ai risk today logo svg bl font
Menu
SUBSCRIBE
SUBSCRIBE
ai risk today logo svg bl font
Menu

Federal Election Commission

Five candidate lawn signs on a suburban street illustrate election fraud targeted by the AI Ads Act.

AI Ads Act Targets Fraudulent AI Election Content

, , , , , , , , ,

The bill would write AI-assisted fundraising fraud into the Federal Election Campaign Act and expand impersonation rules beyond candidates and their representatives.

 

The AI Ads Act would explicitly put AI-generated election content under Section 322 of the Federal Election Campaign Act, which prohibits falsely claiming to represent a candidate or political group, including to solicit donations.

Rep. Ro Khanna (D-Calif.) introduced H.R. 9985. The House referred the bill to the Committee on House Administration. Sen. Adam Schiff (D-Calif.) introduced a Senate version.

The bill would not ban all AI use in political advertising. The content would have to fraudulently present a person or group as speaking, writing, or otherwise acting for a candidate, political party, committee, or political organization.

AI fundraising fraud would fall under the law

Federal law already prohibits a person from pretending to speak, write, or otherwise act for a candidate or political party, or their employees or agents, to solicit political contributions or donations.

H.R. 9985 would expressly state that this prohibition covers content generated in whole or in part with AI. It would also extend the rule to fraudulent solicitations made in the name of a real or fictitious political committee or organization.

The bill would not create a separate requirement to label AI-generated fundraising material. Instead, it would expand the existing federal ban on fraudulent political solicitations.

Bill would write AI into existing impersonation law

Section 322 of the Federal Election Campaign Act already prohibits two forms of fraudulent impersonation. Candidates and their representatives cannot pretend to act for another candidate or political party to damage them. The law also prohibits anyone from pretending to represent a candidate or political party to solicit contributions or donations.

The Federal Election Commission (FEC) said in 2024 that these restrictions apply whether the fraudulent content is created with AI or another technology. However, that guidance only explained how the FEC interprets the existing law. It did not change the statute, which does not mention AI.

H.R. 9985 would amend Section 322 to state directly that either violation can involve content generated “in whole or in part” with AI. This would put the FEC’s interpretation into federal law instead of leaving it only in agency guidance.

Using AI would not create a violation by itself. The content would still have to falsely present someone as speaking or acting for a candidate or political group.

Impersonation rules would cover more people and groups

The bill’s broader change would be to expand who can violate the law and who can be impersonated.

Current law generally limits the campaign-impersonation provision to candidates and their employees or representatives. It also requires the impersonation to damage another candidate or political party.

The AI Ads Act would remove both restrictions. Any person could be held responsible for falsely claiming to represent a candidate or political group, and the government would not have to prove that the impersonation was intended to damage someone.

The bill would also extend both the impersonation and fundraising provisions to false claims of representing real or fictitious political committees and organizations.

Existing enforcement and penalties would apply

The bill would not establish a separate penalty for using AI. Violations would be handled under the existing federal campaign-finance enforcement system.

Anyone could file a formal complaint with the FEC. If at least four FEC commissioners agree, the agency could investigate and try to resolve the case through corrective action and a civil penalty. If no agreement is reached, the FEC could take the case to court. It’s important to note that the FEC generally has six commissioners, but presently it only has two. With four seats vacant, the FEC could not act on complaints.

Knowing and willful violations can also result in criminal prosecution, with fines or imprisonment available under existing federal law.

There would be no automatic penalty simply because political content was created with AI. Authorities would still have to establish that the person fraudulently claimed to speak or act for a candidate or covered political group.

H.R. 9985 remains before the House Administration Committee.

Blue absentee ballot drop box illustrating proposed disclosure rules for AI political ads.

House Bill Would Set Fines Up to $50,000 for Unlabeled AI Political Ads

, , , , , , , ,

The proposal would require separate disclosures for AI political ads containing AI-created or materially altered images, audio, or video.

 

The AI Transparency in Elections Act of 2026 would let the Federal Election Commission (FEC) fine people or groups that pay for AI political ads up to $50,000 per ad if they omit a required disclosure.

Rep. Joseph Morelle (D-N.Y.) introduced H.R. 9922 on July 23, along with Reps. Rick Larsen (D-Wash.), Terri Sewell (D-Ala.), Norma Torres (D-Calif.), and Julie Johnson (D-Texas). The House referred the bill to the Committee on House Administration.

Bill would cover federal political advertising

The bill would apply to paid political advertising across broadcast, print, outdoor, mail, phone, and digital platforms.

It would cover ads that expressly support or oppose a candidate, solicit political contributions, or refer to a candidate during the period starting 120 days before a primary and ending on the general election date. Using a candidate’s voice or likeness would count as referring to that candidate.

The person or group paying for the ad would be responsible for including the disclosure.

Minor AI edits would be excluded

The requirement would apply when generative AI creates or materially alters an image, audio clip, or video.

The bill would exclude minor changes such as cosmetic adjustments, color edits, cropping, or resizing when they do not give a fundamentally different understanding of the original media.

Disclosure formats depend on the medium

The required disclosure would change with the type of content:

  • Images: The statement must say the ad was created or materially altered by AI. It must be permanently attached and use letters at least as large as most of the other text, unless it satisfies existing federal political-ad disclaimer rules.
  • Audio: A clearly audible statement must run for at least four seconds at the beginning or end of the ad. It must identify the person or group that paid and say they used AI to generate the content.
  • Video with audio: The written disclosure must remain on screen throughout the video, and the ad must also include the spoken statement required for audio.

FEC would set the penalty schedule

The FEC would set penalties based on factors including previous violations and how widely an ad was distributed. Each violation could carry a civil penalty of up to $50,000 per covered ad.

The requirements would apply to communications distributed after the bill becomes law. They would take effect even if the commission had not finished writing its regulations. The agency would have 90 days after enactment to issue those rules in consultation with the National Institute of Standards and Technology.

H.R. 9922 remains in the House Administration Committee and has not received a vote as yet.