Vermont · verified October 5, 2026

Vermont AI laws for political campaigns

Disclosure required2 AI call rules4 political AI laws

Since March 5, 2026, anyone who knowingly distributes a deceptive deepfake of a candidate, or AI media that feeds voters materially false election information, within 90 days before a Vermont election must label it with the exact statutory disclosure, or face fines and a candidate injunction suit.

Political ad AI law
Disclosure
When it applies
Within 90 days before an election in Vermont
Required AI disclaimer
Exact wording set by law
Criminal penalty
No, civil only
Candidate can sue
Yes
AI voice calls
2 state rules + TCPA

Political ads and deepfakes

Does Vermont regulate AI in political ads?

Yes. Act 75 of 2026 (S.23, 17 V.S.A. 2031 to 2033 and 2041 to 2042) bars publishing or distributing synthetic media that you know is "deceptive and fraudulent" within 90 days before a Vermont election unless it carries a set disclosure. Covered media is any image, audio, or video of a person's appearance, speech, or conduct created or intentionally manipulated with digital technology, including AI, that realistically depicts a candidate in a way that injures the candidate's reputation, or depicts anyone in a way that tries to sway an election (including a public question) with materially false information. Knowing and intentional violations carry fines from $1,000 up to $15,000, the depicted candidate can sue for an injunction, and the Attorney General or a State's Attorney can investigate and sue. A federal court blocked one AG investigation of an AI satire video in Sept. 2026 (Poitras v. Clark) but left the law in force for everyone else. Key law (S.23 (Act 75 of 2026), 17 V.S.A. 2031 to 2033 (subchapter 4), 2041 to 2042 (subchapter 5)).

S.23 (Act 75 of 2026): An act relating to the use of synthetic media in elections

In effect

Required disclaimer, word for word

“This media has been manipulated or generated by digital technology and depicts speech or conduct that did not occur.”
Citation
17 V.S.A. 2031 to 2033 (subchapter 4), 2041 to 2042 (subchapter 5)
Enacted
March 5, 2026
Effective
March 5, 2026
When it applies
within 90 days before an election in Vermont
Media covered
Images, audio recordings, and video recordings of an individual's appearance, speech, or conduct created or intentionally manipulated with digital technology, including artificial intelligence. Only "deceptive and fraudulent" synthetic media is regulated: media that appears to a reasonable person to realistically depict (A) a political candidate in a way that injures a candidate's reputation, or (B) an individual in a way that attempts to unduly influence an election, including a public question, by giving voters materially false information.
Who it covers
Any person who publishes, communicates, or otherwise distributes the media
Rule
ban unless disclosed
Intent required
The person must know the media is deceptive and fraudulent synthetic media; penalties apply to knowing and intentional violations.
Disclaimer format
Images and video: text in a size easily readable by the average viewer and accessible to people with disabilities to the greatest extent possible; for video, displayed for the full duration. Audio only: read in a clearly spoken manner, at a pitch and pace easily heard by the average listener, at the beginning and end of the audio, and if longer than two minutes, repeated at intervals of no more than two minutes.
Exemptions
  • radio, TV, cable, satellite, websites, streaming platforms, and apps airing the media in a bona fide newscast, news interview, documentary, commentary, or on-the-spot news coverage, if the broadcast clearly acknowledges questions about authenticity
  • broadcasters and platforms that are paid to broadcast the media
  • broadcasters required by federal law to air ads from legally qualified candidates
  • newspapers, magazines, periodicals, and news websites that clearly state the media does not accurately represent the person
  • satire or parody
  • telecommunications and information service providers for content provided by another person
  • interactive computer services (47 U.S.C. 230) for content provided by another person
Civil penalty
Fines for knowing and intentional violations (17 V.S.A. 2033(a)): up to $1,000; up to $5,000 with intent to cause violence or bodily harm; up to $10,000 for a repeat violation within five years; up to $15,000 for both. Obstructing an AG or State's Attorney civil investigation: fine up to $5,000.
Who can sue
A candidate whose appearance, speech, conduct, or environment is misrepresented may seek injunctive or other equitable relief to stop distribution. No damages remedy is created.
Injunction available
Yes
Court challenges
Poitras v. Clark, No. 2:26-cv-234 (D. Vt.), filed July 8, 2026, challenges Act 75 under the First Amendment (facial and as-applied) after the Attorney General served a civil investigative demand over an AI satire video of Rep. Becca Balint. On Sept. 4, 2026 Judge Sessions granted a narrow preliminary injunction: the AG may not enforce that demand (AGO Matter No. 26-SM-101) or take Act 75 action against the plaintiff over that one video. The court expressly did not rule on the law's facial validity, and said the statute's reference to criminal prosecution is likely severable. Act 75 otherwise remains enforceable; a motion to dismiss was pending as of Oct. 2026.
Notes
Signed by Gov. Phil Scott March 5, 2026 and effective on passage. The Attorney General or a State's Attorney may bring any action or injunction to prevent or abate violations of 17 V.S.A. chapter 35 and has civil investigative demand powers (17 V.S.A. 2041 to 2042). Section 2033 sets fines without labeling them criminal, but the notice for an investigative demand must warn that a knowing and intentional violation is subject to criminal prosecution; the federal court in Poitras v. Clark questioned that criminal reference and called it likely severable. The Vermont Association of Broadcasters has published guidance on the law.

Read the official text

Bills to watch

  • H.846 (2026): Platform requirements for synthetic media (per Public Citizen tracker). Failed, May 2026 (session adjourned without passage)

For campaigns and consultants

Before an AI ad runs in Vermont

  • Inside 90 days before any Vermont election, put the exact Act 75 disclosure on any AI-altered or AI-generated image, audio, or video that depicts a candidate unfavorably or could mislead voters about the election.
  • For AI audio (radio spots, prerecorded calls, voicemail drops), read the disclosure at the start and the end, and every two minutes if the audio runs longer than two minutes.
  • For AI video, keep the text disclosure on screen for the entire video, not just the end card.
  • Labeling makes it legal under Act 75 but does not shield you from defamation claims; do not depict an opponent saying things they never said.
  • Keep the usual "paid for by" identification on robotic phone calls and other electioneering communications under 17 V.S.A. 2972; the AI disclosure is in addition to it.
  • The AG is actively enforcing Act 75 (it opened an investigation of a social media AI video in June 2026); a court blocked that one case on First Amendment grounds, but do not treat the law as dead.

AI voice calls and texts

Can a campaign use AI voice calls in Vermont?

Vermont has no statute that specifically regulates AI or artificial voices in political calls, and no ban on political robocalls. Two rules reach campaign calls: robotic phone calls are listed as "electioneering communications" and "mass media activities" in 17 V.S.A. 2901, so they must carry sponsor identification under 17 V.S.A. 2972, and Act 75 covers audio, so an AI voice clone of a candidate (or AI audio giving voters materially false election information) used within 90 days of an election needs the spoken disclosure at the start and end of the call. Federal TCPA rules on artificial voices still apply.

Existing campaign finance law: Identification in electioneering communications (covers robotic phone calls)

In effect
Citation
17 V.S.A. 2901 (definitions) and 17 V.S.A. 2972 (identification)
When it applies
anytime
Media covered
Electioneering communications, defined to include robotic phone calls and mass electronic or digital communications that refer to a clearly identified candidate and promote, support, attack, or oppose that candidate. A "telephone bank" is defined as more than 500 identical or substantially similar calls in a 30-day period.
Who it covers
Any person, candidate, political committee, or political party that pays for an electioneering communication
Disclaimer format
Must contain the name and mailing address of the payer; an audio electioneering communication paid for by a candidate does not need to include the candidate's address.
Exemptions
  • lapel stickers and buttons
  • individuals acting alone who spend $150 or less in a two-year general election cycle
Notes
Not AI-specific. Robotic phone calls are also "mass media activities" in 17 V.S.A. 2901, which triggers separate reporting rules close to an election.

Read the official text

S.23 (Act 75 of 2026): Audio deepfake disclosure (applies to prerecorded calls and voicemail)

In effect

Required disclaimer, word for word

“This media has been manipulated or generated by digital technology and depicts speech or conduct that did not occur.”
Citation
17 V.S.A. 2032(a)(2)
Enacted
March 5, 2026
Effective
March 5, 2026
When it applies
within 90 days before an election in Vermont
Media covered
Audio recordings of an individual's speech created or manipulated with digital technology including AI, when deceptive and fraudulent
Who it covers
Any person distributing the audio
Rule
ban unless disclosed
Disclaimer format
Spoken at the beginning and end of the audio, and every two minutes or less for audio longer than two minutes.
Exemptions
  • telecommunications service providers for content provided by another person
  • satire or parody
Notes
Act 75 regulates media depicting a real individual. A generic AI voice that does not imitate a real person and carries no false election information is not addressed by Act 75 (our reading of the definitions).

Read the official text

Federal floor everywhere: AI-generated voices are “artificial” voices under the TCPA (FCC, February 2024), so they need prior express consent to cell phones and must identify the caller at the start of the call. AI robocall laws in every state

AI and election administration laws

Other Vermont AI laws that touch elections

S.23 (Act 75 of 2026), section 1: AI media that gives voters materially false election information

17 V.S.A. 2031(1)(B) · effective March 5, 2026

Act 75 is not limited to candidate deepfakes: it also covers realistic synthetic media of any individual that attempts to unduly influence an election or public question by giving voters materially false information, which reaches AI-faked messages from election officials or fake voting-procedure announcements inside the 90-day window.

Official text

Federal layer

Federal rules that apply in Vermont

At the federal level, the binding rules that hit AI in campaigns today are the TCPA and FCC robocall rules, not any AI-specific election law. Since February 2024 the FCC treats AI-generated voices as "artificial" voices, so AI voice calls to cell phones need prior express consent (political calls included), AI or prerecorded political calls to landlines are limited to three per 30 days without consent, and every such message must identify the caller, give a callback number and offer an opt-out. The FEC says its existing fraudulent misrepresentation ban covers AI, but it has had no quorum since April 2025 and cannot enforce. Proposed FCC rules on AI disclosure in broadcast political ads and in AI calls were never adopted, and no federal AI political ad disclaimer bill has passed. Federal efforts to preempt state AI laws (a 2025 moratorium the Senate struck 99 to 1, and Executive Order 14365) have not displaced any state election or robocall law as of October 2026.

FCC, FEC and the federal preemption fight

How Vermont compares

25 other jurisdictions share Vermont's status (disclosure):

Alabama · Arizona · Colorado · Delaware · Florida · Idaho · Indiana · Kentucky · Louisiana · Maine · Michigan · Mississippi · Montana · Nevada · New Hampshire · New Mexico · New York · North Dakota · Oregon · Rhode Island · South Dakota · Tennessee · Utah · Washington · Wisconsin

See Vermont on the mapPolitical deepfake laws, all statesDisclaimer generatorVermont 2026 elections

HyperPhonebank

Voter contact that clears Vermont's AI rules before the first dial.

Live professional callers and AI voice programs for voter ID, persuasion, ballot chase and GOTV, scripted to the disclosure and consent rules on this page, with every call graded and a real time dashboard.

Frequently asked questions

Does Vermont regulate AI in political ads?

Yes. Act 75 of 2026 (S.23, 17 V.S.A. 2031 to 2033 and 2041 to 2042) bars publishing or distributing synthetic media that you know is "deceptive and fraudulent" within 90 days before a Vermont election unless it carries a set disclosure. Covered media is any image, audio, or video of a person's appearance, speech, or conduct created or intentionally manipulated with digital technology, including AI, that realistically depicts a candidate in a way that injures the candidate's reputation, or depicts anyone in a way that tries to sway an election (including a public question) with materially false information. Knowing and intentional violations carry fines from $1,000 up to $15,000, the depicted candidate can sue for an injunction, and the Attorney General or a State's Attorney can investigate and sue. A federal court blocked one AG investigation of an AI satire video in Sept. 2026 (Poitras v. Clark) but left the law in force for everyone else. Key law (S.23 (Act 75 of 2026), 17 V.S.A. 2031 to 2033 (subchapter 4), 2041 to 2042 (subchapter 5)).

What AI disclaimer does Vermont require on political ads?

S.23 (Act 75 of 2026), 17 V.S.A. 2031 to 2033 (subchapter 4), 2041 to 2042 (subchapter 5) requires: "This media has been manipulated or generated by digital technology and depicts speech or conduct that did not occur." Format: Images and video: text in a size easily readable by the average viewer and accessible to people with disabilities to the greatest extent possible; for video, displayed for the full duration. Audio only: read in a clearly spoken manner, at a pitch and pace easily heard by the average listener, at the beginning and end of the audio, and if longer than two minutes, repeated at intervals of no more than two minutes.

What is the penalty for a deceptive political deepfake in Vermont?

Civil: Fines for knowing and intentional violations (17 V.S.A. 2033(a)): up to $1,000; up to $5,000 with intent to cause violence or bodily harm; up to $10,000 for a repeat violation within five years; up to $15,000 for both. Obstructing an AG or State's Attorney civil investigation: fine up to $5,000. Private lawsuits: A candidate whose appearance, speech, conduct, or environment is misrepresented may seek injunctive or other equitable relief to stop distribution. No damages remedy is created.

When does the Vermont political deepfake law apply?

within 90 days before an election in Vermont. It covers Images, audio recordings, and video recordings of an individual's appearance, speech, or conduct created or intentionally manipulated with digital technology, including artificial intelligence. Only "deceptive and fraudulent" synthetic media is regulated: media that appears to a reasonable person to realistically depict (A) a political candidate in a way that injures a candidate's reputation, or (B) an individual in a way that attempts to unduly influence an election, including a public question, by giving voters materially false information.

Can a campaign use an AI voice for robocalls in Vermont?

Vermont has no statute that specifically regulates AI or artificial voices in political calls, and no ban on political robocalls. Two rules reach campaign calls: robotic phone calls are listed as "electioneering communications" and "mass media activities" in 17 V.S.A. 2901, so they must carry sponsor identification under 17 V.S.A. 2972, and Act 75 covers audio, so an AI voice clone of a candidate (or AI audio giving voters materially false election information) used within 90 days of an election needs the spoken disclosure at the start and end of the call. Federal TCPA rules on artificial voices still apply. Under federal law, the FCC ruled in February 2024 that AI-generated voices are "artificial" voices under the Telephone Consumer Protection Act, so prerecorded or AI voice calls to cell phones need prior express consent, and every such call must identify the caller at the start.

Does Vermont have other AI election laws beyond political ads?

Yes. S.23 (Act 75 of 2026), section 1, AI media that gives voters materially false election information: Act 75 is not limited to candidate deepfakes: it also covers realistic synthetic media of any individual that attempts to unduly influence an election or public question by giving voters materially false information, which reaches AI-faked messages from election officials or fake voting-procedure announcements inside the 90-day window.

How many political AI laws does Vermont have?

This tracker counts 4 Vermont AI laws touching politics and elections as of October 5, 2026: 1 on elections, 2 on AI or artificial voice calls, 1 on voting and election officials.

Are there pending AI election bills in Vermont?

H.846 (2026) (Platform requirements for synthetic media (per Public Citizen tracker)): Failed, May 2026 (session adjourned without passage).

Sources

Verified October 5, 2026 against official legislature and statute pages. This page summarizes Vermont law for campaigns and is not legal advice.