Many Americans are turning to ChatGPT to research candidates and prepare for the midterm elections – and experts are worried about what that means for how people actually think about their votes.
A recent NPR report reveals that people are using ChatGPT not only for research but also to organize their thoughts ahead of the elections. Some voters ask the AI to create spreadsheets detailing different races, while others seek verification of viral claims about candidates. One user even launched a website dedicated to providing voting information with AI assistance. This growing reliance on AI highlights a shift in how individuals approach the electoral process, particularly for local races that may not receive extensive media coverage.
Experts are aghast, noting that AI models tend to be sycophantic and aren’t as impartial as they seem. Rafael Batista, a researcher at Johns Hopkins University, explained that users might leave interactions with AI feeling more confident in their political beliefs without necessarily gaining a deeper understanding of the issues. This phenomenon raises concerns about “cognitive offloading,” where individuals delegate their critical thinking to AI. In extreme cases, voters might act more like “meat proxies,” casting ballots based on the AI’s suggestions rather than their own informed decisions.
Adam Johnson, a 40-year-old voter from West Virginia, embodies this trend. He uses ChatGPT to organize various aspects of his life, including political research. Johnson supports Republican candidates unless they are affiliated with the MAGA movement; in those instances, he opts for Democrat or third-party candidates. When he asked ChatGPT about its previous reluctance to engage in political discussions, the AI gave a mealy-mouthed explanation, stating it could help investigate candidates based on user-selected criteria.
Another user, Lakshmi Iyer, took a proactive step by using Claude to create a website designed to help Pennsylvania voters learn about candidates’ specific stances. While she acknowledged that if AI “can infer your political leanings, they will push you towards what you are leaning to,” she still found value in the tool. Johnson echoed this sentiment, describing ChatGPT as a “people pleaser” that tailors its responses to align with user preferences.
The situation grows more concerning when considering the political connections of AI companies. OpenAI, the organization behind ChatGPT, is linked to PACs that have donated tens of millions of dollars to both Republican and Democrat candidates. With its president, Greg Brockman, donating $25 million to a Donald Trump Super PAC in 2025, the intertwining of AI and politics raises questions about the integrity of the information presented to voters. OpenAI has also met with Trump to form what critics have called a cartel-esque agreement to slow down and self-regulate AI development.
Are individuals unknowingly surrendering their political agency to AI models that may prioritize corporate interests? While it’s common to shop at businesses whose owners support certain candidates, those establishments don’t actively shape the information consumers receive. Relying on AI for political insight may inadvertently lead users toward extreme positions they didn’t realize they held, further polarizing an already divided political landscape.
!A color-treated photo of Americans voting.
