If someone used an AI tool to clone your voice and make it say things you would never say – could you actually stop them?

Today, a motivated impersonator can grab a short clip of you from social media, feed it into a voice-cloning model, and generate convincing audio that sounds just like you. The same is true for your face: generative systems can drop your likeness into fake videos, ads, or explicit content. Tools you’ve heard of – from ChatGPT-powered ecosystems to Gemini, Claude, and specialized voice and video models – make synthetic media creation faster and cheaper every month.

The law, on the other hand, is still moving at human speed. You technically have rights over your own name, image, and likeness in many US states, but those protections were mostly written long before deepfakes and AI voice cloning existed and they vary wildly from one state to another. That’s the gap the NO FAKES Act is trying to fill.

What is the NO FAKES Act?

The NO FAKES Act – short for “Nurture Originals, Foster Art, and Keep Entertainment Safe” – is a proposed US federal law aimed at stopping unauthorized AI-generated replicas of a person’s voice or visual likeness.

It began as a bipartisan discussion draft introduced in October 2023 by Senators Chris Coons, Marsha Blackburn, Amy Klobuchar, and Thom Tillis, specifically to address generative AI deepfakes of performers and ordinary individuals.Senate press release The idea is to create, for the first time, a federal right over highly realistic, computer‑generated replicas of you.

Key points from the draft and subsequent commentary:

  • It focuses on “newly-created, computer-generated, highly realistic” representations that are readily identifiable as someone’s voice or visual likeness, not just any cartoon or parody.
  • It would make people who create or knowingly distribute unauthorized digital replicas legally liable for damages.Bloomberg Law analysis
  • It covers both living and deceased individuals, treating control over AI replicas as a kind of property right.
  • It is explicitly framed as a response to generative AI, not just traditional image or voice misuse.

As of late 2026, the NO FAKES Act has not yet become law. It remains a live proposal with bipartisan interest, backed by performers’ unions and industry groups that want a national standard instead of state-by-state guesswork.Recent deepfake law tracker

Why do we need it if states already protect your likeness?

You might assume US law already handles this. After all, you’ve probably heard of celebrities suing over unauthorized ads or merch that use their face.

What you’re thinking of is the right of publicity – a state-level right to control the commercial use of your name, image, and similar attributes.Overview of personality and publicity rights The problem is:

  • There is no federal right of publicity in the US today.NO FAKES Act background
  • States differ on:
    • Whether they recognize this right at all
    • Whether it applies to voice or just image
    • Whether it continues after death, and for how long
    • Whether AI-generated “digital replicas” are clearly covered

Some examples:

  • California and New York have relatively strong publicity protections, but their laws were drafted around traditional media like ads, films, and merch, not synthetic replicas generated on demand.ABA analysis of right of publicity and AI
  • Tennessee’s ELVIS Act, effective July 2024, explicitly extends the right of publicity to AI-generated voice and likeness clones and even targets tools whose primary purpose is unauthorized impersonation.ELVIS Act summary
  • Other states are only just beginning to mention AI in their statutes, or focus more narrowly on deepfakes in elections or non-consensual intimate content.NCSL deepfake legislation tracker

In practice, this means that if an AI-generated video of you goes viral across platforms in multiple states, your legal options might depend on:

  • Where you live
  • Where the uploader lives
  • Where the platform is based
  • Which state’s law a court decides to apply

That’s the “patchwork” problem the NO FAKES Act is designed to address.

What exactly would the NO FAKES Act protect?

Think of three layers of “you” that can be replicated:

  1. Body and face – how you look in photos and video
  2. Voice – how you sound, including tone, accent, cadence
  3. Persona – your personality, catchphrases, and mannerisms

The NO FAKES Act is mainly aimed at the first two: highly realistic digital replicas of your voice and visual likeness, regardless of whether you’re a celebrity.

Under the discussion draft and subsequent descriptions:

  • An AI-generated clip that sounds and looks like you, presented as if it were you, would be covered.
  • The law would kick in when that replica is used without your consent and in ways that fall outside protected speech (for example, purely deceptive or commercial exploitation).
  • People harmed could potentially seek:
    • Injunctions (orders to take the content down)
    • Damages for harm to reputation or lost income
    • Possibly profits made from the unauthorized use, depending on the final text

Importantly, the bill tries to carve out First Amendment safeguards so that:

This is a difficult balance: you want creators to be able to use AI for obvious parody or films that clearly cast an actor as a historical figure, while still blocking nakedly deceptive “I never said that” fakes.

How would it interact with state deepfake and publicity laws?

You might wonder: if there’s a new federal rule, does that wipe out all the new state deepfake laws?

Based on current descriptions and analysis:

  • The NO FAKES Act would create a nationwide baseline right over AI-generated replicas of voice and likeness.
  • It would also preempt (override) future state laws in this narrow area – but it is expected to preserve existing state causes of action that were already in place by early 2025.Analysis of state AI voice/likeness laws and NO FAKES preemption

In practice, that likely means:

  • You would always have at least the federal NO FAKES protection for AI replicas, no matter which state you’re in.
  • In states with stronger or specialized laws (like Tennessee’s ELVIS Act or California’s new digital replica rules), you might have additional tools to go after impersonators, especially for non-AI or non-”highly realistic” uses.

For platforms (social media, video sites, music services), federal law would also simplify compliance:

  • Instead of trying to track 50 slightly different state standards, they could build one main policy around the federal rule, plus some state-specific add-ons where needed.
  • That could translate into clearer reporting tools, faster takedowns, and better support for people whose images and voices have been misused.

What does this mean for AI tools you actually use?

Most people now interact with generative AI through accessible front-ends like:

  • ChatGPT (OpenAI)
  • Google Gemini
  • Anthropic Claude
  • Consumer voice-cloning apps and avatar creators
  • Music and video generation tools that can mimic specific styles or voices

These tools already have terms of use that generally forbid uploading content you don’t have rights to, or creating harmful deepfakes of private individuals. But those are contractual, not statutory: they’re platform rules, not laws.

If the NO FAKES Act becomes law, several things likely change:

  • Tool makers may need to:
    • Add more robust identity consent mechanisms (e.g., “I certify I own or have permission to use this voice”).
    • Implement stronger safeguards that reject prompts asking to clone named people without proof of consent.
  • Platforms may face more pressure to:
    • Provide clear reporting workflows when you find a fake of yourself.
    • Respond quickly with takedowns to avoid liability for hosting or redistributing unauthorized replicas.
  • For you as a user:
    • If you are authorizing a clone of yourself (for a podcast, game avatar, or customer-service bot), you may see more standardized consent agreements and licensing language.
    • If someone else fakes you, you gain a clearer legal basis to demand removal and compensation, without having to navigate a maze of state law.

Behind the scenes, researchers are already mapping how rights of publicity and personality law intersect with AI voice and likeness, especially in light of high-profile controversies over AI-generated celebrity voices.Recent academic work on AI and vocal identity The NO FAKES Act would give those frameworks a concrete statutory anchor at the federal level.

What are the main concerns and criticisms?

Not everyone is cheering the NO FAKES Act as-is. Critics raise several concerns:

  • Free speech and creativity: Overbroad rights could chill legitimate creative uses like satire, biography, or transformative art, especially if platforms over-remove content to avoid risk.
  • Scope creep: A new federal right over voice and likeness might expand over time, potentially giving powerful entities more control over how history and public figures can be depicted.
  • Ambiguity: Terms like “highly realistic” or “readily identifiable” can be hard to apply in edge cases, especially as AI-generated media becomes even more photorealistic and customizable.
  • Enforcement inequality: Celebrities and big brands may be best positioned to use the new law aggressively, while ordinary people still struggle to afford lawyers or navigate complex complaints.

These debates mirror broader AI-regulation tensions: how do you protect people from real harms – fraud, reputational damage, harassment – without freezing legitimate expression or over-regulating experimental tech?

What you can do right now

You don’t have to wait for Congress to act to start protecting your own voice and likeness. Here are practical next steps:

  1. Lock down your own usage rights

    • When you use AI voice or avatar tools, read the permissions: are you giving the service the right to reuse your voice or face in other products?
    • Prefer tools that clearly state your continued ownership of your likeness and voice data and allow deletion.

  2. Monitor and respond to misuse

    • Set up simple alerts on your own name and handle (searches, social listening tools) to spot suspicious content that “sounds like you”.
    • If you find a fake, use platform reporting tools immediately – many sites already have categories for “impersonation” or “deepfake,” even before any federal law passes.

  3. Stay informed and, if you care, speak up

    • Follow developments around the NO FAKES Act and related state laws from credible legal or policy trackers, not just viral posts.
    • If this matters to you professionally (you’re a creator, streamer, teacher, or business owner), consider contacting your representatives or industry groups to share how you want your voice and likeness protected in the AI era.

AI can be an incredible amplifier for your creativity – or for someone impersonating you. The NO FAKES Act is one attempt to make sure that when a voice or face appears on screen, it’s there with the real person’s permission. Understanding how it works puts you in a better position to insist that your identity stays yours, even as the technology gets better at pretending otherwise.