New Hampshire Deepfake Laws: AI Images, Voice Cloning & Penalties (2026)

New Hampshire stands out as one of the most aggressive states in the country on deepfake regulation. As of January 1, 2025, RSA 638:26-a criminalizes the creation, distribution, or presentation of harmful deepfakes as a Class B felony, and a companion provision, RSA 507:8-j, gives victims a private civil right of action for damages. Separate laws cover synthetic intimate images (RSA 644:9-a) and AI-generated election deepfakes (RSA 664:14-c). New Hampshire's lawmaking in this space was directly shaped by the January 2024 fake-Biden robocall incident that targeted the state's presidential primary.

Is It Illegal to Make a Deepfake of Someone in New Hampshire?

Yes, depending on the purpose, creating a deepfake of an identifiable person is a felony in New Hampshire. RSA 638:26-a targets deepfakes made to embarrass, harass, entrap, defame, extort, or cause financial or reputational harm. The law covers all three buckets that matter most: sexual and intimate imagery, voice cloning, and political manipulation.

What the law does not cover: purely fictional or satirical depictions that make no realistic claim to truth, news reporting that includes clear authenticity disclaimers, and deepfakes created for entertainment without harmful intent. The statute expressly exempts satire, parody, and traditional impersonation that does not rely on AI tools. Even in those exempt categories, a deepfake that crosses into defamation or harassment can still trigger liability under other statutes.

New Hampshire law does not require the deepfake to be sexually explicit to be criminal. RSA 638:26-a is a general-purpose harmful-deepfake law, broader than the NCII-only approach taken by many states.

Sexual and Intimate Deepfakes

New Hampshire addresses non-consensual intimate deepfakes through two overlapping statutes. RSA 644:9-a covers the nonconsensual dissemination of private sexual images and was amended to include synthetic images: a realistic but false depiction of a person's intimate parts, sexual acts, or sexual activity, disseminated without consent and with the intent to harass, intimidate, threaten, or coerce the depicted person. A violation is a Class B felony.

RSA 638:26-a provides a second, broader criminal hook. Because it covers harmful deepfakes in any medium, including intimate content designed to harass or extort, a prosecutor can charge under either statute or both depending on the facts. Victims also have a civil right of action under RSA 507:8-j, which allows them to sue for damages without waiting for criminal prosecution.

For minors, New Hampshire's child sexual abuse image chapter reaches synthetic content through its definitions. RSA 649-A:2 is the chapter's definitions section: it defines a covered visual representation to include digital, computer, and computer-generated images, including a depiction created, adapted, or modified to appear that an identifiable child is engaging in sexually explicit conduct. The offenses and penalties sit in the sections that follow, and those are what a charge is brought under: RSA 649-A:3 (possession, a class A felony), RSA 649-A:3-a (distribution, up to 20 years), RSA 649-A:3-b (manufacture, up to 30 years), and RSA 649-A:3-c (creation of child intimate visual representations, a class B felony, whose definition of intimate visual representation expressly includes a synthetic image). Because those definitions turn on what the image depicts rather than on whether a real child was photographed, AI-generated material can fall within them.

Election and Political Deepfakes

New Hampshire enacted RSA 664:14-c in 2024, the year the state experienced one of the most prominent deepfake incidents in American political history. In January 2024, two days before the New Hampshire presidential primary, robocalls carrying an AI-generated clone of President Biden's voice urged Democratic voters to stay home. The calls were traced to political consultant Steve Kramer. The FCC issued a $6 million fine, and the New Hampshire Attorney General charged Kramer with 13 felony counts of voter suppression and 13 misdemeanor counts of impersonating a candidate. A jury acquitted Kramer on all criminal charges on June 13, 2025; Kramer has said he does not intend to pay the FCC fine.

RSA 664:14-c prohibits any person, corporation, committee, or other entity from distributing, within 90 days of an election at which a candidate for elective office will appear on the ballot, a message created using AI or generative AI that the distributor knows or should have known is a deepfake of a candidate, election official, or party on the state or local ballot. Two limits are easy to misread. The statute is not confined to political advertising: it reaches any distributed AI-generated message, paid or not. But it is confined by subject: a deepfake of a person who is not a candidate, election official, or party on the ballot does not trigger this disclosure duty, although RSA 638:26-a may still reach it. The required disclosure must state: "This [Image/Video/Audio] has been manipulated or generated by artificial intelligence technology and depicts speech or conduct that did not occur." The law specifies minimum text size, display duration, and audio clarity requirements for the disclosure. Exemptions apply to bona fide news reporting, satire and parody, and impersonation that does not rely on AI. Enforcement is private: a candidate or election official depicted in a violating deepfake may seek an injunction and sue the sponsor for damages, and a prevailing party may be awarded attorney fees.

First Amendment risk is real in this space. A federal court struck down California's analogous election-deepfake law (AB 2839) in its entirety in August 2025 on free-speech grounds, permanently enjoining its enforcement. New Hampshire's disclosure-based approach (requiring a label rather than an outright ban) is generally considered more constitutionally durable than prohibition-only laws, but litigation in this area is ongoing nationally.

AI Voice Cloning and Digital Likeness

New Hampshire has no right-of-publicity statute. Most states protect voice and likeness through a dedicated civil law; New Hampshire relies on common law instead. That gap is meaningfully filled by RSA 638:26-a, whose deepfake definition expressly includes a digitally altered voice in video, audio, or any other media.

The national benchmark for voice-clone legislation is the Tennessee ELVIS Act (Tenn. Code Ann. 47-25-1101, eff. July 1, 2024), which was the first state law to extend right-of-publicity protections specifically to AI voice simulations. New Hampshire takes a different path: rather than creating a standalone civil right for voice and likeness, it criminalizes harmful voice cloning under the deepfake fraud statute (RSA 638:26-a) and provides the civil remedy through a companion provision, RSA 507:8-j. The practical effect for victims is similar: a civil damages claim is available, but the legal basis differs.

Commercial voice cloning without a harmful purpose (for example, a consenting celebrity licensing their voice to an AI company) is not addressed by NH statute and would fall under contract and common law principles. The federal NO FAKES Act, which would create a federal right of publicity for AI digital replicas, remains a proposal only as of June 2026 and has not passed either chamber of Congress.

Federal Law That Applies in New Hampshire

The TAKE IT DOWN Act (Public Law 119-12, signed May 19, 2025) is now the primary federal law on intimate deepfakes. It makes it a federal crime to knowingly publish nonconsensual intimate visual depictions of adults or minors, including AI-generated deepfakes. Penalties reach two years in prison (three for content involving minors). Platforms must remove flagged content within 48 hours of a victim's notice; the FTC enforces compliance.

The FCC's February 2024 ruling (FCC 24-17) declared that AI-generated voices in robocalls qualify as "artificial" voices under the Telephone Consumer Protection Act. That ruling, triggered directly by the New Hampshire Biden robocall, makes AI voice-clone calls to phones without prior express consent illegal nationwide, regardless of state law.

Federal law reaches AI-generated child sexual abuse material through 18 U.S.C. 2256(8)(B), which covers a visual depiction that is indistinguishable from that of a real minor engaging in sexually explicit conduct. One qualification matters. Under 18 U.S.C. 2252A(c), a defendant has an affirmative defense that the material was not produced using any actual minor or minors, a defense Congress added in response to the First Amendment holding in Ashcroft v. Free Speech Coalition. That defense is expressly unavailable only for material described in 18 U.S.C. 2256(8)(C), the category covering an image of an identifiable real minor altered to appear sexually explicit. New Hampshire's own chapter 649-A offenses apply alongside the federal scheme.

Two federal bills are still working through Congress as of June 2026. The DEFIANCE Act (S.1837) passed the Senate by unanimous consent in January 2026 but remains pending in the House; the NO FAKES Act (S.4591) has cleared the Senate Judiciary Committee and awaits a Senate floor vote (as of August 2026). The DEFIANCE Act would add a federal civil cause of action for sexual deepfakes; the NO FAKES Act would create a federal right of publicity for voice and likeness. Neither has been enacted and neither should be relied on as current law.

For general AI regulation, New Hampshire's broader approach to artificial intelligence governance is covered separately at New Hampshire AI Laws, which addresses the full scope of AI-related legislation beyond the deepfake-specific statutes discussed here.

What Victims Can Do

Victims of deepfakes in New Hampshire have several avenues for relief. For criminal matters, a report should be filed with local police or the New Hampshire Attorney General's office. RSA 638:26-a (harmful deepfakes) and RSA 644:9-a (synthetic intimate images) are both felony offenses that law enforcement can prosecute.

For civil remedies, RSA 507:8-j gives victims a private right of action to sue for damages. This is a direct claim for harm caused by a deepfake created to embarrass, harass, defame, extort, or otherwise damage the victim, covering both economic losses and reputational harm. Plaintiffs do not need to wait for a criminal conviction to pursue a civil case.

For online platforms, the TAKE IT DOWN Act's 48-hour removal obligation (effective May 19, 2026) creates a direct path for victims to demand takedowns of nonconsensual intimate deepfakes. Submit a notice to the platform identifying the content; the platform must act within 48 hours or face FTC enforcement action.

Additionally, the Deepfake and AI Voice Cloning Laws by State hub provides a national comparison of remedies, which can be useful if the creator of the deepfake is located in a different state.

Penalties at a Glance

Disclaimer: This page provides general legal information about New Hampshire deepfake laws and is not legal advice. Laws in this area are changing rapidly: New Hampshire enacted two major deepfake statutes in 2024, and federal law added a third layer in 2025. If you have been harmed by a deepfake or face a deepfake-related charge, consult a licensed New Hampshire attorney.

Sources

The following primary legal sources were used to prepare this article.

For broader New Hampshire AI regulation, see New Hampshire AI Laws. For recording consent law in the state, see New Hampshire Recording Laws. For data privacy, see New Hampshire Data Privacy Laws.

More New Hampshire Laws

• New Hampshire AI Meeting Recording Laws
• New Hampshire Data Privacy Laws
• New Hampshire Recording Laws
• New Hampshire Sexting Laws