, the mere possibility has prompted calls for greater transparency and digital literacy. Experts suggest that voters who use AI tools should verify key claims against trusted sources, such as official state voter guides, nonpartisan fact-checking organizations, and local news coverage.
For Johnson, the technology remains a useful supplement rather than a replacement for his own judgment. He described ChatGPT as a significant part of his life, but his hours-long session was ultimately aimed at clarifying his own values and priorities, not outsourcing his vote. That distinction may prove important as the practice spreads. The question facing voters, researchers, and election administrators alike is not whether AI will play a role in elections, but how to ensure it serves as a helpful aid rather than a source of confusion.
As early voting continues and Election Day approaches, the experience of voters like Johnson offers a preview of how technology may reshape civic participation in the years ahead. The coming weeks will likely bring more anecdotal reports of AI-assisted ballot research, along with further scrutiny of how these tools handle political content. For now, the phenomenon stands as another example of how rapidly evolving digital tools are woven into the fabric of everyday decision-making, including the most fundamental democratic act of all.
(0)