Odette CaplanWhen Scarlett Johansson went public in May 2024 about OpenAI's voice assistant sounding uncannily like her — despite her explicit refusal to license her voice — most Americans watched the drama unfold as a Hollywood story. By mid-2026, it has become their story too. The legal battles she has pursued are now reshaping federal law, and attorneys specializing in digital identity say the implications extend well beyond celebrities.
What the Scarlett Johansson Cases Actually Established
Johansson's confrontation with AI began when OpenAI launched GPT-4o featuring a voice called "Sky." She alleged Sky was deliberately designed to mimic her voice — the same voice she had declined to license months earlier after direct outreach from OpenAI CEO Sam Altman. OpenAI pulled the Sky voice but did not admit wrongdoing.
Then, in early 2025, a deepfake video circulated widely showing AI-generated versions of her and other public figures. Produced using generative AI tools by a digital marketing agency, the video was created without her knowledge or consent. It was seen by millions before platforms began taking it down.
The combined effect of these two incidents turned Johansson into the most visible face of a legal problem that affects anyone with a public digital presence: the gap between what AI can do with your image and voice, and what the law currently prohibits.
"We are at an inflection point," Johansson told the press in 2025. "There is a wave coming regarding AI that several countries have responded to responsibly — and the United States has not."
Congress has since moved. The NO FAKES Act (Nurture Originals, Foster Art, and Keep Entertainment Safe Act), introduced in updated form in 2026, would create a federal right of publicity for every American — not just celebrities — making it illegal to produce AI-generated digital replicas of a person's voice or likeness without their consent. As of August 2026, the bill has cleared committee and is awaiting a floor vote.
What Lawyers Say the Law Actually Covers Right Now
Here is where it gets complicated for ordinary Americans. Right of publicity law — your legal right to control commercial use of your identity — currently exists at the state level only, and coverage varies dramatically.
States like California and New York have strong protections, including posthumous rights for deceased public figures. Tennessee passed the ELVIS Act in 2024, the first state law specifically targeting AI voice cloning. Illinois, Texas, and Florida have their own digital likeness statutes. But roughly half of U.S. states have no right of publicity law at all.
"The patchwork is the problem," says a prominent IP attorney quoted in American Bar Association coverage of the Johansson case. "A deepfake video produced in a state without protections, featuring someone who lives in a protected state, creates instant jurisdictional confusion."
Federal copyright law offers limited protection: it covers creative works you've produced — recordings, films, photos — but it does not cover your face or voice as raw material. A deepfake that doesn't use any of your copyrighted content is, in many states, currently legal.
This is precisely why Johansson's advocacy matters beyond her personal cases. She has pushed lawmakers toward a federal floor that would make your likeness federally protected regardless of where you live — or where the AI company generating your digital twin is incorporated.
A Concrete Case: What Happens When It Happens to You
Consider this scenario, which mirrors situations attorneys are already seeing in 2026: a 34-year-old real estate agent in Ohio named Dana posts professional headshots and a video walkthrough tour on her agency's website. A competing agent — or a scammer — runs those images through a commercially available AI tool and generates a fake video of "Dana" endorsing a different brokerage. The fake is posted to Facebook and LinkedIn.
Ohio has no standalone right of publicity statute as of 2026. Here is what Dana's options actually look like:
- If the fake video uses a clip from her original walkthrough video (copyrighted): she can file a DMCA takedown immediately. Platforms must respond within 24-72 hours under existing copyright law. Legal cost to initiate: approximately $0–$500 for attorney-assisted filing.
- If the fake is entirely AI-synthesized (no use of her original footage): she cannot use copyright law. She may have a Lanham Act claim (false advertising/false endorsement), but that requires proving consumer confusion and is expensive — litigation can run $25,000–$100,000+ before settlement.
- If the federal NO FAKES Act passes: she would have a direct federal cause of action, with statutory damages of up to $5,000 per violation and attorneys' fees available — making it financially viable to pursue without a massive upfront investment.
The gap between those three outcomes is enormous. The difference between a free DMCA takedown and a $50,000 lawsuit depends entirely on whether the AI tool was clever enough to avoid using your source footage directly.
An attorney who specializes in digital identity rights can assess which legal pathway applies in under an hour of consultation — and in many cases, a cease-and-desist letter drafted by counsel is enough to end the behavior before it escalates to litigation.
Why This Matters More for Non-Celebrities
Johansson has resources — legal teams, publicists, and a platform to apply public pressure — that most people do not have. Her cases have been litigated publicly because she can absorb the cost of fighting back.
For everyone else, the asymmetry is brutal. The same AI tools that created fake Johansson content are available to anyone for under $50 a month. The cost to weaponize someone's likeness is nearly zero. The cost to fight it back, in the current legal landscape, can be prohibitive.
That asymmetry is precisely what the NO FAKES Act is designed to correct. Statutory damages with fee-shifting provisions mean that a lawyer can take a case on contingency if the violation is clear. Congressional testimony from the AI Policy Institute in June 2026 estimated that without federal protection, an estimated 4.2 million Americans will have non-consensual AI-generated content featuring their likeness created by 2028, based on current adoption curves.
This is not a hypothetical celebrity problem. It is a small-business-owner problem, a job-seeker problem, a teacher problem. Anyone who has posted a professional photo online — which is to say, nearly every working adult in America — has potential exposure.
What You Should Do in 2026
The legal landscape is shifting fast, but there are practical steps you can take now:
Audit your digital footprint. Run your name through AI image search tools periodically. Services like Google's reverse image search and emerging "deepfake detection" platforms can surface synthetic content using your likeness before it spreads widely.
Copyright your original content. Registering original photographs and videos with the U.S. Copyright Office ($65 per application) gives you the strongest legal hook if your likeness is used via deepfake that incorporates your source material. It is the one federal protection that exists today, regardless of your state.
Document everything. If you find fake content, take screenshots with timestamps before filing takedowns. Platforms are inconsistent about preserving evidence during removal processes, and you may need it if you pursue legal action.
Consult a digital identity attorney early. Many attorneys specializing in right of publicity and AI law now offer initial consultations specifically for deepfake and AI likeness situations. In states with strong protections — California, New York, Illinois, Tennessee — an attorney can often resolve violations quickly and at low cost. In states without protection, they can assess your Lanham Act and defamation options, and advise on the federal bill's likely impact on your case.
Watch the NO FAKES Act. If it passes in its current form, it creates a two-year window to file claims for violations that occurred after its enactment date. Keep records of any AI likeness violations you discover now, even if you cannot act on them immediately.
The Scarlett Johansson cases gave AI likeness law its most recognizable face. But the legal framework that emerges from them will protect — or fail to protect — people who will never make headlines. That is the real stakes of a policy debate that started with a Hollywood actress and is ending in everyone's inbox.
ExpertZoom connects Americans with licensed attorneys specializing in digital identity rights, right of publicity, and AI law. If you believe your likeness has been used without your consent, find a qualified attorney in your state today.
Note: This article addresses general legal principles and does not constitute legal advice. Consult a licensed attorney in your state for guidance on your specific situation.
