The Senate Judiciary Committee has scheduled a June 18 markup of S.4591, the NO FAKES Act of 2026, sending the fourth iteration of a federal deepfake bill toward the floor with platform liability capped at $750,000 per work and a $5,000-per-work floor against individual infringers.

Introduced May 20 by Sens. Marsha Blackburn, Chris Coons, Thom Tillis and Amy Klobuchar, with a House companion (H.R.8915) from Reps. María Elvira Salazar and Madeleine Dean, the bill would establish the first federal intellectual property right over a person’s voice and likeness, running for life plus up to 70 years, renewable by heirs in five-year increments. Platforms get a safe harbor only if they register an agent with the Copyright Office and act on takedowns.

The endorser list reads like a who’s-who of the legacy IP stack: Disney, Universal, Warner, Sony Music, WME, CAA, the RIAA, the AFL-CIO and the Recording Academy. The White House’s National AI Framework recommended the bill by name. That’s an unusually unified coalition for a tech-regulation fight, and it explains why the 2023 discussion draft, which was treated as the strongest structural shot at federal deepfake legislation at the time, now looks modest by comparison.

Opponents are sharper. The Electronic Frontier Foundation argues the new property right sweeps in parody, news and criticism, and that artists will sign it away in boilerplate label and studio contracts. The Computer & Communications Industry Association pegs first-year compliance for a startup at roughly $1.64 million: $1.14 million annually plus $500,000 in fingerprinting.

The markup shares its date with a class-action conference in Doe 1 v. X.AI Corp. in the Northern District of California. Statute and tort are converging on the same question from opposite ends, and Washington is choosing which side to underwrite.

Sources

Sources