The NO FAKES Act has been introduced three times before. It died each time. The fourth introduction, S.4591, has done something its predecessors didn’t: it got a markup date.
That distinction matters. A markup vote isn’t a hearing. It’s a committee deciding it has counted the votes. When Sen. Chuck Grassley, who chairs the Senate Judiciary Committee, schedules an Executive Business Meeting rather than calling another round of testimony, that’s a signal, not a guarantee, but a signal, that leadership has assessed sufficient bipartisan support to advance. The scheduling of the June 11, 10:15 AM markup is the most concrete evidence in three years that this bill has momentum.
Where the NO FAKES Act Stands Today
S.4591 establishes a federal intellectual property right protecting individuals’ voices and visual likenesses against unauthorized AI replication. The scope is broad by design. It covers commercial use of AI-generated synthetic voice, AI-generated visual likeness, and deepfake video, regardless of whether the subject is a celebrity, a public figure, or a private person. The right is federal, which means it would create a national floor for enforcement rather than the current patchwork of state laws.
Sponsors span both parties and both chambers. On the Senate side, Sen. Marsha Blackburn (R-TN) and Sen. Chris Coons (D-DE) lead the bill. On the House side, Rep. María Salazar (R-FL) and Rep. Madeleine Dean (D-PA) have introduced companion bill H.R.8915. That bipartisan, bicameral structure is deliberate, and it’s why this iteration has moved further than earlier versions, according to legal analysis tracking the bill’s legislative history.
This is the fourth introduction in approximately three years. Previous versions stalled. None reached markup.
Who Supports the Bill and Why
The entertainment industry’s position is clear. The RIAA has formally endorsed S.4591, and the broader coalition behind the bill includes music industry organizations, film and television labor groups, and performing arts associations. Their argument is structural: AI voice cloning and synthetic likeness technology have developed faster than existing IP law can address. Right of publicity laws exist at the state level, but they vary enormously in scope, enforcement, and coverage. A federal standard would close that gap.
The bipartisan sponsorship structure reflects a genuine cross-party consensus on the underlying problem. Blackburn has been the most consistent congressional voice on AI deepfakes and synthetic media since the issue first emerged. Coons brings the Democratic side of the Senate Judiciary Committee. The addition of House sponsors Salazar and Dean signals that the House companion bill isn’t afterthought coordination, it’s a parallel track designed to compress the House-Senate reconciliation timeline if the Senate bill passes committee.
Individual rights advocates have also supported the bill’s framing. The core protection, your voice and likeness can’t be replicated commercially without your consent, is as applicable to a private person whose voice is scraped and cloned as it is to a recording artist.
Who Has Reservations and Why
Technology platforms haven’t publicly opposed the NO FAKES Act in its current form, but the safe harbor provisions are the fault line. How liability for hosting user-generated content gets structured will determine whether the bill is workable for intermediaries.
Who This Affects
The concern follows a familiar pattern from prior platform liability debates. If a platform hosts a synthetic voice clip that violates someone’s rights under the NO FAKES Act, what’s the platform’s exposure? Does the DMCA-style notice-and-takedown framework apply? Is there a knowledge standard, or is strict liability possible? These questions haven’t been definitively answered in the public version of the bill, and the negotiation going into the June 11 markup is focused on exactly this terrain.
The specific safe harbor terms aren’t publicly confirmed ahead of the markup, the Filter’s verification flagged this as contested but unresolved. What’s confirmed is that the issue remains live as of June 8. The markup may resolve it, or it may produce amendments that send the bill back for further negotiation on that specific point.
The New York State Law Parallel
This federal action doesn’t exist in a vacuum. New York’s AI Synthetic Performers Disclosure Law took effect on June 6, two days before this markup was announced. That law creates state-level disclosure requirements for AI-generated synthetic performers in entertainment productions.
The federal-state interaction is now live. New York’s law is already operative. A federal NO FAKES Act, if enacted, would establish a national standard that either preempts state laws like New York’s or coexists alongside them. That question, preemption or coexistence, isn’t settled in S.4591’s current public text.
For compliance teams operating across multiple states, this creates a sequencing problem. Do you build your synthetic voice and likeness compliance program around New York’s current requirements, around the anticipated federal standard, or around both simultaneously? The honest answer right now is both. New York’s law is in force today. The federal law is a committee vote away from advancing, and still multiple legislative steps from enactment.
The hub has been tracking the broader federal AI preemption debate. The NO FAKES Act fits into that landscape as a narrower, bipartisan carve-out, a federal floor for performer IP protection, rather than a broad preemption measure. That framing has helped it build support across the aisle while the broader preemption debate remains contested.
What Compliance Teams and Content Creators Should Watch After June 11
Three outcomes are possible on June 11. The committee approves S.4591 and sends it to the Senate floor, the scenario the bill’s sponsors are banking on. The committee approves S.4591 with amendments, most likely to the safe harbor provisions, which advances the bill but may require House-Senate reconciliation before enactment. Or the markup is postponed or the vote fails, returning the bill to the prior three iterations’ trajectory.
Each outcome has different compliance implications.
What to Watch
Analysis
The NO FAKES Act's fourth introduction follows the same structural logic as GDPR's eventual passage after years of stalled EU data protection reform: the underlying problem (unauthorized use of personal data / unauthorized AI replication of voice and likeness) didn't go away, and the technology causing it kept accelerating. What changed in both cases wasn't the policy argument, it was the political moment. For the NO FAKES Act, that moment may be June 11.
If the bill advances intact: Senate floor scheduling becomes the next milestone. Floor votes can move quickly or wait months depending on the legislative calendar. The House companion bill H.R.8915 creates a parallel track that could compress the timeline if the House moves independently.
If the bill advances with safe harbor amendments: The specific terms of those amendments will matter enormously for technology platforms. A narrow safe harbor that requires active knowledge of violations is workable. A broad strict liability provision is not.
If the markup stalls: The state-level patchwork, New York now, California’s existing right of publicity laws, and any additional state legislation that follows, becomes the operative compliance framework indefinitely.
The practical implication for today: if your organization uses AI voice cloning, synthetic likeness, or deepfake generation in any commercial context, you need a consent framework now. New York’s law already requires it for covered productions. A federal law would extend that obligation nationally. Building the consent infrastructure before the federal mandate arrives is cheaper than retrofitting it after.
The catch is the safe harbor language. Companies that host user-generated synthetic content, not just producers, may face obligations that aren’t fully defined until the markup concludes. Watch the markup outcome for whether safe harbor language survives intact, is narrowed, or is removed entirely. That’s the variable that determines platform liability exposure under a federal NO FAKES Act.
Four introductions in three years. The first three stalled. The fourth has a markup date, bipartisan sponsors in both chambers, and active entertainment industry endorsement from the RIAA. The June 11 vote won’t end this story, but it will determine whether federal AI performer protection finally has a path to the floor.