By: Maya Schmaling ‘28
Volume XI – Issue I – Fall 2025
I. INTRODUCTION
On January 14th, 2025, the Better Business Bureau (BBB) released a scam alert warning consumers about an eerie new advertising strategy. It had received numerous reports of instances in which consumers had bought a product because they had seen a well-known celebrity endorsing it, only to discover that the product did not exist. In fact, the endorsement did not exist either; the celebrity was generated by artificial intelligence, which, as the BBB warned, is now sophisticated enough to make the endorsement look real. [1] The alert aligns with a growing number of high-profile cases involving AI-generated celebrity endorsements. In 2023, Tom Hanks debunked the authenticity of a video advertisement featuring an AI version of himself promoting a dental plan. [2] In a series of advertisements, Oprah Winfrey and high-profile doctors appeared to recommend a weight loss supplement. [3] Despite the general dubiousness of the products, the attachment of a familiar face to the advertisement led at least some viewers to lose their money. [4]
A product approved by a celebrity tends to be more tempting to potential buyers. In purchasing such products, buyers are faster and more confident in their decisions, and the celebrity presence often outshines the product itself in grabbing attention. [5] The impact is significant: a Harvard Business School study found an approximate 4% increase in sales, or an average of $10 million annually, amongst brands that had used celebrity endorsements. [6] Commissioning an endorsement with a high-profile figure can cost millions, but the commercial advantage offered in advertising provides a strong incentive for those who can afford it. [7] The featured celebrity can also benefit by profiting from their persona, so long as it remains a limited and marketable commodity.
Publicity rights are a legal framework meant to protect against the unauthorized use of another’s likeness, including appearance, voice, and personality, for commercial gain or promotion. Rights of publicity are not exclusive to celebrities—but for commercial use issues, particularly the use of an individual’s likeness in advertising, notable personalities are most often involved. [8] The monetary consequences for companies who violate publicity rights help to preserve the commercial value of a famous persona. However, the effectiveness of this form of regulation comes into question when the bar is dramatically lowered for endorsement falsification, which has happened with the development of AI technology capable of generating true-to-life pictures and voices.
It may seem that non-celebrities have little reason to worry about the issue of AI encroaching upon publicity rights. [9] But buyers in general are harmed by commercial deceit, and the increased regulatory difficulty of publicity right violations has broad legal, social, and economic implications. Complicating the situation is the patchwork of statutory and common law that forms publicity rights across the country, as the right operates as a state-level protection differentiated slightly between states. A solution lies in federal policy, which presents a potential method of widespread regulation more effective than adapting state laws on a case-by-case basis. Relevant policies have been introduced but have yet to pass through congress, and there are inherent challenges to implementing sweeping regulatory legislation. However, it is necessary to address this issue on a federal level.
II. PRECEDENT & DEVELOPMENT
i. Statutory Origins of the Right of Publicity
Although the first lawsuit concerning the right of publicity was not brought until 1953 in Haelen Laboratories, Inc. v. Topps Chewing Gum, Inc., its statutory foundation preceded Haelen by half a century. [10] Because there is no single “Right of Publicity” provided by an overarching federal law, pinning down a singular origin is a complex matter. While it is based in longer-standing privacy law, only an estimated half of states explicitly recognize the right of publicity, and the criteria and stringency of the right varies widely between states. [11]
A federal protection of privacy did not exist until Congress enacted the Privacy Act of 1974, but its statutory roots can also be traced much further back. [12] The advent of cameras and the popularization of photography as a component of advertisement and journalism at the end of the 19th century introduced the issue of exploitative or nonconsensual publicization. The milestone advancement in technology cleared a new and uncomfortable path for the violation of, as Samuel D. Warren and Louis Brandeis wrote in their 1890 Harvard Law Review article, one’s right “to be left alone.” [13] Despite being legally unenforceable, Warren and Brandeis’ article was critical for the development of privacy law. Another decade passed before the law was established by a court, when Warren and Brandeis’ article resurfaced as the basis of the plaintiff’s argument in Roberson v. Rochester Folding Box, Co. (1902). A forerunner of the typical right of publicity cases that were to come, the case centered around Rochester Folding Box Company’s use of Abigail Roberson’s portrait in an advertisement without her consent. In its ruling, the court found that Robinson had been harmed, and she was compensated. But a key detail differentiated the grounds of the ruling in this case from what would develop as the foundation of publicity rights: the nature of the harm the plaintiff had suffered, which was found to be mental anguish. The monetary damages awarded were justified by emotional harm rather than economic harm, as a legal foundation for the commercial value of identity had yet to be established. [14]
The statutory right to privacy was codified by the state of New York in 1903 in sections 50 and 51 of the Civil Right Law. Section 51 addresses damages: “Any person … may also sue and recover damages for any injuries sustained by reason of such use and if the defendant shall have knowingly used such person's name, portrait, picture, likeness or voice in such manner as is forbidden or declared to be unlawful by section fifty of this article.” [15] As in Roberson, the focus is on mental injury rather than economic loss. For decades, mental injury due to the violation of privacy rights remained the actionable basis for nonconsensual portrayal or impersonation in advertisement.
Then, in the 1950s, a chewing gum manufacturer used baseball players’ photographs despite the players’ existing exclusive contracts with another gum manufacturer. The other manufacturer sued, alleging that the players’ privacy rights had been violated. The resulting case, Haelen Laboratories, Inc. v. Topps Chewing Gum, Inc. (1953) was seminal for the right of publicity because the issue that arose indicated a need for another, less personal filter through which to evaluate certain privacy-related cases. The defendant, Topps Chewing Gum, drew on the injury focus in New York law at the time, arguing that Haelen had no valid legal claim because privacy rights essentially hinged on hurt feelings. The court saw a need for expansion. “We think that, in addition to and independent of that right of privacy, a man has a right in the publicity value of his photograph,” wrote Jerome Frank in the court’s majority opinion. “For it is common knowledge that many prominent persons, far from having their feelings bruised through public exposure of their likenesses, would feel sorely deprived if they no longer received money for authorizing advertisements.” [16] The court’s ruling was an application of the right to privacy to the ability to control one’s identity in commerce, formalizing the legal concept of a persona as a type of property that could be monetized. In doing so, the court in Haelen coined the “right of publicity” as an area of law distinct from mental injury-focused privacy law. [17] Haelen marked the first instance of legal recognition of a right to publicity, and over the following decades more than half of states adopted right of publicity laws, either through the passage of statutes or by common law. [18]
ii. The Lanham Act
The Lanham Act, enacted by congress in 1946, is distinct from the right to publicity and has a basis in intellectual property law rather than privacy, although the two overlap with the focus on commerce. [19] In the realm of false endorsements, it serves as an equally important legal protection and is subject to the same uncertainties in the midst of technological breakthroughs. Unlike publicity rights, the Lanham Act is a single, federal statute. In the context of celebrity endorsements, identity is considered a trademark rather than a commercial property, and the Act aims to protect consumers from being misled by false endorsements. Consequently, claimants of false endorsement must prove that the advertisement in question is likely to “cause confusion” and “deceive as to the affiliation, connection, or association of such person with another person, or as to the origin, sponsorship, or approval of his or her goods, services, or commercial activities.” As with publicity rights, a claimants’ likeness must have a commercial value which was exploited by another for commercial gain. [20]
Because both publicity rights and the Lanham Act offer protections against false endorsements, one may employ both in appropriate states. For example, in Waits v. Frito-Lay, Inc. (1992), singer Tom Waits sued Frito-Lay under both the False Endorsement section of the Lanham Act and California’s common-law right of publicity for an alleged impersonation of his distinct singing voice in an advertisement. The court ruled in his favor for both, finding that the advertisement could mislead the public into believing that he endorsed the product and that Waits’ voice was a personal characteristic Frito-Lay had appropriated for commercial gain. [21]
iii. Lohan v. Take-Two Software (2025): Adaptation to other technological shifts
While advancements in AI pose a novel threat to the effectiveness of publicity protection, the right of publicity has persisted through decades of a changing technological backdrop. In the onset of the digital age, similar issues about applying a pre-digital law in a new context emerged. For example, the ease of dissemination made possible by the rise of the internet changed the advertising landscape dramatically. [22] The precedent provided by the resolution of such issues is a useful tool for determining the state of the right at present.
Lohan v. Take-Two Software (2018) serves as an example of publicity rights being interpreted to cover new forms of impersonation. Actress and singer Lindsey Lohan brought a lawsuit against the creators of “Grand Theft Auto V,” a video game featuring a similar actress-singer character named Lacey Jones, alleging that her persona had been pasted onto the character without her consent in violation of the New York Civil Rights Act. Two aspects of the court’s decision in Lohan are notable. It interpreted the language of the Civil Rights Act, which explicitly includes “portraits” in reference to graphical representations of a person, to include digital avatars. [23] However, the court affirmed the lower dismissal of Lohan’s claim on the grounds that the character Lacey Jones could not be reasonably confused with Lohan herself, and rather was a satirical representation of a more general “modern, beach-going young woman.” [24]
The court expanded its interpretation of the original statute to encompass a modern situation, because the fact that the “portrait” was digital did not change its nature as a potential replication of a persona. The court cited similar decisions of incorporation in Burck v. Mars, Inc, Allen v. National Video, Inc., and Ali v. Playgirl Inc., including a quote from Allen that “any recognizable likeness, not just an actual photograph, may qualify as a ‘portrait or picture.’” [25] The consequence of economic loss, at the heart of publicity rights, is as relevant with digital portraits as with any non-digital impersonation. The medium of impersonation, no matter how enhanced by technology, does not itself change the application of the law. As the court put simply, “there can be no appropriation of a plaintiff's likeness for commercial purposes if he or she is not recognizable from the image in question.” Because Lohan was not recognizable from the avatar, the court decided, the commercial value of her persona was not in jeopardy. [26]
It follows that an AI-generated likeness would also fall under this definition of “any recognizable likeness.” Importantly, AI usage shifts the question of recognizability. Even if the court in Lohan had ruled that the character Lacey Jones would be commonly seen as a portrayal of Lohan, there is an inevitable layer of fantasy added by the game’s style constraints. A CGI avatar can closely resemble a real person, but AI is unique in that it can make an indistinguishable portrait with a fraction of the human effort. [27] There are a host of implications introduced if any company can almost instantly and cheaply create an endorsement identical to the real thing.
III. POST-AI INVENTION
i. Lehrman v. Lovo, Inc. (2025)
The legal landscape surrounding photorealistic AI-generated celebrity endorsements lacks a precedential foundation, largely due to the fact that the technology has only recently developed to its current stage in which such a feat is possible. Only since 2024 have video-generating softwares been able to produce synthetic content mistakable for reality—such as OpenAI’s Sora in December of 2024 and Google Deepmind’s Veo 3 in May. Even one year prior, the clearest AI videos were still distorted and fuzzy. Two New York Times articles, published less than a year apart, exemplify the rapid development. The first, in 2023, describes a company called Runway AI as positioned at the frontier of video generation, but the examples it shows are shaky, low quality, and distorted. [28] The second, in 2024, is an interactive quiz titled “Can you tell what’s real?” Readers can guess whether several short-form videos are real or AI. [29] The difference in quality is striking. It would be going too far to say that AI video generation is consistently perfectly indistinguishable from reality, but the rapid development in the past few years signals a near future in which it will be. [30] AI voice and image generation have developed similarly rapidly, and have arguably approached the realism stage even faster than video generation has. [31] All three types are potential components of a false endorsement.
Despite minimal precedent, a few cases in recent years have touched on rights of publicity and AI use. Voice actors Paul Lehrman and Linnea Sage brought a lawsuit against Lovo AI in Lehrman v. Lovo, Inc. (2025) alleging that the company had used their voice recordings to train its voice-“cloning” software without their permission. The plaintiffs accused Lovo of using Lehrman’s voice in ads to promote the software and allowing customers to freely use both actors’ voices. They argued that the implied affiliation with Lovo constituted false endorsement, and made two claims under section 43(a) of the Lanham Act. [32] The plaintiffs alleged that they were well known enough to cause confusion about a partnership with Lovo, and the court disagreed: “Plaintiffs here use their voices in ways that are clearly separable from their identities and personalities.” It decided that the actors’ voices served solely as a product rather than a “source identifier,” or a branding asset associated with a source—and therefore were not protected by the Lanham Act. [33] The plaintiffs argued that they lost profits due to the cheaper cost of using Lovo compared with their hiring fee, but since the loss was not attributed to false claims in Lovo’s advertising, the court found the false affiliation protection provided by the Act inapplicable. [34]
The plaintiffs were successful in their claims under the New York Civil Rights Law, to which the court applied the same logic as in Lohan: “Construing the Civil Rights Law to exclude digital clones would frustrate the statutory purpose, and, for all practical purposes, enable commercial entities to appropriate individuals' identities without restraint.” [35] Lovo is important as one of the first cases in this category of AI false endorsements, and one in which the court considered claims under a spread of privacy and publicity laws in an AI context. While the relevant law was not created with AI context in mind, the decision in Lovo exemplifies its flexibility.
ii. Scarlet Johansson and Lisa AI
In a still-developing situation, Scarlet Johansson reportedly took “legal action” against a company that used an AI-generated version of her voice to promote its app, “Lisa AI: 90s Yearbook & Avatar.” The ad includes a real video clip of Johannson speaking, followed by AI-generated photos of her and her voice endorsing the app. “It’s not limited to avatars only,” the voice says. “You can also create images with texts and even your AI videos.” [36] The nature of Johannson’s legal action is not yet publicly available, and it is unclear whether the case will proceed to court. However, this is significant as an instance of a high-profile celebrity taking on the legal challenge of fighting an AI-created endorsement rather than just clarifying the ad’s falsity, as Hanks and other celebrities had done.
Despite not yet having undergone an evaluation in court, this situation can be contrasted with Lohan in its consumer confusion aspect. Recall that Lohan’s claim was dismissed because “[t]he artistic renderings are indistinct, satirical representations of the style, look, and persona of a modern, beach-going young woman.” [37] There is a blatant difference in realism between an animated avatar and a deepfake. Both can be compared to the person they allegedly represent, but only one can be actually confused with that person, and Johanson’s case exemplifies the latter.
iii. Legislation
Several bills designed to address general fraudulent or malevolent deepfake usage have been introduced in the U.S. Congress. None have passed—with the exception of S. 146 (the TAKE IT DOWN Act), the scope of which is limited to non-consensual intimate content—but all offer potential mitigating frameworks for fake endorsements beyond the potential constraints of older right of publicity laws. [38] H.R. 5586 (The DEEPFAKES Accountability Act) was introduced in the House of Representatives on September 20, 2023. The Act requires that any person who produces a piece of digital media containing a false impersonation “clearly identify such record as containing altered audio or visual elements, or as having been entirely created through generative artificial intelligence or similar technologies.” For media with both AI-generated videos and audios, creators must include a “clearly articulated verbal statement,” an “an unobscured written statement in clearly readable text appearing at the bottom of the image throughout the duration of the visual element” and “a link, icon, or similar tool” that reveals the content as having been artificially generated. [39] In advertising, such a policy would serve to minimize consumer confusion by clarifying that the endorsement is in fact fake.
Similarly, S. 1367 (The NO FAKES Act), introduced in the Senate on April 9, 2025, creates federal liability for entities that engage in the “display, distribution, transmission, or communication of, or the act of otherwise making available to the public, a digital replica without authorization by the applicable right holder.” [40] These acts are ambitious because in order to regulate deepfake content on a federal level, they must also implement federal rights of publicity; no such thing yet exists. [41] However, they present a clean, standardized alternative to the convoluted process of adapting and re-interpreting state-level publicity rights.
IV. ANALYSIS
Due to a lack of directly relevant precedent and the still-early development of current cases, investigation into the impact of AI on publicity rights necessitates inference and prediction based on rulings in similar, but older, cases. Another difficulty arises from the absence of a federal right of publicity law, and thus a uniform set of legal criteria applied to cases across state lines. However, past cases demonstrate a flexibility in the application of rights of publicity across decades and levels of technological development. The potential influence of AI is less about the right of publicity itself having to change, because courts have shown an ability to interpretively adapt statutes to new contexts in order to preserve the fundamental purpose of the right of publicity: protecting identity as a valuable commodity. Instead, the real challenge posed by widespread AI use in celebrity endorsements is the idea of widespread enforcement becoming more difficult. The adaptability of the law is less reassuring if the law cannot accomplish its fundamental purpose because it cannot be wielded.
AI usage in advertisement points to a likely future in which the publicity rights will be violated on a larger scale, because generative technology allows for the cheap and quick creation of celebrity endorsements. Companies have an obvious financial incentive to commission celebrity endorsements if they can afford them, and AI eliminates the cost barrier. Companies have a disincentive to violate publicity rights because of the costs of being sued, but this is where AI changes the calculus. Lawsuits are less of a threat if there are more instances of AI false endorsement in general, because even celebrities cannot afford to bring lawsuits against every offender. Tom Hanks, Oprah Winfrey, and several others put out clarifying statements about the falsity of AI impersonations rather than taking official legal action. Even using this route, there is no guarantee that victims of publicity rights violations can manage, across the board, to locate and debunk every attempt.
A solution to the potential difficulties in regulating publicity right violations through identification and enforcement may lie in reworking regulation at the legislative level. With the lack of oversight ability imposed by the nature of AI, policy solutions could mitigate some of these problems even just by bringing awareness to AI fraud and its consequences. Disclaimers like those required by the DEEPFAKES Accountability Act can tamp down consumer confusion in an age of rampant misinformation. The implementation of federally enforced consequences by bills like the NO FAKES Act could provide a nationwide protection for publicity rights—more needed now than ever as the commercial value of identity, and publicity rights in general, are under unprecedented threat. Ultimately, the adaptability of the law will mean little if enforcement becomes practically impossible. As AI accelerates the scale and anonymity of violations of publicity rights, meaningful protection will require a renewed regulatory framework to strengthen oversight.
Endnotes
[1] Better Business Bureau. “Scam Alert: Celebrity Impersonations Get More Sophisticated with AI Technology.” Better Business Bureau. August 8, 2025. https://www.bbb.org/article/scams/18549-scam-alert-celebrityimpersonations-get-more-sophisticated-with-ai-technology.
[2] Singleton, Tom. “Tom Hanks Warns Dental Plan Ad Image Is AI Fake.” BBC News. October 2, 2023. https://www.bbc.com/news/technology-66983194.
[3] Rozner, Lisa. “Complaints about Deepfake AI Videos More than Doubled This Year, FBI Says. Here Are Warnings from Experts.” CBS News. September 7, 2025. https://www.cbsnews.com/newyork/news/deepfake-artificialintelligence-videos.
[4] Better Business Bureau, “Scam Alert: Celebrity Impersonations.”
[5] Basiouny, Angie. “The Marketing Psychology Behind Celebrity Endorsements.” Knowledge at Wharton. May 30, 2023. https://knowledge.wharton.upenn.edu/article/the-marketing-psychology-behind-celebrity-endorsements/.
[6] Elberse, Anita, and Jeroen Verleun. “The Economic Value of Celebrity Endorsements.” Harvard Business School. 2012. https://www.hbs.edu/faculty/Pages/item.aspx?num=40853.
[7] Gilbert, Sophie. “How Much Do Celebrities Make From Fashion Deals?” The New York Times. June 26, 2025. https://www.nytimes.com/2025/06/26/t-magazine/fashion-brands-celebrity-endorsement-pay.html.
[8] Rothman, Jennifer E. “The Law.” Rothman’s Roadmap to the Right of Publicity. Accessed November 24, 2024. https://rightofpublicityroadmap.com/law/.
[9] Quinn, Julia. “Something Like .00086 of the World Is Famous.” The Atlantic. January 2013. https://www.theatlantic.com/technology/archive/2013/01/something-like-00086-of-the-world-is-famous/267397/.
[10] Vick, Kevin L., and Jean‑Paul Jassy. “Why a Federal Right of Publicity Statute Is Necessary.” Communications Lawyer (American Bar Association). August 2011. https://www.americanbar.org/content/dam/aba/publications/communications_lawyer/august2011/why_federal_right_ publicity_statute_is_necessary_comm_law_28_2.authcheckdam.pdf.
[11] Legal Information Institute. “Publicity.” Wex, Cornell Law School. https://www.law.cornell.edu/wex/Publicity.
[12] Privacy Act of 1974, 5 U.S.C. § 552a.
[13] Warren, Samuel D., and Louis D. Brandeis. “The Right to Privacy.” Harvard Law Review, December 15, 1890. MIT CSAIL. https://groups.csail.mit.edu/mac/classes/6.805/articles/privacy/Privacy_brand_warr2.html.
[14] Roberson v. Rochester Folding Box Co., 171 N.Y. 538, 64 N.E. 442, (N.Y. 1902)
[15] N.Y. Civ. Rights Law § 51
[16] Haelan Labs., Inc. v. Topps Chewing Gum, Inc., 202 F.2d 866 (U.S. App. 1953)
[17] Haelan Labs, 202 F.2d 866
[18] Vick and Jassy, “Why a Federal Right of Publicity Statute Is Necessary.”
[19] 15 U.S.C. §§ 1051–1127 (2023).
[20] 15 U.S.C. § 1125 (2023).
[21] Waits v. Frito-Lay, Inc., 978 F.2d 1093 (U.S. App. 1992)
[22] Young, Susan. “Getting the Message: How the Internet Is Changing Advertising.” Harvard Business School Working Knowledge. May 16, 2000. https://www.library.hbs.edu/working-knowledge/getting-the-message-how-theinternet-is-changing-advertising.
[23] Lohan v Take-Two Interactive Software, Inc., 31 N.Y.3d 111 (N.Y. 2018).
[24] Lohan, Inc., 31 N.Y.3d 111
[25] Lohan, Inc., 31 N.Y.3d 111
[26] Lohan, Inc., 31 N.Y.3d 111
[27] Kramer, et al. “AI-generated images of familiar faces are indistinguishable from real photographs.” National Library of Medicine. October 14, 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12521686/.
[28] Metz, Cade. “Instant Videos Could Represent the Next Leap in A.I. Technology.” The New York Times. April 4, 2023. https://www.nytimes.com/2023/04/04/technology/runway-ai-videos.html.
[29] Thompson, Stuart. “Can You Spot the Fake Video? Take the Test.” The New York Times. September 9, 2024. https://www.nytimes.com/interactive/2024/09/09/technology/ai‑video‑deepfake‑runway‑kling‑quiz.html.
[30] Heaven, Will Douglas. “How Do AI Models Generate Videos?” MIT Technology Review. September 12, 2025. https://www.technologyreview.com/2025/09/12/1123562/how-do-ai-models-generate-videos/.
[31] Vazquez Llorente, Raquel; Gildea, Ross; Anlen, Shirin. “Audiovisual Generative AI and Conflict Resolution: Trends, Threats and Mitigation Strategies.” WITNESS. September 12, 2024. https://www.gen-ai.witness.org/wpcontent/uploads/2024/08/WITNESS-Report_Audiovisual_Generative_AI_and_Conflict-1.pdf
[32] Lehrman v. Lovo, Inc., 790 F. Supp. 3d 348 (U.S. Dist. 2025)
[33] Lehrman, 790 F. Supp. 3d 348
[34] Lehrman, 790 F. Supp. 3d 348
[35] Lehrman, 790 F. Supp. 3d 348
[36] The Independent. “Scarlett Johansson Takes Legal Action Against AI App That Used Her Likeness ‘Without Permission’.” The Independent. November 2, 2023. https://www.the-independent.com/artsentertainment/films/news/scarlett-johansson-ai-lawsuit-ad-b2440034.html
[37] Lohan v Take-Two Interactive Software, Inc., 31 N.Y.3d 111 (N.Y. 2018).
[38] S. 146, 119th Cong. (2025).
[39] H.R. 5586, 118th Cong. (2023).
[40] S. 1367, 119th Cong. (2025).
[41] Salazar, Maria Elvira. “Congresswoman Salazar Introduces the NO FAKES Act.” U.S. House of Representatives. April 11, 2025. https://salazar.house.gov/media/press‑releases/congresswoman‑salazar‑introduces‑no‑fakes‑act‑0.