Minnesota officials warn of danger that AI could pose to elections

Top Democrats in Minnesota are sounding alarm over the danger artificial intelligence could pose to elections and urging Congress to regulate the rapidly evolving technology.

While AI could be helpful for day-to-day operations, leading Democrats say it should not be used to deceive or manipulate voters. The Legislature passed a law earlier this year regulating so-called "deepfake" technology, prohibiting it from being used to depict sexual content or manipulate elections. Now, Minnesota Democrats are urging the federal government to follow suit.

"I think the fear is that in a high-stakes … presidential contest, that there will be an incentive for bad actors to use AI in ways that distort and destabilize the election system," Minnesota Secretary of State Steve Simon said in an interview. "That's what we're up against."

Deepfakes of famous politicians doing things they never actually did have already started circulating ahead of next year's election. Florida Gov. Ron DeSantis' presidential campaign attempted to attack former President Donald Trump with an ad that used fake AI-generated images depicting him hugging Dr. Anthony Fauci. And U.S. Sen. Elizabeth Warren, D-Mass., was the victim of a separate deepfake video that falsely depicted her saying Republicans should not be allowed to vote in next year's election.

Simon shared his concerns about such content during a U.S. Senate committee hearing in Washington, D.C., on Wednesday. The hearing, focused on AI and how it could impact elections, was chaired by Minnesota Democrat Amy Klobuchar.

Klobuchar and a bipartisan group of senators, including GOP Sen. Josh Hawley of Missouri, have introduced legislation to ban deceptive AI-generated content in political ads. She also has pushed a bill that would require a disclaimer on political ads that use AI-generated content.

"With AI, the rampant disinformation we have seen in recent years will quickly grow in quantity and quality," Klobuchar said during the hearing. "We need guardrails to protect our elections."

"When we do end-of-the-year budget work and the like, that would be a vehicle for this," Klobuchar said. "That's what I'm working with my Republican co-sponsors on."

At the same time, politicians aren't dismissing AI's potential to help improve efficiency and productivity.

Simon said advanced chatbots could aid his office and other secretaries of state in administering elections.

"Every election, we run a phone center where we take in thousands and thousands of voter calls on various matters," Simon said. "It is certainly possible that a really highly trained and highly accurate chatbot could theoretically help with some of those questions."

AI is "not inherently a bad thing" if used the right way, Simon said. He added that his office isn't currently using the technology.

"I spend about two to three hours a day raising money on the phone," Martin said. "At least 60% of that time is leaving messages."

The party is also considering adding an online chatbot that can assist people who visit its websites, Martin said.

"We will use it in a very limited capacity," Martin said of the DFL's AI use. He added that he's concerned some political campaigns will use AI to deceive people.

A spokeswoman for the Republican Party of Minnesota did not respond to multiple requests for comment about whether the party will use AI next year.

The Minnesota House Republican Campaign Committee, which will seek to help Republicans win back control of the state House next year, has no plans to use AI but is exploring its options.

Craig, who represents Minnesota's highly competitive Second Congressional District, said in a statement that she supports "increased transparency requirements when generative AI is used in campaign advertisements."

"We should respect voters enough to be clear about where the information in campaign ads is coming from," Craig said.