A new kind of viral content keeps showing up online: AI-generated ad videos featuring the faces of public figures, created without their knowledge or consent. The results are often impressive enough that the creator turns the video into a portfolio piece, selling "AI-powered ad creation" services to brands at a fraction of traditional production costs.
It looks like innovation. But beneath the polish sits a question most creators skip past entirely: does impressive output justify using someone's face without permission?
Legally, the answer is no, and the legal framework that explains why has a name: the right of publicity.
What Is the Right of Publicity?
The right of publicity is the legal principle that protects individuals from the unauthorized commercial exploitation of their identity, their name, image, voice, signature, or other recognizable attributes. It is particularly relevant to public figures, whose names, images, and other recognizable attributes often hold significant commercial value.
The core idea is simple: a person's identity has economic value, and that value belongs to them, not to whoever happens to have the technical means to reproduce it.
How It's Recognized Around the World
United States: A Patchwork of State Laws
There's no single federal right of publicity statute in the US. The right is recognized in roughly thirty-five states, mostly through statutes, New York protects it only by statute, while California recognizes it through both common law and statute.
California's version is among the strongest. Under California's Civil Code section 3344, individuals are entitled to control the commercial use of their name, voice, signature, photograph, or likeness, and violations can result in damages including actual damages, profits from the unauthorized use, and statutory damages.
The legal foundation goes back decades. The Supreme Court solidified the basis for the right of publicity in Zacchini v. Scripps-Howard Broadcasting Co. in 1977. A landmark case that's especially relevant to AI mimicry involved a different kind of impersonation: in Bette Midler v. Ford Motor Co., a sound-alike singer was used in place of Bette Midler for a commercial, and the court held that when a distinctive voice deliberately imitates a known performer to sell a product, this violates the right of publicity. The same logic now extends naturally to AI-cloned voices and faces.
Tennessee's ELVIS Act: A Direct Response to AI
Recognizing that traditional laws weren't built for generative AI, Tennessee took direct legislative action. On March 21, 2024, Tennessee enacted the Ensuring Likeness Voice and Image Security (ELVIS) Act, expanding the state's existing publicity law to specifically address unauthorized AI-generated voice clones and deepfakes, and it may serve as a model for other states addressing generative AI exploitation of "integrity, identity, and humanity."
Federal Movement in the US
Because state laws vary so widely, there's growing pressure for a unified federal standard. Right of publicity laws generally create a private right of action for the unauthorized commercial use of someone's name, image, and likeness, so if a company uses a well-known athlete's face in a TV commercial without permission, that athlete could sue. Separately, the Lanham Act establishes a federal cause of action for "false endorsement," which applies when someone's identity is used in a way likely to confuse consumers into believing they endorse a product, overlapping with and reinforcing state-level publicity claims.
Beyond the US
In today's global digital economy, applying right of publicity laws that are governed only at the state level is increasingly difficult, a gap that matters enormously when AI-generated content can be produced in one country and distributed worldwide within seconds. Many other jurisdictions address the same harm through different legal doors: image rights, personality rights, data protection law, or copyright over a person's portrait, but the underlying principle is consistent globally: a person's face is not free creative material just because it's technically reproducible.
Real Cases That Show This Isn't Theoretical
This isn't a hypothetical legal debate, it's already playing out in courtrooms and headlines.
Tom Hanks and the AI dental ad. In 2023, actor Tom Hanks alerted fans that an AI-generated replica of him was being used to advertise a dental plan without his permission, a textbook case of identity misappropriation that could trigger both state right-of-publicity claims and federal false endorsement claims.
Matthew McConaughey's trademark strategy. Rather than rely solely on right-of-publicity law, McConaughey's legal team secured eight federal trademark registrations designed to deter unauthorized AI-generated versions of his identity, including a sound mark for his catchphrase "Alright, alright, alright!" complete with a formal description of its pitch and cadence. This gives his team a way to argue that an AI-generated imitation of his voice or image could confuse consumers into thinking he endorsed a product.
The Avatar likeness lawsuit. A more complex case now testing the boundaries: a lawsuit against James Cameron involves a likeness that belonged to a minor, was processed through digital production technology, and generated billions of dollars across a global franchise, all without the knowledge or consent of the person whose face made the character possible. Legal observers note the outcome could help define what "likeness rights" mean in the age of AI and generative media.
These cases share a common thread: the technology being "AI-generated" rather than a direct photograph didn't make the use any less actionable.
Why "It's Just AI-Generated" Isn't a Legal Defense
A common justification creators give is that the content isn't a real photo, just an AI output, so it shouldn't count as misappropriation. This reasoning falls apart for a few reasons:
The whole point is recognizability. Ad creators use a public figure's face precisely because audiences recognize them. If the face weren't recognizable, the commercial value would disappear too, which means recognizability is exactly what makes it a violation.
AI still relies on someone's real likeness as input. Generative tools need reference images of a real person to produce a convincing output. That source material is itself protected.
Courts have already addressed this logic. Even inadvertent appropriation through AI of a person's name, likeness, voice, signature, identity, or persona can be actionable under common law. The method of creation, camera versus algorithm, doesn't change the underlying harm.
The commercial purpose is what matters. In some jurisdictions, like Illinois, the right of publicity protects against unauthorized use of someone's identity specifically when that use is for a commercial purpos, meaning merely generating a deepfake isn't automatically a violation, but using it to sell something almost always is.
What's at Stake, for Creators and Brands Alike
The risk isn't limited to the person who built the AI tool. A single unauthorized AI-generated voice clone or deepfake video can damage someone's commercial brand within hours of release, and enforcement is increasingly aggressive: image scraping for commercial use violates publicity rights even when the source images were publicly available, and bot-generated content is increasingly fabricating celebrity endorsements without permission.
For brand owners purchasing these "AI-powered ad" services, the exposure is real too. If a brand commissions and publishes a video using someone's unauthorized likeness, it can be named alongside the creator in any resulting legal action, ignorance of how the content was made isn't typically a defense once it's live and generating commercial benefit.
What Responsible Use Actually Looks Like
None of this means generative AI has no place in advertising. The issue is consent, not the technology itself.
- Use models or talent who've explicitly licensed their likeness for AI use, with clear contractual terms covering scope and duration.
- Build original, fictional characters with AI rather than recognizable replicas of real people.
- If you want a specific public figure, go through proper channels, their management or agency, the same way you would for a traditional endorsement deal.
- Disclose AI-generated content clearly. Transparency protects audiences and reduces legal exposure simultaneously.
The Bottom Line
Generative AI has genuinely expanded what's creatively possible in advertising. But technical capability was never the same thing as legal permission. A person's face remains theirs to control commercially, and that protection doesn't disappear just because the method of reproduction shifted from a camera to a prompt.
Before turning an AI-generated likeness into a sales pitch for your services, ask the only question that actually matters: did this person agree to this? If the answer is no, the polish of the output doesn't change what it is, it's not innovation, it's infringement.
Comments
Post a Comment