State Deepfake Statutes (MN, WA, TX, CA)

US state-level laws regulating political and election deepfakes fall into two dominant regulatory modes: outright prohibition within a pre-election window (Minnesota, Texas) and required disclosure of synthetic media depicting candidates (Washington, California's surviving 2024 law). As of early 2026, 26 states have enacted some form of political or election deepfake law (Brennan Center, 2025 tracker), and the body of law is the subject of an active First Amendment dispute centered on the Kohls litigation in California and Minnesota. Texas, Minnesota, Washington, and California are the four most-cited statutes and have generated the most litigation; they are treated in detail below.

A separate state landscape addresses sexual nonconsensual deepfake imagery rather than political content. As of September 2025, 48 of 50 states had enacted laws addressing AI-generated sexual nonconsensual abuse imagery, with Michigan becoming the 48th on August 26, 2025 when Governor Gretchen Whitmer signed HB 4047 / 4048 (a misdemeanor for creation; a felony if posted, monetized, or created with intent to harass). Only New Mexico and Missouri had neither pending nor enacted statutes (Source: 404media.co). The federal law on sexual NCII deepfakes is TAKE IT DOWN Act — Source Summary; the state political-deepfake statutes covered here address a distinct domain, and the state landscape now provides near-universal coverage of sexual NCII deepfakes alongside the federal regime.

Regulatory modes

Two approaches predominate across the states. Under the prohibition model, a criminal or civil penalty attaches to publishing a deepfake within a defined pre-election window, with intent elements; Minnesota and Texas take this approach. Under the disclosure model, a visible disclaimer is required when synthetic media depicts candidates, and depicted candidates may seek injunctive relief; Washington and the 2024-era California laws take this approach. Of the 26 state laws, 24 adopt the disclosure model; only Minnesota, Texas, and partially Colorado prohibit.

Statutes by state

Texas — HB 1758 (2019)

HB 1758 amended Texas election law to add the criminal offense of fabricating a deceptive video of a candidate. It was passed by the Texas Legislature and signed by Governor Abbott, taking effect September 1, 2019, making it the first deepfake statute in the US. The law makes it a criminal offense to create and publish a deepfake video within 30 days of an election with intent to injure a candidate or influence the election result. The offense is a Class A misdemeanor, enforced by the Texas Attorney General and local prosecutors. In 2023 and 2024, Texas SB 751 and related bills extended prohibitions to audio deepfakes and non-consensual intimate imagery, the latter since federalized by TAKE IT DOWN Act — Source Summary.

Minnesota — Minn. Stat. § 609.771 (2023, expanded 2024)

Minnesota's law regulating election-related "deep fakes" was passed by the Minnesota Legislature and signed by Governor Walz in 2023, then expanded in 2024. It makes it a crime to disseminate a deepfake within 90 days of an election where the depicted person does not consent, the publisher knows the material is a deepfake, and publication is made with intent to injure a candidate or influence the result. The penalty is a gross misdemeanor, rising to a felony on a repeat offense or where the deepfake depicts sexual conduct. The 2024 amendments added non-election sexual-deepfake offenses, largely preempted by TAKE IT DOWN Act — Source Summary but retained for state prosecution. The statute is the subject of an active First Amendment challenge in federal court (Kohls v. Ellison, D. Minn., 2024), with an 8th Circuit appeal pending as of early 2026.

Washington — RCW 29A.32.031 (2023)

Washington's statute on synthetic media in election-related communications was passed by the Washington Legislature and signed by Governor Inslee in 2023. It uses a disclosure mechanism, requiring clear disclosure when synthetic media (termed "fabricated media" under the statute) is used in election-related communications covering elections for state, local, and federal offices. Candidates depicted in non-compliant synthetic media may seek injunctive or equitable relief and damages. Enforcement runs through private action and the state Attorney General.

California — AB 2839, AB 2655, AB 2355 (2024)

In 2024 the California Legislature passed, and Governor Newsom signed, a package of laws addressing "materially deceptive media" in elections. AB 2839 prohibited knowingly distributing materially deceptive election-related content within 120 days of an election. AB 2655 required large online platforms to label or remove materially deceptive content about candidates during election windows. AB 2355 required disclosure for AI-generated political advertising. A federal judge in the Eastern District of California struck down AB 2839 on First Amendment grounds in Kohls v. Bonta (2024), treating it as a prior restraint and content-based restriction on political speech and rejecting the state's "knowingly false" framing as insufficient. AB 2655 is subject to ongoing challenge, and AB 2355, the AI-ad disclosure law, remains in effect and interacts with California AI Transparency Act (SB 942). As of 2026, California's prohibition path has been substantially constrained while the AI-ad disclosure law survives.

Comparison of the four statutes

First Amendment litigation

The running First Amendment confrontation is the most-litigated feature of this body of law. In Kohls v. Bonta (E.D. Cal. 2024), the court struck down California's AB 2839 as a prior restraint and content-based restriction on political speech, rejecting the state's "knowingly false" framing as insufficient. In Kohls v. Ellison (D. Minn. 2024, appeal 8th Cir. 2025), the same challengers contest Minnesota's prohibition statute on the same grounds. The Tennessee ELVIS Act (2024), not covered in detail here, protects voice and likeness under a different doctrinal frame (right of publicity) but is part of the same synthetic-media regulatory push.

Legal commentators including Eugene Volokh, Rebecca Tushnet, and the Electronic Frontier Foundation argue that disclosure laws are more likely to survive First Amendment review than outright prohibitions, a view associated with the post-2024 drift among state legislatures toward the Washington and California AB 2355 model over the Minnesota and Texas model.

Other state models

Beyond the four detailed statutes, Colorado uses the term "deepfake" and adopts an election prohibition. Michigan, New York, Florida, Arizona, Georgia, and others have enacted disclosure-based election statutes. The Tennessee ELVIS Act protects voice and likeness under a right-of-publicity frame. In total, 26 states had some form of political-deepfake law as of early 2026 (Brennan Center tracker).

A survey published August 7, 2026 counted 29 states with election-related AI deepfake laws in effect as of June 23, 2026, and two states whose similar laws had been struck down by courts. The survey groups the statutes by temporal scope: Minnesota and Texas prohibit political deepfakes for a set number of days before an election, while Maryland's ban runs year-round (Source: axios.com). The count is measured on a different date and by a different tracker than the Brennan Center figure above, so the two are not directly comparable; both are recorded rather than reconciled. The unevenness the survey documents — coverage that depends on the voter's state and, within a state, on proximity to election day — is the pattern the federal preemption debate at AI Federalism turns on.

Relation to other US synthetic-media law

The state election-deepfake laws and the federal TAKE IT DOWN Act divide the synthetic-media field by harm domain: federal law governs sexual NCII, and states govern political deepfakes. Neither domain is addressed by the other. Combined, they produce a US synthetic-media legal landscape that is comprehensive in coverage but fragmented across statutes and levels of government.

The August 2026 survey records the Take It Down Act, which took effect in May 2026, as Congress's first measure addressing non-consensual intimate imagery (Source: axios.com).

California AI Transparency Act (SB 942)'s machine-readable disclosure could serve as an enforcement primitive for political-deepfake laws, but the state election statutes largely predate SB 942 and do not leverage it.

Relationships

• related: TAKE IT DOWN Act — Source Summary — federal analog for sexual NCII
• related: California AI Transparency Act (SB 942) — watermarking / provenance instrument that could interoperate
• related: California SB 243 — Companion Chatbots — adjacent state synthetic-media / minors statute
• related: AI and the First Amendment — constitutional framework
• related: US AI Regulatory Approaches Compared — adds election-integrity dimension
• related: Techno-Federalism: How Regulatory Fragmentation Shapes the U.S.-China AI Race — state action filling a federal gap

Sources

• TechPolicy.Press, Regulating Election Deepfakes: A Comparison of State Laws
• R Street, Update on 2025 State Legislation to Regulate Election Deepfakes
• Multistate.AI, Can State Laws Actually Stop Political Deepfakes?
• Brennan Center, Regulating AI Deepfakes and Synthetic Media in the Political Arena (2025 tracker)
• Minnesota Secretary of State, MN Deep Fake Overview
• Mitchell Hamline Law Review, Election Integrity and the First Amendment
• First Amendment Encyclopedia (MTSU), Political Deepfakes and Elections
• Conference Board (CED), Federal Judge Strikes Down California Deepfake Law
• Fordham Law Review Vol. 93, Issue 1, Article 7 (2024)
• Global Law Today, Deepfake Disclosure Laws: How States Are Tackling Synthetic Media (2025)