Federal AI rules for political campaigns

February 8, 2024

The FCC confirmed that the Telephone Consumer Protection Act's limits on calls using an "artificial or prerecorded voice" cover AI technologies that generate or clone human voices. Any call that uses an AI voice therefore needs the called party's prior express consent unless it is an emergency call or fits an existing exemption. The ruling says the TCPA has no carve out for technologies that claim to be the equivalent of a live agent, relying on the FCC's 2020 Soundboard Ruling that a live person choosing which recorded clips to play does not make the call a live call.

Why campaigns care: An AI voice agent that talks to voters is treated like a robocall, not a live volunteer, so cell phone calls need prior express consent even for political calls. Each unlawful call can bring $500 in statutory damages, up to $1,500 if willful, under 47 U.S.C. 227(b)(3).

Official source47 CFR 64.1200 (Cornell LII)47 U.S.C. 227 (Cornell LII)

December 30, 2020

Calls to cell phones made with an autodialer or an artificial or prerecorded voice (now including AI voices) need prior express consent, and political calls get no exemption (47 CFR 64.1200(a)(1)(iii)). Artificial or prerecorded voice calls to residential landlines that are not commercial, such as most political calls, need no consent, but the caller may make no more than three such calls to a line within any 30-day period and must honor opt-out requests (64.1200(a)(3)(ii)). Every artificial or prerecorded voice message must name the responsible entity at the start and give a callback phone number during or after the message (64.1200(b)(1) and (b)(2)), and messages sent under the residential exemptions must offer an automated voice or key press opt-out within two seconds of the identification, with a toll-free opt-out number when left on voicemail (64.1200(b)(3)). Political calls are not telephone solicitations, so the National Do Not Call Registry and the solicitation calling-hours rule in 64.1200(c) do not apply to them.

Why campaigns care: An AI voice phonebank must open by naming the campaign or committee, give a callback number, provide a working opt-out, and stay within three calls per 30 days to each landline without consent. Calls to cell phones without prior express consent are the main liability risk.

Official sourcedocs.fcc.govwww.federalregister.gov

November 21, 2022

The FCC ruled that ringless voicemail delivered straight to a wireless subscriber's voicemail box is a call made using an artificial or prerecorded voice. It therefore needs the same prior express consent as any other prerecorded or artificial voice call to a cell phone. The FCC said the ruling applies to any provider that uses the subscriber's mobile number to drop a message into that mailbox.

Why campaigns care: Dropping AI-voiced or prerecorded political voicemails onto cell phones without ringing them is not a loophole. It needs prior express consent like any other robocall to a cell phone.

Official sourceFederal Register, Dec. 14, 2022 publication

August 8, 2024

The FCC proposed defining an "AI-generated call" and requiring callers to disclose at the start of a call that it uses AI-generated technology. It also proposed that consent to receive robocalls must specifically cover AI-generated calls, and asked how to protect beneficial uses such as AI tools that help people with speech or hearing disabilities make calls.

Why campaigns care: If adopted, AI voice calls would need an on-call AI disclosure and AI-specific consent language. Today it is not binding, but it signals where federal rules could go.

Official sourceWilson Sonsini summary of the FCC AI robocall NPRM

July 25, 2024

The proposal would have required broadcasters and other FCC-regulated programmers (cable, satellite TV and satellite radio) to ask political ad buyers whether an ad contains AI-generated content, to air an on-air disclosure when it does, and to note it in the public political file. It would not have covered online ads, which the FCC does not regulate.

Why campaigns care: There is no federal rule requiring an AI disclaimer on broadcast political ads. State laws, not the FCC, are what currently require AI disclosures in political ads.

Official sourcewww.federalregister.govCarr dissent, FCC 24-74

August 21, 2024

Lingo Telecom, the voice provider that carried the January 2024 AI-cloned Biden robocalls to New Hampshire voters, agreed to pay a $1,000,000 civil penalty and adopt a compliance plan. The FCC said Lingo gave the spoofed calls the highest STIR/SHAKEN caller ID attestation level without verifying them, making them less likely to be blocked.

Why campaigns care: Carriers and dialer platforms are now under pressure to vet who is sending political robocalls and whose caller ID is used. Expect stricter know-your-customer checks before AI or prerecorded political calls are carried.

Official sourceFCC DA 24-790 Order and Consent Decree

September 26, 2024

The FCC fined political consultant Steve Kramer $6,000,000 under the Truth in Caller ID Act for spoofed robocalls that used an AI clone of President Biden's voice to tell New Hampshire voters not to vote in the January 2024 primary. Kramer did not respond to the earlier Notice of Apparent Liability. Unpaid FCC forfeitures can only be collected through a Justice Department lawsuit.

Why campaigns care: This is the leading federal enforcement action against an AI political robocall. It shows that a consultant who orchestrates the calls, not just the carrier, can face multimillion dollar penalties.

Official sourceFCC 24-59 Notice of Apparent Liability

December 11, 2025

The order tells the FCC Chairman to open a proceeding on whether to adopt a federal reporting and disclosure standard for AI models that preempts conflicting state laws. Chairman Carr said the FCC would open a proceeding and was open-minded on the outcome. Separately, a September 2025 FCC notice of inquiry on wireline deployment asked whether state or local laws limiting uses of AI should be preempted, and a bipartisan group of 23 state attorneys general told the FCC it lacks that authority.

Why campaigns care: Any FCC attempt to preempt state AI disclosure laws would face legal challenge. For now, state rules on AI disclosure in calls and ads still apply.

Official sourceEO 14365 Section 6broadbandbreakfast.com

September 30, 2026

The order lets callers treat an opt-out as applying only to the category of informational robocalls it responded to, rather than to every robocall from that caller, and lets callers designate exclusive opt-out methods (an automated voice or key press mechanism on the call, standard reply keywords for texts, or a designated website or phone number). It replaces the twice-delayed "revoke all" rule that had been set to take effect January 31, 2027. The further notice seeks comment on opt-out timing, two-way texting and treatment of affiliates.

Why campaigns care: Campaigns using consent-based robocalls, AI calls or texts must still honor opt-outs quickly, but can steer them through a designated in-call or keyword opt-out channel once the rule takes effect.

Official sourcewww.burr.comwww.globalpolicywatch.com

November 20, 2025

A Belknap County Superior Court jury acquitted Kramer of 11 felony voter suppression counts and 11 misdemeanor counts of impersonating a candidate; his defense argued the unsanctioned primary was not covered and the call did not name Biden as a declared candidate. In the federal civil suit brought by the League of Women Voters of New Hampshire and voters, the court entered a default judgment against Kramer under the Voting Rights Act and state law, awarded $7,500 to each of three voter plaintiffs ($22,500 total), and permanently barred him from deceptive or intimidating AI robocall campaigns in any U.S. election. A consent judgment entered on 2025-05-23 resolved claims against other defendants.

Why campaigns care: Section 11(b) of the Voting Rights Act (voter intimidation) can reach deceptive AI robocalls, and private groups can sue for damages and nationwide injunctions even when criminal charges fail.

Official sourceDefault judgment and report and recommendation (LWV)NHPR: Kramer found not guilty (June 13, 2025)

June 20, 2025

In Facebook v. Duguid (592 U.S. 414, decided 2021-04-01) the Court held that an "automatic telephone dialing system" must have the capacity to store or produce numbers using a random or sequential number generator, which narrowed autodialer liability. In McLaughlin Chiropractic v. McKesson (606 U.S. 185, decided 2025-06-20) the Court held that district courts in TCPA enforcement cases are not bound by the FCC's interpretation of the statute and must decide the law's meaning themselves.

Why campaigns care: Live, human-dialed calls generally avoid the TCPA consent rules, but AI or prerecorded voice calls do not, because the artificial voice alone triggers the law. After McLaughlin, courts give FCC rulings such as FCC 24-17 respect but decide independently, though the statutory text already covers "artificial" voices.

Official sourceFacebook, Inc. v. Duguid (opinion)

September 19, 2024

Acting on Public Citizen's July 2023 petition (REG 2023-02), the FEC decided not to start a rulemaking on deceptive AI in campaign ads. Instead it approved an interpretive rule stating that the existing ban on fraudulent misrepresentation applies no matter what technology is used, including AI-generated content. That federal ban covers a candidate or agent pretending to speak or act for another candidate or party in a damaging way, and anyone fraudulently soliciting contributions in another's name.

Why campaigns care: There is no federal AI disclaimer requirement for campaign ads, but using AI to falsely speak for an opposing candidate or party, or to raise money in someone else's name, can violate federal campaign finance law.

Official sourceFEC weekly digest, September 16 to 20, 2024CRS Insight IN12222: AI and Campaign Finance Policy

February 11, 2026

The FEC needs four commissioners to issue rules, advisory opinions or enforcement decisions. It lost its quorum on April 30, 2025 and has had only two commissioners (Shana Broussard and Dara Lindenbaum) since Commissioner Trey Trainor's resignation took effect on October 3, 2025. President Trump nominated two Republicans on February 11, 2026, but they have not been confirmed.

Why campaigns care: The FEC cannot open enforcement on AI misrepresentation complaints, issue advisory opinions on AI ads, or adopt any AI rule for the 2026 midterms. Complaints can still be filed and the statute of limitations keeps running.

Official sourceCRS R45160: FEC Membership and Policymaking Quorum, In BriefSenate: Nominations in Committee (Civilian)

March 31, 2025

The bill would ban knowingly distributing materially deceptive AI-generated audio or video of federal candidates to influence an election or raise money. Depicted candidates could sue for injunctions and damages. It includes exceptions for news coverage and satire or parody.

Why campaigns care: If enacted it would be the first federal ban on deceptive AI deepfakes of federal candidates. It has not moved out of committee.

Official sourceCongress.gov H.R.5272GovInfo bill status XML S.1213

July 23, 2026

The bill would amend the Federal Election Campaign Act to require political ads to carry a statement when generative AI was used to create images, audio or video in the ad. A Senate version (S.3875) passed the Senate Rules Committee in the 118th Congress but was not enacted.

Why campaigns care: This is the main federal proposal for an AI disclaimer on political ads. Until it passes, only state laws require such disclaimers.

Official sourceGovInfo bill status XML H.R.9922GovTrack S.3875 (118th Congress)

June 11, 2026

The bill would prohibit distributing false AI-generated election media, defined to include AI-made text, images, audio or video with materially false information about federal elections. It would also amend the National Voter Registration Act to bar removing voters from rolls based on unverified voter challenge databases.

Why campaigns care: It would reach AI-generated content that misleads voters about voting, not just candidate deepfakes. It has Democratic sponsors only and has not moved.

Official sourceGovInfo S.4774 introduced text

June 24, 2026

The bill would create a federal right to control unauthorized digital replicas of a person's voice or likeness, with liability for people and platforms that knowingly make or distribute them, and a notice-and-takedown system for online services. It includes exceptions for news, commentary, parody and other First Amendment uses. The White House March 2026 framework endorsed a federal digital replica right, and reports in September 2026 described it as part of a possible preemption package.

Why campaigns care: A cloned voice of a real person in an ad or call could create federal civil liability if this passes. It is the furthest advanced AI likeness bill in Congress.

Official sourceGovInfo bill status XML S.4591GovInfo bill status XML H.R.8915

December 4, 2025

The bill would require any robocall that uses AI to emulate a human being to say at the start of the call that AI is being used. It would also increase TCPA penalties for using AI to impersonate a person or organization with intent to defraud or cause harm.

Why campaigns care: If enacted, AI voice phonebank calls would need an upfront AI disclosure under federal law. Several states already require similar disclosures.

Official sourceCongress.gov S.3354 textRep. Sorensen press release on reintroduction

July 1, 2025

The House-passed reconciliation bill included a 10-year moratorium on enforcing state and local AI laws, later tied to federal broadband (BEAD) funding in the Senate draft. A Blackburn and Cantwell amendment removed it by a 99 to 1 vote before the bill passed. A second attempt to add preemption to the annual defense bill (NDAA) was dropped in early December 2025.

Why campaigns care: State AI election laws, including deepfake disclaimer and AI robocall laws, stayed fully enforceable. Congress has twice declined to preempt them.

Official sourcewww.nextgov.comStateScoop: moratorium omitted from 2026 defense bill

September 30, 2026

H.R.5388, the American Artificial Intelligence Leadership and Uniformity Act (Rep. Baumgartner, introduced 2025-09-16), would impose a temporary moratorium preempting certain state AI laws. A bipartisan House discussion draft, the Great American Artificial Intelligence Act (released 2026-06-04 by Reps. Obernolte, Trahan and others), would preempt for three years state laws specifically regulating AI model development but not laws on AI use. Democrats' GUARDRAILS Act (Sen. Schatz, S.4216, 2026-03-26; Rep. Beyer, H.R.8031, 2026-03-20) would void EO 14365. In September 2026 a frontier AI bill negotiated by Sens. Cruz, Thune and Klobuchar, and a reported White House offer pairing preemption with the Kids Online Safety Act and NO FAKES Act, remained unreleased.

Why campaigns care: Proposals so far focus on AI model development, so state election deepfake and AI call laws would likely survive most drafts, but campaigns should watch for any year-end package.

Official sourceGovInfo bill status XML S.4216 (GUARDRAILS Act)GovInfo bill status XML H.R.8031 (GUARDRAILS Act)

December 20, 1991

Except for the federal technical standards in subsection (d), the TCPA and FCC rules do not preempt any state law that imposes more restrictive intrastate requirements on, or prohibits, the use of autodialers, artificial or prerecorded voice messages, or telephone solicitations. States such as Florida, Oklahoma and Maryland have their own "mini-TCPA" laws on automated telemarketing calls, and many states have their own rules on political robocalls and, increasingly, AI voice disclosures.

Why campaigns care: Federal TCPA compliance is only the floor. An AI or prerecorded political call must also satisfy the rules of the state being called, which can be stricter.

Official source

July 23, 2025

The plan says federal AI-related funding should not go to states with "burdensome AI regulations" while not interfering with states' right to pass prudent laws. It directs OMB to have agencies consider a state's AI regulatory climate in discretionary funding decisions, and asks the FCC to evaluate whether state AI rules interfere with its duties under the Communications Act. It also calls for tools to counter malicious deepfakes used as evidence in the legal system.

Why campaigns care: It set up the funding pressure and FCC preemption ideas that the December 2025 executive order later formalized. By itself it changes no campaign obligations.

Official source

December 11, 2025

The order directs a DOJ AI Litigation Task Force to challenge state AI laws (within 30 days), a Commerce evaluation listing "onerous" state AI laws (within 90 days), an NTIA policy notice making states with such laws ineligible for BEAD non-deployment broadband funds, an FTC policy statement on state laws that require altering truthful AI outputs, and an FCC proceeding on a federal AI reporting and disclosure standard that would preempt conflicting state laws (within 90 days after the Commerce list). It also orders a legislative recommendation for federal preemption that would not preempt state child safety laws, AI compute and data center infrastructure rules (other than general permitting), or state procurement and use of AI.

Why campaigns care: Only courts or Congress can actually displace state AI election laws. State deepfake and AI disclosure rules for political ads and calls remain enforceable unless a court blocks a specific law.

Official sourceFederal Register, Dec. 16, 2025 (EO 14365)FCC Chairman Carr statement on the AI executive order

March 20, 2026

The White House asked Congress to preempt state AI laws that impose undue burdens and set one minimally burdensome national standard. It said preemption should not reach states' generally applicable laws (including laws protecting children, preventing fraud and protecting consumers), state zoning of AI infrastructure, or a state's own use of AI. It also recommended a federal right against unauthorized AI digital replicas of a person's voice or likeness with exceptions for parody, satire and news, and more law enforcement against AI impersonation scams.

Why campaigns care: The fraud and consumer protection carve outs suggest state laws against deceptive AI impersonation may survive any federal preemption bill, but the final text is up to Congress.

Official sourcewww.morganlewis.com

April 24, 2026

Attorney General Bondi created the Task Force on January 9, 2026 to challenge state AI laws on Commerce Clause, preemption or other grounds. Its first visible step was moving to intervene in xAI's April 2026 suit against Colorado's AI Act (SB 24-205); on April 27, 2026 the court approved a joint motion suspending enforcement. Colorado then repealed and replaced that law with SB 26-189 (signed May 14, 2026), a narrower automated decision-making law effective January 1, 2027.

Why campaigns care: So far the federal litigation campaign has targeted broad AI governance laws, not state election deepfake or AI robocall laws. No state political AI law has been challenged by DOJ.

Official sourcewww.nortonrosefulbright.comwww.jenner.com

March 11, 2026

The order requires Commerce to publish an evaluation identifying state AI laws that conflict with federal policy, at minimum laws that require AI models to alter truthful outputs or compel disclosures that may violate the First Amendment. That list is meant to trigger DOJ referrals, BEAD funding consequences and the FCC proceeding. Members of Congress urged Commerce to include Colorado's AI Act and New York's RAISE Act.

Why campaigns care: Without a published list, no state has been formally designated, and the BEAD and FCC steps tied to the list have not been triggered.

Official sourcewww.butzel.comwww.hklaw.com

March 11, 2026

The order directs NTIA to issue a policy notice making states with onerous AI laws ineligible for leftover BEAD non-deployment broadband funds (about $20 billion nationally) to the extent federal law allows. NTIA delayed the notice past its March 2026 deadline; in June 2026 the NTIA Administrator told Congress no final decision had been made, and in September 2026 NTIA said further non-deployment guidance was not coming soon.

Why campaigns care: No state has lost broadband money over its AI laws, so there is no current federal funding pressure forcing states to repeal AI election rules.

Official sourcerollcall.combroadbandbreakfast.com

July 1, 2026

Issued under EO 14365, the statement says AI companies that distort outputs toward undisclosed objectives, including ideological ones, may be deceiving consumers in violation of Section 5 of the FTC Act. It says complying with a state law is no excuse for an undisclosed distortion and suggests state laws that require such alterations may be preempted, naming Colorado's AI Act as an example.

Why campaigns care: It is part of the federal push against state AI laws but is aimed at AI developers and state algorithmic bias laws, not campaign ads. It does not change disclosure rules for political AI content.

Official sourceFTC proposed policy statement (PDF)Arnold and Porter analysis