Representative Shontel Brown of Ohio has reintroduced the Securing Elections from AI Deception Act, a bill that would require disclaimers on AI-generated campaign content and bar the use of AI to defraud people of their right to vote. The measure, which Brown first introduced in 2024, now has 30 House cosponsors and would be enforced by the Federal Trade Commission across federal, state and local campaign material. It is endorsed by the NAACP, the Lawyers' Committee for Civil Rights Under Law, and NETWORK Lobby for Catholic Social Justice.
Under the bill, AI-generated images, video, audio or text depicting a candidate, or conveying information about how and when to vote, would have to carry either the statement "AI Disclaimer: This output was generated by artificial intelligence" or a circled "AI" symbol modeled on the copyright mark. It would also prohibit using AI to deceive voters about the time, place or method of voting, eligibility rules or the counting of ballots, with fines or imprisonment for violations. The FTC would be directed to educate the public on the label and report to Congress on its uptake.
Brown's office cites a New York Times finding that more than $130 million in AI-generated ads have run this election cycle, and an analysis by Wesleyan University that 69% of ads containing AI content carry no disclaimer. The congresswoman framed the problem plainly: "voters deserve to know that the campaign content they see and hear is real." The bill's backers argue that older voter-suppression tactics, such as the false robocalls that targeted Black voters in Ohio in 2020, become more persuasive when paired with generative tools.
The federal push arrives while a patchwork of state laws is already being stress-tested, with mixed results. In Oregon, a 2024 statute requires a disclaimer when AI edits create "a materially different understanding or impression" than the original. A complaint filed this week accuses a political action committee opposing Multnomah County commissioner Julia Brim-Edwards of posting an apparently AI-altered photo that aged and darkened her image without disclosure; the PAC's designer says it was edited in Photoshop, not AI, while metadata reviewed by OPB indicated it was "made with Google AI."
That case shows how hard enforcement is in practice. Oregon has investigated only two complaints under the law, and in the one it has resolved it found no violation, ruling that cosmetic edits to a candidate's own photos did not create a materially different impression. A separate Oregon episode, involving a video of commissioner Nafisa Fai that appeared to show an upside-down flag, grew murkier still: a news outlet reported the video was AI-edited, then retracted the entire story over doubts that the image offered as proof of tampering was itself genuine.
The deeper tension is that disclosure law depends on reliably telling real from synthetic, a line that is blurring as the tools improve and as accusations of fakery become their own tactic. Brown's bill would set a single national standard where states now differ, but it faces the same practical limits on the ground, and its prospects are uncertain in a Congress that, as AI Daily has reported, remains split over whether a federal rule should override state laws at all. For this cycle, the labels will stay voluntary, and most AI ads will carry none.