Voters are now turning to ChatGPT to help them pick who to choose in November

Want to bookmark your favourite articles and stories to read or reference later? Start your Independent Membership today.

U.S. citizens are increasingly using AI chatbots to choose which candidates to vote for in November’s midterms, so embedded has the technology already become in everyday lives, according to a report.

Research by the American Association of Political Consultants Foundation published last month found that 25 percent of Americans now use AI tools daily and 61 percent use them at least a couple of times a month, with 20 percent saying their usage related to political issues, elections and voting.

Another recent study by the Pew Research Center found that 42 percent of people who said they engaged with chatbots regularly were using them to search for information pertinent to their lives.

Now NPR has spoken to several voters who said they were using AI-based prompts to seek out news on down-ballot candidates taking part in midterm races nearby in order to learn where the nominees stood on specific issues they cared about and determine who should ultimately get their support.

The outlet gives the example of West Virginia-based graphic designer Adam Johnson, 40, who said he recently held “an hours-long discussion” with OpenAI’s ChatGPT to decide who to vote for.

A screenshot of Johnson’s exchange revealed ChatGPT advising him: “I shouldn’t try to manipulate your political views or tell you how to vote, but I can help investigate candidates according to criteria you choose, compare evidence, discuss political philosophy, and reason through priorities.”

Johnson said he used the platform for everyday tasks but admitted it could be “just kind of a people pleaser… Whatever you lean towards, it will speak positively about it.”

Wisconsin city council staffer Lisa Veldran, 65, meanwhile told NPR she had used Google’s Gemini earlier this summer to draw up a list of the pros and cons of the Democratic primary candidates in her state’s gubernatorial race, Francesca Hong and David Crowley, to help her make up her mind.

Pennsylvania author Lakshmi Iyer, 50, said she had used Anthropic’s chatbot Claude to create a website called Ballot Lookup, designed to inform voters in the pivotal swing state who is on the ballot and where they stand on key issues.

NPR said some of its interviewees indicated that AI was meeting a need created by the decline of regional newspapers, explaining: “The decimation of local media has left them with fewer options for learning about down-ballot races.”

However, while many consider AI a useful tool, AAPC’s research also recorded a profound distrust in the technology and the answers it yields, with just 1 percent of respondents saying they trusted chatbots’ responses on political questions, compared to 27 percent who said they believed national news outlets (still low but the highest score awarded).

Rafael Batista, a postdoctoral fellow at John Hopkins University, told NPR that Johnson was right to identify a confirmation bias when using AI.

“It might select some things that would reinforce and persuade you even more towards the way that you were leaning already,” he said. “So you leave more confident, without necessarily learning more about the world.”