AI laws on voting, voters and election officials

• AB 2655 (2024): Defending Democracy from Deepfake Deception Act of 2024 (platform duties), Cal. Elec. Code 20510 to 20520, effective January 1, 2025 (struck down or blocked in court). Required large online platforms (1 million or more California users) to remove reported materially deceptive election deepfakes within 72 hours during set pre-election windows and to label others with "This [image/audio/video] has been manipulated and is not authentic."
• AB 502 (2026): Deceptive portrayals of election officials and voting equipment, Cal. Elec. Code 20012(b)(1)(B) to (D), (c)(3), effective September 19, 2026. Bars knowingly distributing, with malice, digitally fabricated content showing an elections official or elected official misrepresenting election conduct, or showing ballots, voting machines or voting sites in a materially false way, from 30 days before to 38 days after an election. Elections officials may sue for injunctions and damages.
• SB 1001 (2018): Bolstering Online Transparency (BOT) Act, Cal. Bus. and Prof. Code 17940 to 17943, effective July 1, 2019. Unlawful to use an automated online account (bot) to communicate with a person in California online with intent to mislead about its artificial identity in order to influence a vote in an election, unless the bot clearly and conspicuously discloses that it is a bot.

• Senate Substitute for HB 2056 (2025): False representation of an election official, K.S.A. 25-2438 (crime of false representation of an election official). It is a crime to falsely represent yourself as an election official with intent to cause a person to believe you are one, including by using the official seal or insignia of the Secretary of State or a county election office in any communication with voters. Not AI-specific, but it reaches AI-generated calls, texts, or videos that impersonate election officials.

• SB 4 (2025), 2025 Ky. Acts ch. 66: State government AI governance plus election synthetic media remedy, KRS 42.720 to 42.742 (state AI governance) and KRS 117.322, effective March 24, 2025. The same act that created the election deepfake remedy also created an Artificial Intelligence Governance Committee and requires the Commonwealth Office of Technology to set standards and publicly disclose certain state uses of AI. Its emergency clause states a purpose of protecting candidates and election officers from fraudulent misrepresentations.

• HB 459 (2026), Act 806: Campaign finance reporting for electioneering communications, La. R.S. 18:1483(33), 18:1501.1(A)(1), effective August 1, 2026. The same act that created the AI disclosure rule broadened "electioneering communication" to cover printed, digital, and broadcast express advocacy and its functional equivalent, and requires persons other than candidates and committees who spend over $1,000 on electioneering communications to file reports. AI-generated ads by outside groups are therefore reportable spending.

• HB 333 (2024), Chapter 126: Election Disinformation and Improper Influence Related to Voting, Md. Code, Election Law 2-110, effective June 1, 2024. Requires the State Board of Elections to run a public portal for reporting election misinformation and disinformation, review submissions, issue corrections, and refer cases to the State Prosecutor.
• SB 141 (2026), Chapter 444: State Administrator powers against election misinformation, including deepfakes, Md. Code, Election Law 2-110(c), effective June 1, 2026. On a credible report of election misinformation or disinformation, including a deepfake, the State Administrator must publicly correct it and may seek an injunction to remove it from online platforms (but not against platforms for third-party content) and subpoena records about who spread it and to whom. The State Board may sue the responsible person, campaign, or PAC for its correction costs.

• HB 1432 (2024): Fraudulent use of deepfakes (general law that reaches candidates), N.H. RSA 638:26-a (crime); RSA 507:8-j (civil action), effective January 1, 2025. A general deepfake crime, not an election ad rule. Knowingly creating, distributing, or presenting a deepfake of an identifiable person to embarrass, harass, entrap, defame, extort, or cause financial or reputational harm is a class B felony at any time of year, with no disclaimer safe harbor, so a deepfake made to damage a candidate's reputation is covered. Satire, parody, bona fide news reports, and paid carriers that keep the creator's disclaimer on an election communication are exempt.
• Criminal case: State v. Kramer (2024 to 2025): Biden AI robocall prosecution, N.H. RSA 659:40 (voter suppression) and RSA 666:7-a (impersonation of a candidate), as charged. The Attorney General charged Steve Kramer under existing voter suppression and candidate impersonation laws for the AI Biden robocalls. A jury acquitted him on all 22 counts tried on June 13, 2025, after the defense argued the unsanctioned Democratic primary was not covered. Existing voter suppression law still applies to AI calls that lie about how or when to vote.
• Federal civil case: League of Women Voters of New Hampshire v. Kramer (D.N.H.): Civil suit over AI voter intimidation robocalls, Voting Rights Act section 11(b) and state law claims. Voters and the League sued Kramer, Lingo Telecom and Life Corporation. Consent judgments resolved the claims against the corporate defendants in 2025, and on November 20, 2025 the court entered a default judgment against Kramer awarding damages and permanently barring him from deceptive or coercive robocall schemes against voters.
• FCC enforcement (2024): Federal penalties for the New Hampshire AI robocall, FCC forfeiture order against Steve Kramer; FCC consent decree with Lingo Telecom. Lingo Telecom agreed to pay $1 million and adopt know-your-customer and STIR/SHAKEN compliance measures (announced August 21, 2024). The FCC imposed a $6 million forfeiture on Kramer in September 2024 for spoofed caller ID; Kramer has said he will not pay.

• A 3540 (2024 to 2025), P.L. 2025, c. 40: Deceptive audio or visual media (deepfakes) used to further crimes, including improper influence in political matters, N.J.S.A. 2C:21-17.7 and 2C:21-17.8, effective April 2, 2025. General law, not an election ad rule. Creating, soliciting, disclosing, or using realistic deceptive audio or visual media (including voice clones) to attempt or further any crime is a third degree crime with fines up to $30,000; the listed predicates expressly include threats or improper influence in official and political matters (N.J.S.A. 2C:27-3) and harassment, so it reaches deepfakes or AI calls aimed at officials, candidates, or voters for those purposes. Knowingly or recklessly disclosing unlawfully created media is a fourth degree crime, and victims can sue for actual damages of at least $1,000 per knowing or reckless violation, punitive damages, and fees.

• Publication of false information in political advertisements (not AI-specific), N.D.C.C. 16.1-10-04. Makes it a class A misdemeanor to knowingly, or with reckless disregard for the truth, publish untrue, deceptive, or misleading material about a candidate's prior public record or a ballot issue in political ads, including by text message or telephone call. Not AI-specific, but a false AI ad, call, or text about an opponent's record can be prosecuted under it.

• Oregon Laws 2021 ch. 291: False publication intended to mislead electors (not AI-specific), ORS 260.537. Bars knowingly or recklessly circulating, including by telephone or electronic means, a false statement of material fact meant to mislead voters about the election date, ballot return deadline, registration deadline or methods, ballot drop locations, voter qualifications, or a voter's registration status, within 30 days before a primary or special election or 60 days before a general election. The Secretary of State can get an injunction and courts must impose a civil penalty up to $10,000. It is not AI-specific but squarely covers AI robocalls or deepfakes that lie about how to vote.

• SB 649 (2025), Act 35 of 2025: Digital forgery (general crime, not election-specific), 18 Pa.C.S. 4101.1, effective September 5, 2025. A general criminal law, not an election law. It is a first degree misdemeanor (third degree felony in a scheme to defraud, coerce, or steal) to create and distribute a "forged digital likeness" (a computer-generated image or voice of a real, identifiable person that materially misrepresents them, is likely to fool a reasonable person, and was made without consent) as genuine, with intent to defraud or injure. It reaches political deepfakes and voice-clone robocalls of candidates when made with that intent, at any time of year; taking reasonable action to tell viewers or listeners the content is not genuine is an affirmative defense, and constitutionally protected activity is excepted.

• 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.

• PDC rulemaking (WSR 24-12-019; WSR 26-01-209): Campaign finance disclosure of AI and synthetic media, WAC 390-16-037(3), 390-16-063(6), 390-18-035, 390-18-050(7)(g)(iv), effective June 24, 2024. Committees must list synthetic media use, the AI or software vendor, the ad seller, and audience targeting on expenditure and independent expenditure (C-6) reports, and commercial advertisers must keep records of any AI technology they provided to make synthetic media for political ads, open to public inspection. Synthetic media disclaimers are in addition to normal sponsor identification.
• SHB 1205 (2025), Laws of 2025 ch. 51: Knowing distribution of a forged digital likeness, RCW 9A.60.045(1)(c), effective July 27, 2025. General criminal law, not election-specific: knowingly distributing a forged digital likeness (a digitally created or altered image or voice recording of an identifiable person, indistinguishable from the real thing and likely to deceive) as genuine, with intent to defraud, harass, threaten, or intimidate or for another unlawful purpose, is criminal impersonation in the second degree, a gross misdemeanor. It can reach malicious deepfakes of candidates or election officials, but it expressly does not reach content of political, newsworthy, or public interest, including commentary, satire, and parody.

No law yet (39)

Alabama · Alaska · Arizona · Arkansas · Colorado · Connecticut · Delaware · District of Columbia · Florida · Georgia · Hawaii · Idaho · Illinois · Indiana · Iowa · Maine · Massachusetts · Michigan · Minnesota · Mississippi · Missouri · Montana · Nebraska · Nevada · New Mexico · New York · North Carolina · Ohio · Oklahoma · Rhode Island · South Carolina · South Dakota · Tennessee · Texas · Utah · Virginia · West Virginia · Wisconsin · Wyoming

HyperPhonebank

Voter contact that clears every state'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.