What’s Real, What’s Fake: The Right of Publicity and Generative AI

Since the release of ChatGPT in November 2022, there has been a seemingly endless onslaught of coverage about the chatbot and generative AI software like it. These stories range from the fear-inducing (the world will be taken over by AI-controlled bots) to the ridiculous (a lawyer relied on ChatGPT to draft a brief and had to face a very mad judge when it turned out that the software fabricated cases). In the legal realm, much of the focus of conversations around ChatGPT has been on the intersection between copyright law and AI, and rightly so. From questions around whether a work generated by AI qualifies for copyright protection to investigations around if an AI app’s ingestion of copyrighted works is covered by fair use, generative AI technology has introduced a lot of big issues, and answers are being decided in real time.

One often overlooked area of intellectual property where AI is poised to have a big impact is the right of publicity. By way of background, the right of publicity allows individuals to control the commercial exploitation of their identity and reap the rewards associated with their fame or notoriety by requiring others to obtain permission (and pay) to use their name, image, or likeness. As the law in this area develops along with the technology, there are several key issues practitioners need to be aware of.

The challenge with generative AI is that it makes the creation of a credible simulacrum of a celebrity much, much easier. In the past, this would have required finding a real person who could sound like a celebrity or be done up like a celebrity. Generative AI allows users to skip this. For example, earlier this year, an anonymous creator operating under the name Ghostwriter uploaded a song in the style of musicians Drake and The Weeknd. The song, which effectively mimicked the real artists, quickly went viral and was played millions of times on TikTok, Spotify, and YouTube before being removed at the behest of Universal Music Group, which is home to both the artists.

While in this case, there was no action for violation of Drake and The Weeknd’s right of publicity, this will not necessarily be the case going forward, particularly as the technology continues to evolve and its output becomes even more sophisticated.

Practitioners representing entities that make and use generative AI need to be aware of the contours of the right of publicity so they can minimize risk of such claims or appropriately address them when they arise. Here are five things to keep in mind as the technology develops:

1. Right now, there is no federal cause of action for the right of publicity. For the time being, the right of publicity is a creature of state law, with about two-thirds of states recognizing some form of this claim. This means that there is some variation between states and that many right of publicity claims are brought in state—not federal—court.
2. It’s not just a replica of a celebrity’s face or body that is protected. Rather, the courts may give plaintiffs a bit of leeway to proceed where a defendant has used characteristics that call a celebrity to mind. For example, the Ninth Circuit, applying California law, recognized that an ad featuring a robot dressed in a long gown, wearing a blonde wig, and turning block letters as part of something that looked like a game show was meant to depict Vanna White and infringed her right of publicity by appropriating her identity. Similarly, the Ninth Circuit found that Bette Midler could bring a right of publicity action against an advertiser and its ad agency for creating a commercial with a singer who was hired to sound like Midler.
3. The output of a generative AI platform may be protected by the First Amendment. This means that, as with other forms of intellectual property, an individual’s interest in protecting their right of publicity must be balanced against the rights of users to creative or self-expression. This is true even if the product of generative AI is being sold for a profit. Taking a non-AI example, the Sixth Circuit concluded that Tiger Woods could not maintain a right of publicity claim against the painter of a painting entitled The Masters of Augusta that commemorated Wood’s 1997 victory at the Masters. In this case, the Sixth Circuit found that, although the painter sold prints of the painting, the work was protected by the First Amendment because there was substantial “creative content” that outweighed Woods’s interest in profiting from his image.
4. It is unclear if Section 230 applies. This section of the Communications Decency Act can limit a platform’s liability for its users’ conduct. However, at least for now, there is no case law saying that this law applies to AI-generated materials. There are also questions of what that would look like, such as who is the creator of a work generated by an AI platform based on user prompts and whether right of publicity claims fall within an exception to Section 230 for intellectual property claims. Until these details become clearer, platforms should be wary of hosting AI-generated materials featuring a celebrity.
5. The nature and extent of the license granted by a celebrity may mean that the celebrity’s right of publicity claims are preempted. Thus, for example, in late 2022 the Second Circuit Court of Appeals affirmed a district court’s dismissal of a lawsuit against Sirius XM by John Melendez, who used to appear on The Howard Stern Show under the moniker “Stuttering John.” The plaintiff claimed that Sirius XM breached his right of publicity by airing past episodes and ads that featured him. The court found that, in light of Sirius XM’s license to these materials, the plaintiff’s claims were preempted by the federal Copyright Act.

Practitioners need to monitor each of these areas for new developments, particularly as the technology continues to improve and change.