Do AI political ads need a label? The federal, state and platform rules before the November 3, 2026 midterms

Do AI political ads need a label? For the November 3, 2026 midterms, the answer depends on where the ad runs and who runs it, because no federal rule requires one. A patchwork of state laws, some of them blocked in court, and the ad policies of Google and Meta fill the gap. On October 21, 2026, Rep. Shontel Brown (D-Ohio) announced a bill, introduced on October 19, that would create a federal label for AI-generated campaign content, enforced by the Federal Trade Commission. Most of the content these rules target comes from image and video generators.

This explainer is based on Rep. Brown's announcement, Federal Election Commission and Federal Communications Commission documents, Public Citizen's state tracker, court orders, platform policy pages and reporting by the Associated Press, checked October 22, 2026, with the bill text and FEC status re-checked October 23. It describes the rules; it isn't legal advice. In the EU, by contrast, the AI Act's Article 50 transparency duties for AI-generated content have applied since August 2, 2026, as our EU AI Act explainer sets out.

The short version

• Federal: no AI-specific disclosure rule. The FEC's existing ban on fraudulently misrepresenting a candidate applies to AI, it said in 2024. The FCC proposed broadcast AI disclosures in 2024 but never adopted them.
• States: Public Citizen lists 33 states with election deepfake laws, as of October 22, 2026. Most require a label within a set period before an election; a few ban deceptive deepfakes outright.
• Courts: in 2025 a federal judge struck down California's 2024 deepfake law AB 2839 and barred enforcement of AB 2655 against X and Rumble. Hawaii's law is also listed as struck down.
• Platforms: Google and Meta require advertisers to disclose AI in election ads. TikTok bans paid political ads entirely.
• New bill: Brown's Securing Elections from AI Deception Act (H.R. 10841) would require a standard "AI Disclaimer." It has 30 cosponsors, was referred to the House Energy and Commerce Committee, and hasn't passed either chamber.

Is there a federal law requiring AI labels on political ads?

No. Congress hasn't passed one, and neither federal election regulator has written one.

The FEC. On September 19, 2024, the Federal Election Commission declined to start an AI rulemaking requested by Public Citizen. Instead it adopted an interpretive rule saying the federal ban on fraudulent misrepresentation, 52 U.S.C. § 30124, covers AI-made content because "the statute, and the Commission's implementing regulation, is technology neutral." That law bars candidates and their agents from pretending to speak for another candidate or party to damage them, and bars anyone from falsely claiming to speak for a candidate to raise money. It doesn't require labels.

The FEC also can't do much enforcement. In an April 6, 2026 filing, it said it had been without a quorum since May 1, 2025, and needs four votes to open investigations. The two nominees announced on February 11, 2026 were still unconfirmed as of August 2026, the law firm Ashurst Perkins Coie reported, and the FEC's press releases had announced no new commissioners as of October 23, 2026.

The FCC. In 2024, under then-Chair Jessica Rosenworcel, the FCC proposed requiring broadcasters, cable operators and satellite providers to ask whether political ads contain AI-generated content, and to air an on-air disclosure if so (MB Docket No. 24-211, published August 5, 2024). The proposed wording was: "The following message contains information generated in whole or in part by artificial intelligence." The proposal would not have covered online ads. No final rule has been published. The AP reported on September 25, 2026 that Congress and federal regulators still hadn't acted, citing Travis Ridout of the Wesleyan Media Project.

The FCC has acted on AI robocalls: a February 2024 ruling treats AI-generated voices as artificial voices under the Telephone Consumer Protection Act, which requires consent for many calls, Newsweek reported on October 8, 2026.

What would Rep. Brown's bill do?

The Securing Elections from AI Deception Act, H.R. 10841, was introduced on October 19, 2026 with 30 House cosponsors, referred to the House Energy and Commerce Committee and announced by Brown on October 21. It is a new version of a bill Brown first introduced in the previous Congress, in June 2024 (H.R. 8858). It would:

• require a disclaimer on AI-generated images, text, video or audio that simulates or alters a candidate for federal, state or local office, or that gives information about when, where or how to vote or register;
• let the label be either "AI Disclaimer: This output was generated by artificial intelligence" or a circled "AI" symbol resembling the copyright mark;
• have the FTC enforce the label as an unfair or deceptive practice, educate the public and report to Congress (the FTC's investigative tools are explained in our story on its civil investigative demands to OpenAI and Anthropic);
• ban using AI to deprive or defraud people of their right to vote, with fines or imprisonment.

The NAACP, the Lawyers' Committee for Civil Rights Under Law and NETWORK Lobby endorsed it. "This isn't complicated: voters deserve to know that the campaign content they see and hear is real," Brown said. Introduction alone changes nothing for the November 3 election. For another federal AI bill making the rounds, see Anthropic's proposed AI safety bill.

Which states require AI labels on political ads?

Public Citizen, which tracks election deepfake legislation, lists 33 states with at least one enacted law as of October 22, 2026. That count includes laws later struck down or expired.

Last verified: October 22, 2026

The laws differ a lot. Pluribus News reported in July 2026 that most require a label within a set number of days before an election and let candidates go to court to get unlabeled ads taken down. Minnesota and Texas prohibit distributing deceptive deepfakes within a window before an election, and Maryland bans them year-round. Michigan and Oregon exempt satire and parody. The AP reported that New York's law requires disclosure of AI in political communications, while Tennessee's covers only deepfakes that impersonate candidates.

Why were California's laws struck down?

Both were challenged in Kohls v. Bonta, a case brought by Christopher Kohls and joined by X Corp. and Rumble, among others. On August 20, 2025, Senior US District Judge John A. Mendez entered a final judgment that AB 2655, which required large platforms to block or label deceptive election content, is preempted by Section 230 of the Communications Decency Act as applied to X and Rumble. He permanently barred the state from enforcing it against them. In a separate ruling reported by the Daily Journal on September 2, 2025, Mendez struck down AB 2839, which barred distributing deceptive media of candidates near elections, writing that "AB 2839 discriminates based on content, viewpoint, and speaker." He had already blocked it with a preliminary injunction on October 2, 2024.

The AB 2839 ruling turned on the First Amendment, which protects political speech strongly. California's separate labeling law for political ads, AB 2355, remains in effect, Pluribus reported.

What do Google, Meta, YouTube and TikTok require?

Last verified: October 22, 2026

Platform rules apply only on those platforms, and they're enforced by the companies, not by election officials. Meta's own enforcement can misfire, as the No Kings Instagram suspension showed.

Are campaigns actually labeling AI ads?

Mostly not, according to the Wesleyan Media Project, which tracks political advertising. The AP reported on September 25, 2026 that the project had identified at least 164 political ads made or enhanced with AI this cycle, and 69% didn't disclose it. Ads in states with AI disclosure laws carried a disclaimer 29% of the time, versus 32% in states without such laws. Eighty percent came from Republican candidates or groups.

Gizmodo reported on October 16, 2026, citing a later Wesleyan release, that the count had reached 220 ads in 34 states with more than $137 million in spending, and that 27% included an AI disclaimer. The project's release itself wasn't available to check.

What's confirmed, reported and uncertain

• Confirmed: the FEC's 2024 interpretive rule and its lack of a quorum as of April 2026, with no new commissioners announced by October 23; the FCC's unadopted 2024 proposal; the Public Citizen count; the California AB 2655 judgment; platform policies; Brown's bill text (H.R. 10841).
• Reported: the AB 2839 ruling (Daily Journal); how state laws differ (Pluribus News, AP); Wesleyan's ad counts (AP, Gizmodo); FEC nominee status as of August 2026.
• Uncertain: whether the FEC regains a quorum before November 3; whether Brown's bill gets a hearing; how states will enforce their laws against ads already aired.