The War on Digital Replicas: The NO FAKES Act Advances
In a recent AI Watch article, we looked at New York's Synthetic Performers Act and the obligation to disclose when an advert features a digitally created human who is not a real person. This article turns to the mirror image: the use of digital replicas of people who very much are real. A United States federal bill on that very issue - the NO FAKES Act - has just taken its most significant step yet towards individuals having rights over their own voice and visual likeness, and the ability to sue others for the unauthorised use of their voice or likeness. Although this is a US bill, this latest step comes at a moment when the UK government is actively weighing up whether to give individuals a new right over their digital likeness.
What is the NO FAKES Act?
On 20 May 2026, a bipartisan group of US legislators reintroduced the latest version of the Nurture Originals, Foster Art, and Keep Entertainment Safe Act 2026 (the “NO FAKES Act”) onto the Senate floor. This is the fourth time the bill has surfaced in roughly three years, with previous versions in 2023, 2024 and 2025 failing to become law. The underlying aim has remained consistent throughout these versions: to give every individual a federal, licensable property right in their own voice and visual likeness, together with a means of enforcing (through a notice and takedown procedure) against the unauthorised digital replicas that generative AI has made much easier to produce.
The 2026 version carries new revisions and new co-sponsors across leading media and technology companies that may well give it the strongest chance yet of enactment. Senators Marsha Blackburn (Republican), Chris Coons (Democrat), Thom Tillis (Republican) and Amy Klobuchar (Democrat) joined House Representatives María Salazar (Republican), Madeleine Dean (Democrat), Nathaniel Moran (Republican) and Becca Balint (Democrat) in co-sponsoring what is now one of the most closely watched federal AI bills.
Individuals or companies found to be in breach of the NO FAKES Act may be fined $5,000 or $25,000, respectively, per work embodying the unauthorised digital replica. Online services providing unauthorised digital replicas may be fined $5,000 per copy, display, transmission or instance of the unauthorised digital replica, capped at $750,000 per replica, or $750,000 per product or service.
A meaningful step forward
On 18 June 2026, the Senate Judiciary Committee unanimously advanced the NO FAKES Act. This is likely to form part of a wider package of AI and children’s safety bills – including the Kids Online Safety Act and the App Store Accountability Act – that Senator Blackburn is developing with the US Government. As Senator Blackburn put it, "It is imperative that we put this national standard in place for voice and visual likeness protection of creators," adding that "America needs one set of rules for AI, and NO FAKES is a critical component of that rulebook."
What's new in the 2026 bill?
First, and most significantly, the bill introduces a counter-notification process modelled on the US Digital Millennium Copyright Act (“DMCA”) that allows users to contest the removal of digital replicas. In practice, where an individual has exercised their right to request the removal of a digital replica, the user who placed the material on the online service may submit a written counter-notification disputing the takedown and seek to restore the content.
Once a counter-notification is received that complies with the notification requirements, the online service provider that opted to remove the material must promptly send the complainant a copy of the counter-notification and then, within 14 days of the counter-notification, restore the content – unless an eligible complainant brings an action, in which case the material must be removed again. The mechanism is policed at both ends: a counter-notifier who knowingly and materially misrepresents the position faces liability for the greater of $25,000 per counter-notification or actual damages plus fees, with a "failure to consider in good faith" itself treated as a “knowing material misrepresentation”.
Secondly, the 2026 bill adds a new liability exclusion for nonprofit libraries or archives and accredited nonprofit educational institutions, with respect to activity undertaken without any purpose of direct or indirect commercial advantage, extending to their employees and network-access providers. This sits alongside other First Amendment carve-outs in the bill.
The First Amendment tension
These additional protections are not incidental - they go to the heart of the bill's most contested issue. The First Amendment to the US Constitution protects freedom of speech. However, both the NO FAKES Act and the Kids Online Safety Act have come under fire over First Amendment concerns, and a coalition of free speech groups (including the Center for Democracy & Technology and the Electronic Frontier Foundation) wrote to Judiciary Committee leadership urging members not to advance the bill in its current form, with Republican Senators Ted Cruz, Mike Lee and Eric Schmitt raising concerns about the bill's effects on free speech.
Senator Cruz illustrated his concerns by pointing to former reality television star Spencer Pratt's AI-generated campaign videos during his failed Los Angeles mayoral run – including portrayals of himself as Batman and Luke Skywalker – which often ran without any disclosure that the content was AI-generated. Senator Cruz said these “hysterical” campaign videos were " a good example of what should be protected and not fall within a bill like this". This captures the central tension: how to protect individuals against unauthorised replicas without sweeping up parody, satire and political commentary or endangering freedom of speech.
Where does UK law stand?
While Washington makes progress towards a federal likeness right, the UK has no equivalent. As the government’s own Report on Copyright and AI (March 2026, the “Report”) confirms, there is no general personal image or personality right – only a patchwork of data protection, online safety, criminal offences, defamation and IP-related rights such as trade marks and passing off. This patchwork leaves gaps: copyright protects the work rather than the identity of the person within it; performers’ rights are tied to a recording of a performance and so may not protect against a wholly-new AI-generated performance where the performer’s voice has been cloned; passing off (in the context of image rights) helps well-known artists with goodwill but rarely lesser-known performers or members of the general public; and a voice or likeness is difficult to protect meaningfully via a trade mark.
The harm is not hypothetical – think of the ‘fake Drake’ track ‘Heart on My Sleeve’ or the allegedly unauthorised recreation of Stephen Fry’s voice for narration. Criminal law has been strengthened for sexually explicit cases, but many non-consensual replicas fall outside those offences. The direction of travel is clear from the Report: the government has proposed to explore options to address the risks of unauthorized digital replicas (including considering a possible new digital replica or personality right), while investigating approaches adopted internationally (including in both the EU and USA, with the Report expressly flagging the NO FAKES Act). A US federal standard established through the enactment of the NO FAKES Act would only intensify the pressure on the UK government to act.
Why this matters for UK and international businesses
If enacted, the NO FAKES Act would represent the first nationwide publicity right in the United States – a powerful new tool for rightsholders, including estates, and their licensees. For now, the bill still needs to survive committee and floor votes and, in the meantime, the patchwork of state right-of-publicity laws (including some, like California's, that specifically address digital replicas) remains the operative regime, alongside emerging routes such as federal trade mark law.
A federal US likeness right could have extraterritorial reach in practice: global content, talent deals and platform distribution are rarely limited by jurisdictional lines, and UK or international studios, agencies, brands and AI vendors whose output reaches US audiences could find themselves caught by the NO FAKES Act.
Businesses need to take a holistic approach to developing standard operating procedures and workflows, both in relation to replicas of identifiable individuals (covered by the NO FAKES Act) and fabricated humans who do not resemble specific individuals (covered by the New York law discussed in our earlier article).
Finally, the NO FAKES counter-notification model signals where platform compliance may be heading: a DMCA-style, process-driven regime for likeness disputes that businesses will need to operationalize and develop audit trails for in order to defend their actions.
Practical takeaways
For businesses that do or are likely to create, license, publish, share, distribute or host content containing digital replicas, the following measures are worth considering:
- Adopt a holistic approach by mapping exposure of the business to existing and potential likeness rights across key jurisdictions, with processes to monitor the progress of relevant legislation (such as the NO FAKES Act) and update working practices as the position evolves.
- Prepare for a DMCA-style counter-notification workflow, including designated agents, intake processes and the 14-day restoration timeline, if the NO FAKES Act is enacted.
- Build digital-replica identification and disclosure into creative workflows and approval processes, and require active flagging by agencies, studios and AI vendors.
- Allocate likeness clearance, disclosure responsibility and liability clearly in contracts.
- Review talent, influencer and production agreements to address digital replicas and likeness use consent.
- If the NO FAKES Act is enacted, document takedown and counter-notification decisions, given the significant misrepresentation penalties on both sides of the process.
A clear picture soon emerges: whether the human on screen is entirely fabricated or a replica of a real person, the law is moving steadily towards transparency, consent and accountability. Digital replicas are no longer a niche hypothetical; they are quickly becoming a mainstream compliance issue. The NO FAKES Act may still face a difficult path through Congress, but businesses that act now to build clearance, disclosure and clear contractual responsibility into their workflows will be far better placed if and when likeness protection becomes a global regulatory norm.
