Senate committee passes three bills to safeguard elections from AI, months before Election Day

The Senate Rules Committee passed three bills that aim to safeguard elections from deception by artificial intelligence, with just months to go before Election Day. The bills would still need to advance in the House and pass the full Senate to become law, creating a time crunch for rules around election-related deepfakes to take effect before polls open across the country in November.

The three election bills passed by the Senate Rules Committee on Wednesday mark an early step at the federal level to take action on AI in elections. Chair Amy Klobuchar (D-MN), who sponsors the bills, noted that states have already moved forward on this issue for state-level elections. For example, 14 states have enacted a form of labeling of AI content, according to Klobuchar.

The measure with the most support in the committee, the Preparing Election Administrators for AI Act, which passed 11–0, would direct the Election Assistance Commission (EAC) to work with the National Institute of Standards and Technology (NIST) to create a report for election offices about relevant risks of AI to disinformation, cybersecurity, and election administration. It also included an amendment requiring a report on how AI ends up impacting the 2024 elections.

The two other bills, the Protect Elections from Deceptive AI Act and the AI Transparency in Elections Act, passed 9–2 out of the committee. The first would prohibit AI deepfakes of federal candidates in certain circumstances when used to fundraise or influence an election and is co-sponsored by Sens. Josh Hawley (R-MO), Chris Coons (D-DE), and Susan Collins (R-ME). The second, co-sponsored by Sen. Lisa Murkowski (R-AK), would enforce a disclaimer on political ads that have been substantially created or altered by AI (it would not apply to things like color editing or resizing, for example). While the Protect Elections from Deceptive AI Act could not regulate satire, Klobuchar noted that the AI Transparency in Elections Act would at least let voters know when satire ads are AI-generated.

Ranking Member Deb Fischer (R-NE), who opposed the latter two bills, said they were “over-inclusive, and they sweep in previously unregulated speech that goes beyond deepfakes.” Fischer said the Protection Elections from Deceptive AI Act would restrict unpaid political speech, adding that “there is no precedent for this restriction in the 50-year history of our federal campaign finance laws.” Fischer also said that state legislatures are a more appropriate venue for these kinds of election regulations rather than the federal government.

But key Democrats on the committee urged action. Senate Intelligence Committee Chair Mark Warner (D-VA) said he’s, “in many ways, afraid in 2024 we may be less protected than we were in 2020.” He said that’s because “our adversaries realize that interference in our elections is cheap and relatively easy,” and Americans “are more willing to believe certain outrageous theories these days.” Compounding that is the fact that “AI changes the whole nature and game of how a bad actor … can interfere using these tools.”

“If deepfakes are everywhere and no one believes the results of the elections, woe is our democracy,” Schumer said during the markup. “I hope my colleagues will think about the consequences of doing nothing.”

At a press conference on the AI roadmap after the markup, Schumer noted the committee passage and said they’d “like to get that done in time for the election.”