Deepfake laws in the US, federal bills and state rules.

What are the deepfake laws in the US?

Deepfake laws in the US are a patchwork. There is no single federal ban yet, but a growing set of federal bills and state statutes now target unauthorized AI copies of a person's face and voice. As of 2026 the direction is clear: deepfakes that use someone's likeness without consent, especially for commercial or sexual purposes, are increasingly illegal.

This guide covers the federal picture, the state rules, and what all of it means if you protect talent. It is current as of 2026 and changes often, so treat it as a map, not the final word.

Is there a federal deepfake law?

There is no comprehensive federal deepfake law in force yet, but two pieces stand out. The NO FAKES Act is the big one: it would create a federal property right over a person's voice and visual likeness and hold people and platforms liable for unauthorized digital replicas. It advanced through the Senate Judiciary Committee with bipartisan support and is moving toward a full vote; if it passes, it becomes the first nationwide standard for AI likeness.

The TAKE IT DOWN Act is already law as of 2025. It targets non-consensual intimate imagery, including AI-generated versions, and requires platforms to remove it quickly once notified. Until the NO FAKES Act passes, the enforceable law for most commercial likeness misuse still comes from the states and from existing right-of-publicity claims.

Deepfake laws by state

States moved faster than Congress, and the rules vary widely, so where the person and the infringer are located matters. California is the strongest: AB 1836 plus Civil Code 3344 and 3344.1 mean digital replicas of deceased performers need estate consent, on top of a strong right of publicity for the living and the dead. Tennessee's ELVIS Act (2024) was the first state law to name AI voice cloning directly, protecting a person's voice as well as their likeness.

New York carries a postmortem right of publicity and a disclosure rule for synthetic performers in ads, and Washington's 2026 forged-digital-likeness law adds civil penalties for AI content made to deceive. Most other states cover deepfakes indirectly through right-of-publicity or privacy law. The practical takeaway: in much of the country, using someone's AI likeness commercially without consent is already actionable, even without a law that says the word deepfake.

The EU AI Act and other markets

Outside the US, the European Union's AI Act adds a transparency duty: from 2026, AI-generated or manipulated content that looks real generally has to be labeled as artificial. That is a disclosure rule rather than a likeness-ownership rule, but it pushes in the same direction, less hiding of synthetic media. Countries with strong image-rights traditions, like Spain with its right to one's own image, already give people a route to act against unauthorized likeness use.

What these laws mean for talent and estates

The law is catching up, but enforcement still lands on the rights holder. A statute only helps if you can point to a specific unauthorized use and prove it. That is the same story across every rule above: the claim exists, the evidence is the bottleneck.

For talent, agencies and estates, the useful posture is not to wait for the NO FAKES Act. It is to find the AI uses of a likeness now and document them in a form that holds up, so that whichever law applies, you are ready to act. That is what talicense does. For the underlying rights these laws enforce, see our guides on the right of publicity and personality rights. See how PROTECT works →