Half in new poll expect AI-related misinformation to affect 2024 election outcome

A little more than half of Americans believe that misinformation spread by artificial intelligence (AI) will impact the outcome of the 2024 presidential election.

A new Axios-Morning Consult AI poll found that 53 percent of respondents said misinformation spread by AI would impact who wins the upcoming presidential election, with frequent watchers of Fox, CNN and MSNBC sharing that view.

Those who voted for former President Trump in 2020 were twice as likely to say AI will decrease their trust in the election results than those who voted for President Biden. Of those who say AI misinformation will impact the election, 47 percent voted for Trump in 2020 and 27 percent voted for Biden.

Regarding election advertising that utilizes AI, which is already changing the 2024 election, 35 percent of respondents said AI will decrease their trust in election advertising. Of that group, 42 percent were Trump voters and 33 percent were Biden voters.

The latest results highlight the doubt about election integrity in the first presidential election following the Jan. 6, 2021, attack on the Capitol. However, the distrust in AI crosses party lines, with no presidential candidate having “a lot” of trust from their own party to oversee AI.

The poll found that 35 percent of Democrats had “a lot” of trust in Biden to oversee AI regulation. For Republicans, 40 percent of those who voted for Trump had “a lot” of trust in his handling of AI.

Overall, 1 in 3 adults said they are “very concerned” about AI development. The poll also found that 34 percent of respondents said humans are smarter than AI, compared to 22 percent who said AI is smarter.

However, nearly two-thirds said humans will definitely or probably reach a point where we will lose control of AI. Of that group, 54 percent said we will reach that point within five years, while 90 percent said it will be reached within the next 25 years.

The poll was conducted Aug. 10-13 among a sample of 2,203 U.S. adults and has a margin of error of plus or minus 2 percentage points.