AI Deepfake Porn Laws Expand Across the U.S., But States Differ on What Is Illegal

AI-generated nude and sexual deepfakes can now lead to criminal charges under federal law and in a growing number of states, as lawmakers move to punish realistic sexual images created or shared without a person's consent.

The rules remain a patchwork. Federal law largely targets publication, while states including Florida, Texas and Michigan have criminalized some forms of creation or production before an image is ever posted. Other laws focus on distribution, threats, intent to harm, or images involving minors.

Congress entered the field in May 2025 with the TAKE IT DOWN Act, the first federal criminal law directly addressing the nonconsensual publication of intimate images, including realistic AI-generated sexual deepfakes.

The law defines a digital forgery as an intimate depiction of an identifiable person created or altered using software, machine learning, artificial intelligence, or other technology that a reasonable person would view as indistinguishable from an authentic image.

For adults, federal prosecutors generally must show that a qualifying digital forgery was published without consent and that other requirements under the statute were met, including those involving harm or an intent to cause harm. A conviction can carry up to two years in prison.

Offenses involving minors can carry up to three years. The law also covers certain threats to publish sexual deepfakes when they are intended to intimidate, coerce, extort or cause mental distress.

Federal prosecutors have already begun using the law.

In May 2026, prosecutors in Brooklyn charged two men accused of publishing thousands of AI-generated pornographic images and videos depicting celebrities, elected officials and private individuals without their consent. The Justice Department followed in June by seizing CFAKE.com and SOCFAKE.com, websites authorities said hosted large numbers of digitally forged nude and sexual images of women.

The federal law also places new responsibilities on online platforms. Since May 19, 2026, covered platforms have been required to provide a process for victims to request removal of nonconsensual intimate images, including qualifying AI-generated material.

Platforms generally must remove reported material and known identical copies within 48 hours after receiving a valid request. The Federal Trade Commission is responsible for enforcing those requirements.

The FTC moved quickly after its enforcement authority took effect. On May 20, the agency sent warning letters to 12 websites offering so-called nudify tools that can digitally remove clothing from photographs and create sexualized images. The agency warned that companies failing to comply with federal removal requirements could face enforcement proceedings and civil penalties.

State laws can reach further than the federal statute, particularly when someone creates a fake nude image but never publishes it.

Florida provides one of the clearest examples. State law makes it a third-degree felony to willfully generate an altered sexual depiction of an identifiable person without that person's consent. Soliciting such an image can also be a felony, as can maliciously promoting it or possessing it with the intent to promote it.

That means potential criminal exposure in Florida can begin when the image is created, rather than when it reaches social media or another public forum.

Texas law also covers AI-generated photographs and other sexual deepfake media. A person can commit an offense by knowingly producing or electronically distributing qualifying sexual deepfake material depicting someone without that person's effective consent. The law separately addresses threats intended to coerce, extort, harass or intimidate.

New York takes a different approach. Its unlawful dissemination law generally focuses on intentionally publishing or distributing an intimate image, including one created or altered digitally, with an intent to cause harm and knowledge that the depicted person did not consent. The offense is a Class A misdemeanor.

Michigan's Protection from Intimate Deep Fakes Act, which took effect in 2025, reaches nonconsensual creation as well as dissemination and provides civil remedies for people depicted in sexual deepfakes.

Washington criminalizes knowingly disclosing a fabricated intimate image when the person knows or should know the individual depicted did not consent and that the disclosure would cause harm. State law expressly includes images created or altered using artificial intelligence. A first offense is generally a gross misdemeanor, while repeat violations can become Class C felonies.

California has developed criminal and civil restrictions of its own. Recent legislation expanded protections involving fabricated sexually explicit material and allows lawsuits over the creation or disclosure of certain sexual deepfakes. Potential liability can also extend to people or businesses that knowingly facilitate or recklessly aid prohibited conduct.

The differences among the laws matter because creating, possessing and distributing an AI sexual image are legally separate acts.

Someone who generates a fake nude image and keeps it on a device could fall outside a law focused on publication while violating another state's prohibition on creation. Sending the image to another person can qualify as distribution under some statutes, while posting it publicly can bring additional criminal or civil consequences.

Consent can also apply differently at each stage. Permission to use an ordinary photograph does not generally mean permission to turn it into sexual material. Consent to create an intimate image also does not automatically authorize someone to publish it later. The TAKE IT DOWN Act recognizes that distinction.

Images involving children carry additional legal risks.

The National Center for Missing & Exploited Children reported more than 400,000 CyberTipline reports with a generative AI connection in 2025. More than 182,000 reports fell into categories involving the generation or possession of AI child sexual abuse material, attempts to generate it or AI use involving existing CSAM files. NCMEC has categorized more than 158,000 images and videos as AI-generated CSAM since it began tracking the technology in 2023.

Some incidents involve offenders altering photographs of actual children. Others involve students using nudify applications to create fake nude images from ordinary school or social media photographs.

The technology has also appeared in sextortion schemes in which someone threatens to distribute a fabricated sexual image unless a child provides money, additional images or sexual material.

States have responded by broadening child exploitation laws. Washington, for example, defines a fabricated depiction of an identifiable minor to include realistic images created or altered through AI that make a recognizable child appear to engage in sexually explicit conduct that did not occur.

Whether an image depicts an actual child can be legally significant. Governments have broad authority to prohibit sexual exploitation material involving real children. Entirely synthetic images depicting fictional people can raise different constitutional questions, particularly when no actual child was used to create the material.

Courts are only beginning to address how those established rules apply to increasingly realistic AI-generated images.

The technology also changes what police and prosecutors may need to establish during an investigation. Authorities can seek to determine who generated an image, which program was used, where the original photograph came from, whether the depicted person is identifiable, whether consent existed and who ultimately sent or posted the file.

Prompts, account records, messages, metadata, cloud storage and payment information can become evidence. In investigations involving suspected AI-generated child sexual abuse material, authorities may face an additional question: whether the apparent child depicted in the image exists at all.

State legislation continues to expand. The National Conference of State Legislatures previously reported that at least 19 states had enacted measures addressing sexually explicit deepfakes, while states continued considering additional restrictions in later legislative sessions. At least half of states enacted some form of deepfake legislation in 2025, although those measures also covered elections and other uses of synthetic media.

The growing number of laws means an image being fake does not by itself shield the person who created or distributed it from criminal or civil liability. The consequences can depend on whether a real person is identifiable, whether consent was given, what was done with the image, and which federal or state law applies.

As AI-generated photographs and videos become more difficult to distinguish from genuine recordings, courts will continue determining how privacy, free speech, and child exploitation laws apply to synthetic sexual material. Federal and state laws currently impose different rules on creation, publication, threats, images involving minors, and platform removal duties.