There are protections in place in Minnesota targeting artificial intelligence election misinformation for the first time heading into a presidential election.

A state law passed in 2023 makes it a crime to share AI-created content within 90 days of an election if the person being depicted did not consent and the intent is to harm a candidate’s reputation or influence election results, according to Secretary of State Steve Simon.

“I view AI as not necessarily a new problem, but I see it as a new way to magnify a very old problem, which is election misinformation and disinformation,” said Simon. “There’s every incentive for people to shade the truth in all sorts of ways, and AI is no different.”

Minnesota is one of more than a dozen states that already have a law in place to address false AI election content.

“We’re on the lookout for it,” said Simon. “To my knowledge, we have not seen AI-related sort of deepfake images in the campaign context, but it’s still early. We’ve got six plus weeks to go in the election, and anything could happen.”

Senator Amy Klobuchar, meanwhile, is working to get federal regulations in place too ahead of Election Day.

“AI has some great opportunities here but you’re going to have all of these scams going on, scams on voters where they can’t tell if it’s the person or not,” she explained.

Earlier this year, officials investigated robocalls mimicking President Joe Biden’s voice discouraging voters from going to the polls ahead of the New Hampshire primary. More recently, Taylor Swift, in her endorsement of Vice President Kamala Harris, raised concerns about AI misinformation after she said she learned of a post of herself falsely endorsing Donald Trump in the presidential race.

Trump has also been the subject of deep fake images.

“What we’re starting to see in our elections is scary,” said Klobuchar.

She is pushing to get three bipartisan bills, which she introduced over the last year, passed.

One piece of legislation would ban the use of AI-generated material to falsely depict federal candidates in ads to influence elections. Another would require political ads created or altered by AI to be labeled as such, unless the alterations are minor.

The third bill would require the Election Assistance Commission to issue guidelines to help election administrators address AI’s impact on elections.

“We need to do something on the federal level,” said Klobuchar. “Right now, we have some Republican support, not all Republican support, so I’m continuing to work to build support to get some kind of compromise passed.”

Simon supports the proposed federal laws.

“I hope they make it past the finish line because I think that added protection will help us,” said Simon.

Last month, Simon and top election officials from other states successfully petitioned social media platform X to update its AI chatbot after learning it was sharing false election information. Now, users who ask it elections-related questions are forwarded to the non-partisan CanIVote.org.

“Whether it’s AI-driven or not, we are really pushing people to trust good sources,” he said.