For the first time, Congress is close to telling every American something the law has never quite said: your face and your voice belong to you, even when a machine is the one copying them.
On June 18, the Senate Judiciary Committee unanimously advanced the NO FAKES Act of 2026 — the Nurture Originals, Foster Art, and Keep Entertainment Safe Act — by voice vote, sending the bill toward the full Senate floor. Six weeks later, as lawmakers head into the summer stretch, S. 4591 stands as the most serious federal attempt yet to give individuals control over how artificial intelligence recreates them, and the clearest test to date of whether Washington can regulate synthetic media without trampling free speech.
A new federal right
At the center of the bill is what its authors call a "digital replication right" — a new federal intellectual property protection covering a person's voice and visual likeness. Under the legislation, producing or distributing a computer-generated replica of someone without consent becomes a violation an individual can sue over. Crucially, the right is not limited to celebrities. It applies to the songwriter, the schoolteacher, and the eighth-grader alike.
"Nobody — whether they're Tom Hanks or an 8th grader just trying to be a kid — should worry about someone stealing their voice and likeness," said Senator Chris Coons (D-Del.), a lead sponsor, framing the bill as protection for ordinary people rather than famous ones.
The right can be licensed, transferred, and inherited. It survives death and passes to heirs, executors, and estates, though the post-mortem protection expires no later than 70 years after a person dies — a structure that borrows from copyright and puts real economic value in the hands of estates and, potentially, the entertainment companies that contract with them.
The bill is deliberately bipartisan. Coons is joined by Republican Senator Marsha Blackburn (Tenn.), along with Senators Amy Klobuchar (D-Minn.) and Thom Tillis (R-N.C.), and House companions from Representatives Maria Salazar (R-Fla.) and Madeleine Dean (D-Pa.). It has been reworked repeatedly since a 2023 discussion draft, reintroduced in 2025, and revised again in May 2026 into the version now moving.
"AI should empower innovation — not give scammers and online predators a free pass to exploit someone's voice and visual likeness without permission," Blackburn said, tying the measure to fraud and child-safety concerns as much as to Hollywood.
Takedowns, staydowns, and $750,000
For platforms, the mechanics matter as much as the principle. NO FAKES builds a notice-and-takedown pathway modeled loosely on the Digital Millennium Copyright Act: an individual flags an unauthorized replica, and a service that removes it promptly can claim a safe harbor from liability. The 2026 version goes further, adding a "notice-and-staydown" obligation — once notified, a platform must make reasonable efforts to keep identical infringing content from reappearing.
That is where the numbers get large. Platforms that fail to comply face liability of up to $750,000 per work. For a large user-generated-content service, staydown implies building content-fingerprinting infrastructure for synthetic-identity material that does not yet exist at scale.
The industry coalition behind the bill is unusually broad. SAG-AFTRA — whose members drove the AI protections at the heart of the 2023 actors' strike — has rallied support, with more than 16,000 members signing an open letter to Congress. The RIAA, the Recording Academy, the Motion Picture Association, Universal Music Group, and, notably, AI developers including OpenAI, Google, Amazon, Adobe, and IBM have all backed versions of the measure. "RIAA proudly endorses the NO FAKES Act of 2026, a widely supported consensus bill," said Chairman and CEO Mitch Glazier, calling it the product of a bipartisan, bicameral process.
The free-speech fight
That consensus is not universal. A coalition of civil-liberties groups — the Electronic Frontier Foundation, the ACLU, the Center for Democracy & Technology, Public Knowledge, Fight for the Future, and others — wrote to the Judiciary Committee urging it not to advance the bill as written. Their core objection: the takedown regime creates a "heckler's veto" over lawful speech, giving platforms every financial incentive to remove content first and adjudicate later.
The EFF warned the bill "could silence satire, commentary, and news," arguing it offers platforms no protection for the genuinely hard judgment call of whether a given clip is parody, criticism, or reporting. Guess wrong, and the penalty is $750,000. A law meant to shield ordinary people, the groups cautioned, could become "another layer of internet censorship" — and a tool for the powerful to bury unflattering speech. Even senators who voted to advance the bill flagged concern over whether it adequately protects "legitimate First Amendment speech," and agreed to keep negotiating before a floor vote.
The stakes are sharpened by the current legal vacuum. There is no federal likeness right; protection today is a patchwork of state right-of-publicity laws — strong in California and Tennessee, thin or nonexistent elsewhere — plus narrower state deepfake statutes covering election ads and nonconsensual intimate images. NO FAKES would supersede much of that with a single national standard, which is precisely why both its backers and its critics see it as a landmark.
What to watch
The next moves are procedural but decisive. Watch whether sponsors add explicit First Amendment carve-outs — for news, parody, and documentary use — to peel off civil-liberties opposition before a floor vote. Watch the staydown language, the most likely target of platform lobbying. And watch the calendar: with an election year approaching, a consensus bill this far along could pass quickly, or stall in the Senate's crowded queue and reset in the next Congress.
"Nobody — whether they're Tom Hanks or an 8th grader just trying to be a kid — should worry about someone stealing their voice and likeness."- Chris Coons, U.S. Senator (D-Del.), lead sponsor