Eleven women sued three New York City strip clubs over the same basic insult. Their photographs, taken for lingerie catalogs and calendars, had been dropped into club advertisements on websites and social media. Nobody asked them. Nobody paid them. They said they would never have agreed. When the Second Circuit finished with the federal claim in Electra v. 59 Murray Enterprises, Inc., 987 F.3d 233 (2d Cir. 2021), one plaintiff had won it. The winner was Carmen Electra. The other ten had their Lanham Act claims thrown out, because the record did not show that the people looking at those ads knew who they were.

Even Electra's win was thin. The district court declined to award her damages, fees, or costs on the claim, and her lawyers waived the point on appeal by barely arguing it.

That result surprises almost everyone the first time they hear it, including lawyers who do a lot of trademark work. It is also the single best way to understand false endorsement. The claim feels like a right to control your own face, but the statute asks whether consumers are being deceived about who stands behind a product, and a deception about you requires an audience that has some idea who you are.

The short answer

How a trademark statute came to protect people

Section 43(a) was not written with Woody Allen in mind. As enacted in 1946, it reached false designations of origin and false descriptions of goods and services. Federal courts gradually read it to cover false claims of authorization. Geisel v. Poynter Products, Inc., 283 F. Supp. 261 (S.D.N.Y. 1968), let Dr. Seuss challenge dolls sold under his pseudonym, and a Cher case against a magazine that implied she read and endorsed it followed in 1982.

The theory arrived in mature form in 1985. National Video ran an ad for its V.I.P. card featuring a man with a high forehead, tousled hair, and heavy black glasses, leaning on a counter with cassettes of Annie Hall and Bananas beside him. The clerk gasped at him. He was a professional Woody Allen lookalike named Boroff, and in one version of the ad a tiny disclaimer said "Celebrity double provided by Ron Smith's Celebrity Look-Alike's." Chief Judge Motley's opinion in Allen v. National Video, Inc., 610 F. Supp. 612 (S.D.N.Y. 1985) opens by paraphrasing Groucho Marx: the plaintiff wouldn't belong to any video club that would have him as a member.

Her reasoning is still the foundation. She started with Justice Frankfurter's line from Mishawaka Rubber that trademark law protects the "commercial magnetism" of a symbol, and reasoned that a celebrity has a similar commercial investment in the "drawing power" of a name and face. The celebrity's stake and the public's interest in not being deceived point the same way. She then did something clever. New York's privacy statute requires a "portrait or picture," and she doubted that a photograph of a different human being qualified as a matter of law. The Lanham Act has no such limit. Evocation of a persona that would not satisfy the New York statute can still create confusion. She applied the ordinary trademark confusion factors, treated Allen's persona as the mark and Boroff's resemblance as the similarity between marks, and granted summary judgment on the Lanham Act claim.

Congress then rewrote the statute. The Trademark Law Revision Act of 1988 split Section 43(a) into its present two prongs. Subsection (A) now prohibits a use of "any word, term, name, symbol, or device" likely to cause confusion "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 by another person." Subsection (B), 15 U.S.C. § 1125(a)(1)(B), covers false advertising. The Supreme Court in Lexmark International, Inc. v. Static Control Components, Inc., 572 U.S. 118 (2014) calls these "two distinct bases of liability." In Waits v. Frito-Lay, Inc., 978 F.2d 1093 (9th Cir. 1992), the Ninth Circuit read the 1988 language and its legislative history as codifying the earlier cases and held, as a matter of first impression in that circuit, that false endorsement claims, "including those premised on the unauthorized imitation of an entertainer's distinctive voice," are cognizable.

This is why the vocabulary is odd. Courts describe a person as having a "trademark-like interest" in a persona, as Jackson v. Odenat, 9 F. Supp. 3d 342 (S.D.N.Y. 2014) and others put it, and they call that persona the mark. In Souza v. Exotic Island Enterprises, Inc., 68 F.4th 99 (2d Cir. 2023), the Second Circuit put it flatly: "In a false endorsement case like this one, the 'mark' in question is the identity of the purported endorser herself." Nobody has to use their face on a product or register anything. The rights arise from the goodwill attached to the identity.

For the wider history of identity rights, including the 1953 baseball card case that named the right of publicity, see The Right of Publicity: Name, Image, Likeness, Digital Replicas, and the Limits of Owning Yourself.

False endorsement and the right of publicity are cousins, not twins

Plaintiffs almost always plead both, and sloppy opinions occasionally blur them. They protect different things. The right of publicity, recognized in Haelan Laboratories, Inc. v. Topps Chewing Gum, Inc., 202 F.2d 866 (2d Cir. 1953) and given Supreme Court attention in Zacchini v. Scripps-Howard Broadcasting Co., 433 U.S. 562 (1977), protects the value of identity itself. Its defendant has taken something. False endorsement protects consumers and the goodwill that makes them trust an endorser. Its defendant has told a lie, or something that functions like one.

The practical differences are large enough that they should shape a case from the first day.

Question False endorsement, § 43(a)(1)(A) State right of publicity
Source of law Federal statute, uniform text, circuit variations in the tests State statute or common law; some states have little or nothing
Core wrong Likely consumer confusion about sponsorship or approval Commercial use of identity without consent
Does anyone need to recognize the plaintiff? Usually, among the targeted audience Often not. A private person can sue under many statutes, though damages may be modest
Does a disclaimer help? Possibly, because it can reduce confusion. A tiny one will not Generally no, because consent is the issue
Must the defendant's state of mind matter? Not for liability. It is one confusion factor and bears heavily on money Depends on the statute. Cal. Civ. Code § 3344 says "knowingly"
Geographic reach Nationwide wherever the use is in commerce Often limited. N.Y. Civ. Rights Law § 51 requires use "within this state"
After death Survives through successors if confusion is shown. Fifty-Six Hope Road Music, Ltd. v. A.V.E.L.A., Inc., 778 F.3d 1059 (9th Cir. 2015) Depends on the state. Some recognize long post-mortem terms; others none
Limitations period None in the statute; laches, usually measured against a borrowed state period Often short. New York's is one year with a single publication rule
Federal forum Yes, and it carries state claims under 28 U.S.C. § 1367 Only through diversity or a companion federal claim
Remedies Profits, damages, costs, exceptional-case fees, injunction Statutory damages in some states, fees in some, punitive damages in some

Three of these rows deserve a sentence more.

Recognition first. Souza says that where any attractive face will do, the unauthorized use "may invade rights granted by other statutes or common law sources" but "creates no risk of consumer confusion as conceived under the Lanham Act." In other words, the Second Circuit knows it is leaving a wrong without a federal remedy, and it is doing so deliberately.

Then time. New York's one-year period under C.P.L.R. § 215(3) sank most of the state claims in both Electra and Souza. The Souza plaintiffs argued that the short period applied only to privacy and not to publicity, and lost, because in New York publicity is encompassed in the statutory privacy right. A plaintiff who waits eighteen months may therefore have a state wrong with no state remedy and a federal remedy with no recognizable plaintiff. And once the federal claim falls, the district court can decline supplemental jurisdiction over whatever is left, as Souza affirmed.

Finally, consent. Under New York law, Electra held that a sweeping release given to a catalog company did not protect strangers to that release. The clubs were not third-party beneficiaries, and their contractors never obtained a license. Written consent in favor of one party does not allow others to use the image for advertising, however broadly it reads. That part of the case is a substantial win for models, and it explains why six plaintiffs returned to the district court with live state claims after the Lanham Act claim was gone.

For the state-law side in depth, see The Right of Publicity and the First Amendment: Transformative Use, Newsworthiness, and the Limits of Owning a Persona and Right of Publicity Basics. For the post-death question, see The Post-Mortem Right of Publicity: Who Controls a Deceased Person's Name, Image, and Voice.

The elements, and how the circuits test confusion

The statutory formula is simple enough. In Burck v. Mars, Inc., 571 F. Supp. 2d 446 (S.D.N.Y. 2008), a Southern District of New York judge listed the elements this way: the defendant, in commerce, made a false or misleading representation of fact, in connection with goods or services, that is likely to cause consumer confusion as to origin, sponsorship, or approval. Facenda v. N.F.L. Films, Inc., 542 F.3d 1007 (3d Cir. 2008) frames it in the language of trademark ownership: a legally protectable mark, ownership of it, and a use likely to create confusion about sponsorship or approval. The plaintiff bears the burden throughout.

The action is in the confusion factors. Each circuit started with its ordinary trademark test and adjusted it, sometimes uneasily, because those tests assume two companies with two products.

Circuit Framework What it does with fame Notable feature
Second Eight Polaroid factors Strength of the mark means recognizability among the relevant consumers Electra and Souza treated strength, actual confusion, and bad faith as dispositive on near-identical facts, but Souza warns district courts to address all factors in ordinary cases and reviews the balance de novo
Third Ten Lapp factors, restated in Facenda as eight modified Downing factors Level of recognition among the segment for whom the product is intended Adds how long the defendant used the work before any confusion appeared
Ninth Sleekcraft as tailored in White v. Samsung Electronics America, Inc., 971 F.2d 1395 (9th Cir. 1992) and Downing v. Abercrombie & Fitch, 265 F.3d 994 (9th Cir. 2001) Level of recognition among the targeted segment; "goods" means the reasons for the plaintiff's fame Wendt v. Host International, Inc., 125 F.3d 806 (9th Cir. 1997) calls likelihood of confusion predominantly factual, which makes summary judgment harder
Sixth Trademark factors, with ETW and Parks v. LaFace Records, 329 F.3d 437 (6th Cir. 2003) applying Rogers to art and music Recognizes false endorsement but rejects a general trademark in all images of a person Skeptical that a person's image functions as a mark as a general rule

The Downing list is the one most lawyers carry around. It asks about (1) the plaintiff's level of recognition among the segment of society for whom the defendant's product is intended, (2) the relatedness of the plaintiff's fame to the product, (3) the similarity of the likeness used, (4) actual confusion, (5) marketing channels, (6) likely purchaser care, (7) the defendant's intent in selecting the plaintiff, and (8) likelihood of expansion of the product lines. Courts admit that the eighth factor rarely means much here. White said it did not appear apposite at all.

The factors that do the work are recognition, similarity, intent, and evidence of confusion. Purchaser care often favors the plaintiff because nobody investigates endorsement before buying a beer. Wendt made that point about an airport bar where a customer might buy a single drink.

Two procedural notes matter for practitioners. In the Second Circuit, Souza confirmed that likelihood of confusion is reviewed de novo as a question of law and that older language about deference should not be read to expand a district court's discretion on summary judgment. In the Ninth, Wendt and Downing reversed summary judgments because a jury could reasonably find for the plaintiff on most factors. A lawyer choosing a forum for a marginal recognition case should notice that difference. Also, under Facenda, the plaintiff need not satisfy the false advertising distinction between literal and implied falsity. Likelihood of confusion is the standard under subsection (A), and the Third Circuit declined to muddle the two bodies of law.

For the factors outside the endorsement setting, see Trademark Likelihood-of-Confusion Analysis: A Practical Checklist.

The founding cases, read closely

Waits: the voice that said no

Tom Waits does not do commercials, and he had said so publicly for years. In 1988 the Tracy-Locke agency, developing a campaign for SalsaRio Doritos, liked his song "Step Right Up," itself a parody of commercial hucksterism. It auditioned singers until it found Stephen Carter, a Dallas musician and fan who had spent a decade perfecting a Waits imitation. The commercial's musical director warned Carter that he sounded so much like Waits it could pose legal problems. The agency's executive producer tried to get him to back off the imitation, but the client and the creative team did not like the result. The ad ran on the radio with the imitation intact.

The jury awarded $375,000 in compensatory damages and $2 million in punitive damages on California voice misappropriation, under the doctrine of Midler v. Ford Motor Co., 849 F.2d 460 (9th Cir. 1988), plus $100,000 under the Lanham Act. The Ninth Circuit affirmed on the state claim and held that false endorsement was cognizable. It also rejected the argument that only a competitor could sue, reasoning that the wrongful appropriator competes with the celebrity for the use of the celebrity's identity. Then it vacated the $100,000 as duplicative, because the jury had been told it could award the fair market value of his services on both claims. Fees survived, because the jury's finding of oppression, fraud, or malice made the case exceptional.

Two lessons remain. A plaintiff who has publicly refused to endorse anything is in a strong position, because the imitation falsely conveys a change of heart. And the Lanham Act damages may disappear into the state award, which is one reason to think hard about what each claim is for.

White: the robot

Samsung ran a series of ads in the voice of a future in which its products still existed. One showed a robot in a wig, gown, and jewelry beside a game board, turning letters, over the caption "Longest-running game show. 2012 A.D." Vanna White sued. The Ninth Circuit reversed summary judgment on her Lanham Act claim in White v. Samsung Electronics America, Inc., 971 F.2d 1395 (9th Cir. 1992), walking through the Sleekcraft factors. It said that "mark" means the celebrity's persona and "strength" means the level of recognition she enjoys, and added the sentence that recognizability cases now cite: if she is unknown to the segment of the public at whom the ad was directed, that segment could not be confused about whether she endorsed the product.

The court noted that the ad was a spoof and held that this did not end the inquiry. The series had included other celebrities who had been paid. Looking at the series as a whole, a jury could conclude that beneath the humor lay an intent to suggest that White, like them, was endorsing Samsung. Judge Kozinski's dissent from denial of rehearing en banc, which warned against overprotecting intellectual property, is still the best critique ever written of the persona theory, and every defendant in this field should read it for its rhetorical value. It did not change the result.

Wendt: the robots at the bar

The actors who played Norm and Cliff on Cheers sued over animatronic figures placed in airport bars built on the show's set. The district court had looked at the robots, looked at the actors, and found no similarity at all except that one figure, like one actor, was heavier than the other. The Ninth Circuit reversed in Wendt v. Host International, Inc., 125 F.3d 806 (9th Cir. 1997). A trial judge cannot stop at the similarity factor. People had come up to the actors saying some variation of "Hey George, I just had a drink with you in Kansas City," and the survey the district court had dismissed as "not a good survey" had to be reconsidered. The actors conceded they held no rights in the characters Norm and Cliff, and the defendants argued that this should end things. It did not, because the claim concerned the actors' own likenesses.

Abdul-Jabbar: the name he had stopped using

During the 1993 NCAA men's basketball tournament, an Oldsmobile ad asked who held the record for being voted the tournament's most outstanding player, answered "Lew Alcindor, UCLA, '67, '68, '69," and then boasted that the Eighty-Eight had made a consumer magazine's best buy list three years in a row. It closed with "A Definite First Round Pick." General Motors argued that Kareem Abdul-Jabbar had abandoned his birth name. Abdul-Jabbar v. General Motors Corp., 85 F.3d 407 (9th Cir. 1996) rejected that with a line worth remembering. One's birth name "is an integral part of one's identity" and is not kept alive through commercial use. The court also held that a jury could find implied endorsement in the commercial, because celebrity endorsement in television advertising is so established by custom that many viewers assume it. A newspaper poll does not carry the same implication. Had GM stopped at the trivia, it could probably have defended the reference as nominative fair use. It went further and used the star's record to make a claim for the car.

Burck: the blue cowboy

The Naked Cowboy of Times Square wears boots, a hat, and underpants, and carries a guitar. Mars ran an animated blue M&M dressed exactly the same way on video billboards in Times Square. In Burck v. Mars, Inc., 571 F. Supp. 2d 446 (S.D.N.Y. 2008), the court dismissed the New York privacy claim, because the statute protects the name, portrait, or picture of a living person and a cartoon of a costume is none of these. The false endorsement claim survived. Parody may defeat confusion when it is clear enough, the court said, but whether this one was a parody and whether it was strong enough to negate confusion raised factual questions. The commercial setting mattered.

Oliveira and the limit on sound

There is a counterweight, and plaintiffs often overlook it. Astrud Gilberto sang "The Girl from Ipanema," and Frito-Lay used the actual recording, under license from the rights holders, in a chip commercial. She claimed a trademark in her signature performance. In Oliveira v. Frito-Lay, Inc., 251 F.3d 56 (2d Cir. 2001), the Second Circuit affirmed dismissal of the Lanham Act claim. It could imagine such a right but refused to recognize a previously unknown one, because doing so would expose people who had paid every known rights holder to immense unforeseen liabilities. It did vacate the judgment on her state claims so that she could refile them in state court. Pirone v. MacMillan, Inc., 894 F.2d 579 (2d Cir. 1990) is the sibling case for images. Photographs of Babe Ruth in a baseball calendar, the court said, indicate the contents of the calendar and not its source, and a photograph of a human being is not inherently distinctive.

Read Waits and Oliveira together and a useful rule of thumb appears. Imitating someone to borrow their reputation looks like false endorsement. Licensing a performance they gave, and using it as content, usually needs something more.

Recognizability: the model cases

The strip club cases are a strange body of law. A single plaintiffs' firm has filed many of them, often for overlapping groups of models. Souza lists a long footnote of examples across the country. They have produced most of the modern law on recognition, because they present the issue in its purest form. The ad contains a photograph and no name.

What Electra decided

The advertisements in Electra did not name the models. The clubs' contractors said they took pictures from a catalog supplied by a costume company and believed they had rights. The clubs never asked for a specific person. The court approved the district court's decision to analyze each plaintiff's public prominence as the strength of her mark, for a reason that goes to the heart of the matter: the ads provided no information identifying the plaintiffs other than their pictures. It relied on Bondar v. LASplash Cosmetics, No. 12 Civ. 1417, 2012 WL 6150859 (S.D.N.Y. Dec. 11, 2012), for the idea that misappropriating a completely anonymous face could not support a false endorsement claim, because consumers would not infer that an unknown model was endorsing a product as opposed to lending her image for a fee.

It also affirmed exclusion of the plaintiffs' survey, which asked about only a small fraction of the images and gave respondents no way to indicate they did not know. And it found no bad faith, since the record showed at most a failure to investigate whether a contractor had rights. The clubs did not gain profits from the ads on this record, and the plaintiffs' incomes had not fallen.

What Souza added

Two years later the same arguments came back with some of the same plaintiffs. The record in Souza v. Exotic Island Enterprises, Inc., 68 F.4th 99 (2d Cir. 2023) included Playboy Playmates of the Month, one Playmate of the Year, peak annual modeling earnings from around $18,300 to around $107,000, and social media followings from several thousand to a few million. The court still found no strong mark. The plaintiffs' declarations that they had "achieved celebrity status" were vague, conclusory, and identical.

The opinion is the most useful thing written on the subject, for four reasons.

First, it explains why strength means recognizability here. Inherent distinctiveness is awkward when applied to how much one person resembles another, and a face falls nowhere on the familiar spectrum from arbitrary to generic. Their identity, the court said, inherently is their mark.

Second, it gives a clear account of what a survey must show. The excluded survey of 812 strip club patrons reported that almost half felt they recognized the plaintiffs in some manner. The court was unimpressed. There was no control group, the definition of recognition was too generous, and several plaintiffs' full faces were not shown. A person's sense that a face seems vaguely familiar says little about whether that person infers sponsorship by that particular individual. The confusion portion, which reported that 75 percent agreed all the women had agreed to sponsor the clubs, failed because respondents gave impressions of all images collectively and were never offered a "don't know" option or told not to guess.

Third, it separates two kinds of deception. Suppose patrons did think the women in the posts worked at the club. That may be a deceptive practice, the court said, but it is not a trademark injury, because the patron's expectation does not rest on the models' goodwill. The misconception "goes to the nature of the product itself, not the mistaken belief that someone whose imprimatur the consumer values has vouched for that product." That sentence should be taped to the wall of anyone drafting one of these complaints.

Fourth, it contains a limit. Footnote 10 cautions that the court does not hold that all false endorsement claims rise or fall on these three factors. The court called the circumstances exceptional because they were effectively identical to Electra.

Is recognition a threshold, a factor, or a matter of proof?

Formally it is a factor. Functionally, in the Second Circuit, it is close to a threshold, and courts elsewhere are divided on how early to apply it.

So the pleading stage may be survivable in some courts on allegations alone, and summary judgment is where the record matters. Plaintiffs should not take comfort from that. Evidence that does not exist at the complaint stage often does not exist later.

Recognition among whom?

Celebrity is the wrong word, and courts have said so. In Bondar, as quoted in Lovo, the statute "does not require celebrity, only a likelihood of consumer confusion," and a level of recognition short of celebrity can be enough. What counts is the audience the defendant is targeting. Three examples show the range.

In Downing v. Abercrombie & Fitch, 265 F.3d 994 (9th Cir. 2001), Abercrombie used a 1965 photograph of surfers in a catalog and sold matching shirts. A declaration from a surf historian called the plaintiffs legends in the surf community, and the Ninth Circuit reasoned that young customers buying surf wear would know them. It reversed summary judgment.

In Hauf v. Life Extension Foundation, 547 F. Supp. 2d 771 (W.D. Mich. 2008), a young cancer survivor's recovery story appeared in membership-drive materials from a nonprofit that sold him supplements. The court held that he need not prove he was a celebrity and might show notoriety among cancer patients or advocates of alternative medicine, the very people the seller was courting.

In Stayart v. Yahoo! Inc., 623 F.3d 436 (7th Cir. 2010), by contrast, a woman found her name in search results linked to pharmaceutical and pornographic sites. She lacked a commercial interest in her name, and her claim failed for that reason. This was a pre-Lexmark standing decision, but the instinct persists. Someone with no commercial reputation is unlikely to bring herself within a commercial statute.

The audience question also limits the plaintiff's own proof. A million Instagram followers spread across the world may say little about patrons of a bar in Newburgh. Souza noted how fleeting the plaintiffs' links to New York were. Counsel should think about the defendant's customers before commissioning anything.

Who may sue: Lexmark and its uncertain reach

Lexmark held that a false advertising plaintiff must come within the zone of interests the statute protects, meaning a commercial interest in reputation or sales, and must show injury proximately caused by the misrepresentation. It rejected the older tests for standing. A consumer who buys a disappointing product is out. The Ninth Circuit in Fifty-Six Hope Road treats Waits as abrogated by Lexmark on other grounds, which appears to refer to its standing discussion.

The open question is whether Lexmark governs subsection (A) at all. Lexmark itself involved only false advertising. The Fourth Circuit applied its framework to a false association claim in Belmora LLC v. Bayer Consumer Care AG, 819 F.3d 697 (4th Cir. 2016). The First Circuit, in Clemente Properties, Inc. v. Pierluisi-Urrutia, No. 23-1922 (1st Cir. Jan. 16, 2026), assumed without deciding that the test could apply because no one had challenged it, then added in a footnote that Lexmark may well have intended a test for false advertising only. It also corrected a common error. The test is injury to sales or business reputation, so a plaintiff who pleaded reputational harm to a licensing business did not need to allege lost sales. Souza analyzed Lexmark under its false advertising count and did not apply it to false endorsement.

The false advertising count in the model cases shows how demanding the inquiry is. The Souza plaintiffs claimed two injuries. The first was lost work from being linked to a strip club, which the court said would likely satisfy Lexmark if there had been any evidence for it. There was none. No one was shown to have seen the posts who might have hired them, and there was no temporal correlation, no expert, and no lost job. Interrogatory responses simply said clients never explain why they do not call. The second injury was the fee they should have been paid. That, the court said, sounds in trademark infringement and theft of services, and "fails to check any of Lexmark's boxes." It noted that other courts had gone the other way in a footnote, so the point is not universally settled.

A plaintiff's practical conclusion is to plead and prove a commercial interest in the identity, and reputational harm in a form a jury can see. For the false advertising side more broadly, see False Advertising Under the Lanham Act: Literal Falsity, Materiality, and Competitor Standing.

Dastar and the claims that are really about something else

Dastar Corp. v. Twentieth Century Fox Film Corp., 539 U.S. 23 (2003) held that "origin of goods" in Section 43(a)(1)(A) refers to the producer of tangible goods and not the author of ideas or communications embodied in them. Otherwise the Lanham Act would become "a species of perpetual patent and copyright." The holding concerned the word "origin," and false endorsement rests on "sponsorship, or approval." So Dastar does not abolish the claim. Its spirit does a good deal of work anyway.

Lovo quotes it for the proposition that the plaintiffs' theory would create a cause of action, in effect, for plagiarism. A defendant facing a claim that amounts to "you used my performance without paying me" should reach for it. Fifty-Six Hope Road shows the price of neglecting it. The defendants there waived Dastar, waived Rogers, and waived aesthetic functionality, and the Ninth Circuit affirmed a verdict over Bob Marley T-shirts while pointedly noting that the outcome was in part a function of those waivers. Judge Christen's separate opinion stressed the same point, calling the holding narrow and dictated by the standard of review and the defenses actually pursued.

A copyright is not a complete answer in the other direction. In Facenda v. N.F.L. Films, Inc., 542 F.3d 1007 (3d Cir. 2008), the NFL owned the copyright in narrations by the Philadelphia broadcaster John Facenda and reused small portions of them in a cable program about the video game Madden NFL 06. The Third Circuit held the program was commercial speech, more like an infomercial than a documentary. Facenda had signed a release allowing the NFL to use his recordings as it saw fit, "provided, however, such use does not constitute an endorsement of any product or service." The court sent the Lanham Act claim to trial, and it held that the release's carve-out meant the release would not bar the claim if the use was in fact an endorsement. It affirmed summary judgment for the estate on the Pennsylvania publicity claim and rejected copyright preemption of it. For how Section 301 sorts these claims generally, see Copyright Preemption Under Section 301: When Federal Law Wipes Out Your State Law Claim.

Roberts v. Bliss adds a useful boundary. The actress in a viral street-harassment video sued after TGI Friday's licensed the clip and superimposed giant appetizers over her image. In Roberts v. Bliss, 229 F. Supp. 3d 240 (S.D.N.Y. 2017), the court dismissed the false endorsement claim. A persona or distinctive movement might in principle qualify, but because of the superimposed food she was not pictured or represented in the ad at all, and she did not plausibly allege that her movements were so distinctive as to identify her. A person has to actually be there, in some recognizable sense, before anyone can be confused about her.

Expressive works: Rogers, and what Jack Daniel's left of it

The test

Ginger Rogers sued over Fellini's film Ginger and Fred, about two Italian cabaret performers who imitated her and Fred Astaire. In Rogers v. Grimaldi, 875 F.2d 994 (2d Cir. 1989), the Second Circuit held that the Lanham Act should reach artistic works only where the public interest in avoiding confusion outweighs the interest in free expression, and that this balance normally forbids applying the Act to a title unless the title has no artistic relevance to the work or explicitly misleads as to source or content. Survey evidence that some people inferred her involvement could not overcome that. Explicit misstatements would be different. The court's examples were "Jane Fonda's Workout Book" and the phrase "an authorized biography."

Courts spread the test beyond titles. Mattel, Inc. v. MCA Records, Inc., 296 F.3d 894 (9th Cir. 2002) used it for "Barbie Girl." ETW used it for a painting of Tiger Woods at the Masters and refused, in the same opinion, to give Woods trademark rights in every image of himself. E.S.S. Entertainment 2000, Inc. v. Rock Star Videos, Inc., 547 F.3d 1095 (9th Cir. 2008) used it for a virtual strip club in Grand Theft Auto, and said the level of artistic relevance merely must be above zero. Parks v. LaFace Records, 329 F.3d 437 (6th Cir. 2003) shows the test can bite. OutKast's song "Rosa Parks" used her name in a title for a song the composers said was not about her, with a chorus about moving to the back of the bus. The Sixth Circuit held that whether the title had artistic relevance could not be resolved as a matter of law. The court repeated the warning from Rogers that "[p]oetic license is not without limits."

The avatar cases

Madden NFL gives the cleanest comparison in intellectual property. Jim Brown claimed the game used his likeness as an unnamed avatar on historic teams. Brown v. Electronic Arts, Inc., 724 F.3d 1235 (9th Cir. 2013) applied Rogers, held that a great player's likeness had at least some relevance to a realistic football game, and found nothing explicitly misleading to consumers. A suggestion at an academic conference that players had signed licenses was the wrong audience. The claim was dismissed on the pleadings.

Other former players sued EA over the same game on a right of publicity theory. Davis v. Electronic Arts Inc., 775 F.3d 1172 (9th Cir. 2015) held that EA's First Amendment defenses failed, following Keller v. Electronic Arts Inc., 724 F.3d 1268 (9th Cir. 2013) and the Third Circuit's Hart v. Electronic Arts, Inc., 717 F.3d 141 (3d Cir. 2013), because the game put players in the very setting that made them famous, doing what they did to become famous. So the same avatars could produce a dead Lanham Act claim and a live state claim from the same court. Rogers asks whether consumers are explicitly misled. Transformative use, drawn from Comedy III Productions, Inc. v. Gary Saderup, Inc., 25 Cal. 4th 387 (2001), asks whether the identity was turned into something new. Those questions have nothing to do with each other.

The Supreme Court's narrowing

VIP sold a dog toy called Bad Spaniels, shaped like a Jack Daniel's bottle. The Ninth Circuit had applied Rogers because the toy was funny. A unanimous Supreme Court reversed. Rogers, the Court said, has always applied where a mark is used not to designate source but to perform some other expressive function. The Court held only that the test "is not appropriate when the accused infringer has used a trademark to designate the source of its own goods," meaning it "has used a trademark as a trademark." The Court took no position on whether Rogers is correct in its proper sphere. Confusion about source is the "cardinal sin" of trademark law, and the Court noted that a mark used as a source identifier does not get a First Amendment threshold test.

Later courts have followed the distinction in both directions. Punchbowl, Inc. v. AJ Press, LLC, 90 F.4th 1022 (9th Cir. 2024) held that a company using PUNCHBOWL as the name of its own media business could not use Rogers at all. Vans, Inc. v. MSCHF Product Studio, Inc., 88 F.4th 125 (2d Cir. 2023) reached a similar result on a shoe. But the case that matters most here is Hara v. Netflix, Inc., No. 23-3768 (9th Cir. July 28, 2025).

Hara and the survival of Rogers in endorsement cases

The drag performer Vicky Vox claimed an animated version of her appeared for ten seconds in Netflix's Q-Force, as an unspeaking character who snaps open a fan, as well as in a teaser and a still. Friends and fans contacted her, confused. She argued that Jack Daniel's removed Rogers for any use likely to confuse about endorsement, since the same statute covers both source and endorsement.

The Ninth Circuit disagreed. Jack Daniel's, it held, makes source designation the thing that defeats Rogers, "not other uses such as affiliation or implied endorsement." Rogers itself was an endorsement case. The use of her likeness helped ground a West Hollywood bar in realism, in the same way Jim Brown helped Madden. Under the test, the likeness was relevant and nothing was an explicit misstatement. The court also extended the protection to the teaser and still image, relying on Twentieth Century Fox Television v. Empire Distribution, Inc., 875 F.3d 1192 (9th Cir. 2017) for the idea that protected works may be advertised and marketed. Banner ads that YouTube automatically placed around the teaser did not turn it into an advertisement for those products.

Hara is one circuit's reading, though it is the most direct appellate authority on the point. Plaintiffs may argue elsewhere that a likeness placed prominently in marketing functions as a source signal. Expect the line to be fought over promotional uses, where the expressive work and the advertisement blur.

The ordinary advertisement is outside the shield

Nothing in this line helps a nightclub, a car dealer, or a supplement seller. Facenda refused to apply Rogers to a program it classified as commercial speech, using the factors of whether the speech is an ad, whether it refers to a specific product, and whether the speaker has an economic motive. Burck evaluated a parody in a billboard under ordinary confusion analysis. A helpful way to sort uses is by asking what the identity is doing.

What the identity is doing Example Likely framework
Serving as the defendant's brand or product name A beverage called "[Star]'s Choice" Ordinary confusion analysis after Jack Daniel's
Promoting an unrelated product or service A club's Instagram post, a car ad Ordinary confusion analysis
Appearing inside a film, song, or game for expressive reasons Hara, Brown, Rogers Rogers, in the circuits that use it
Promoting the expressive work it appears in A trailer or teaser Rogers per Hara and Empire, with some litigation risk
Decorating merchandise A T-shirt of a dead reggae star Contested. Confusion analysis if Rogers is waived or inapplicable, as in Fifty-Six Hope Road
Describing the person accurately Trivia, news, a doll "of" a princess Nominative fair use

For the expressive-work defense in more depth, see Trademark Fair Use: Descriptive Use, Nominative Use, and Expressive Works After Jack Daniel's and, from the clearing side, Clearing Expressive Works Against Right of Publicity Claims: A Practical Guide.

Nominative fair use

Sometimes the only way to talk about a person is to name her. Under New Kids on the Block v. News America Publishing, Inc., 971 F.2d 302 (9th Cir. 1992), a defendant in the Ninth Circuit gets the defense if the person or product cannot readily be identified without the mark, the use goes no further than necessary, and nothing suggests sponsorship or endorsement. Abdul-Jabbar lost on the third element because a car commercial carries an implied endorsement. In Cairns v. Franklin Mint Co., 292 F.3d 1139 (9th Cir. 2002), Franklin Mint won on all three for its Princess Diana plates and dolls. "Authentic" referred to an accurate replica of a dress, and "Certificate of Authenticity" referred to the plate. The Fund's false endorsement claim failed. The case also ended with a $2,308,000 fee award against the Fund on its other claims, which is a warning to anyone who thinks that a sympathetic plaintiff is immune from a fee motion.

The statutory classic fair use defense in 15 U.S.C. § 1115(b)(4) protects use of a term descriptively and in good faith, and it rarely fits identity. In the Ninth Circuit, nominative fair use replaces the confusion test instead of supplementing it, as Cairns explains. Other circuits handle it differently. See Asserting or Defeating a Trademark Fair Use Defense: A Practical Guide.

AI voices and digital replicas

What is actually happening

Current voice cloning works by training or tuning a speech model on recordings of a target speaker. The system extracts a mathematical representation of how that person sounds, including pitch, timbre, cadence, and accent, and combines it with a text-to-speech engine so that any typed script comes out in that voice. Sample requirements have fallen from hours of studio audio to, in some products, a short clip taken from a podcast or a video. Face and body replicas operate similarly. The output file usually carries no reliable label of its origin unless someone adds provenance metadata, and that can be stripped when the file is uploaded and re-encoded by a social platform.

This matters legally in three ways. Evidence of copying now lives in a training log, a voice ID, a prompt, or a vendor account, and not in a casting file. The person pressing the button may have no idea whose voice the model was built from. And the costs of producing a false endorsement have collapsed, which means defendants are often small, anonymous, or abroad.

Lovo: what the Lanham Act will not do

The first thorough opinion is Lehrman v. Lovo, Inc., No. 24-CV-3770 (JPO) (S.D.N.Y. July 10, 2025). Two voice actors alleged that Lovo, a text-to-speech company, recruited them through a freelance marketplace for what they were told were limited research uses, and then sold cloned versions of their voices under other names. Judge Oetken accepted that voices can qualify as a persona in principle and that celebrity is not required. Then he dismissed the Lanham Act claims on a different ground.

He reasoned that the celebrity endorsement cases "sit uneasily" with the distinction between marks and products, and treated them as something of an exception. These plaintiffs, unlike celebrities, sold their voices as services. Clients paid for tone, quality, timbre, and delivery. So the voices were protectable only to the extent that they functioned primarily as source identifiers, and the plaintiffs had not pleaded secondary meaning. Relying on Pirone, he said a voice that works as content in a recording is like Babe Ruth's picture on a calendar. "To be clear, Plaintiffs are not out of luck here because they are not famous enough." Even extremely famous people cannot assert the claim for use of a likeness as a product.

He also worried about the consequences. If the theory were right, a newcomer who happened to sound like an established voice actor could be sued, and since unregistered rights can be assigned, a company might end up owning someone's identity. The cases about imitating Woody Allen, he said, involve pretending to be him, and not merely working in the same trade while resembling him.

The rest of the opinion matters as much. The New York Civil Rights Law claims survived, and so did New York General Business Law claims, contract claims, and one plaintiff's copyright claim based on use of her actual recording. Copyright claims based on model training were dismissed with leave to amend. The false advertising theory was dismissed too, as a disguised version of the affiliation claim. The opinion is a district court ruling at the pleading stage, with leave to amend on part of it, so check its later history before relying on it.

Where the Lanham Act still fits

Lovo is about using a voice as a substitute product. It does not touch the fact pattern most people fear, in which a cloned voice speaks as a known person and urges you to buy something. That is Waits with cheaper equipment. The defendant is using the identity's goodwill, the ad is commercial speech, and there is no expressive shield. If anything, a clone is more similar to the plaintiff than a skilled impersonator would be, and the intent to evoke is often obvious from the prompt.

Disclosure language deserves a skeptical look. An ad that says "AI voice" in small print is doing what National Video's lookalike disclaimer did. Allen held National Video's tiny-print disclaimer insufficient as a matter of law, and said that to be effective it "would have to be bolder and make clear that plaintiff in no way endorses National, its products, or its services." The court's injunction let the lookalike keep working only in ads that made clear Allen had nothing to do with the project. A prominent, unmistakable disclosure may reduce confusion. It does nothing for a state claim that turns on consent.

The state statutes

Tennessee's ELVIS Act, effective in 2024, added voice to its protected attributes and created liability around tools whose primary purpose is producing unauthorized replicas. See Tenn. Code Ann. § 47-25-1105. California protects voice in Cal. Civ. Code § 3344, and Cal. Civ. Code § 3344.1 addresses replicas of deceased personalities in some audiovisual works. New York's N.Y. Civ. Rights Law § 51 names voice, and N.Y. Civ. Rights Law § 50-f covers deceased performers. New York has also enacted a disclosure statute for advertising. Under N.Y. S.8420-A (2025), signed in December 2025 and in force from June 2026, a person in the business of dealing in property or services who produces an ad must conspicuously disclose a synthetic performer, if the person has actual knowledge of it, with a civil penalty of $1,000 for a first violation and $5,000 for later ones. The statute defines a synthetic performer as a digitally created performer who is not recognizable as any identifiable real performer, and it does not apply to audio ads, so it reaches invented AI people and not clones of real ones. The act exempts ads for expressive works in defined circumstances and preserves rights under Sections 50, 50-f, and 51. For the details of this legislative wave, read Name, Image, Likeness, and Digital Replicas: The ELVIS Act, the NO FAKES Act, and Deepfake Law.

NO FAKES

The NO FAKES Act of 2026, S. 4591 and H.R. 8915 in the 119th Congress, would create a federal right in a person's voice and visual likeness against unauthorized digital replicas, with a notice-and-takedown system for online services. Earlier versions died in committee. The Senate Judiciary Committee advanced the current bill by voice vote on June 18, 2026, and it was placed on the Senate calendar. It has not become law, and the House companion was last reported to be waiting for action in committee. If enacted, it would fill the very gap Lovo identified, because a replica right would not depend on source identification or confusion. Practitioners should read the final text carefully, particularly on preemption and exceptions, instead of working from summaries. The U.S. Copyright Office, Copyright and Artificial Intelligence, Part 1: Digital Replicas (2024) recommended federal legislation along these general lines.

Platforms

Section 230 of the Communications Decency Act does not limit intellectual property law. See 47 U.S.C. § 230(e)(2). A Lanham Act claim against a platform is not barred on that basis, but it faces ordinary hurdles on use and contributory liability. Whether state publicity claims fall within the same exception is a real circuit split. Perfect 10, Inc. v. CCBill LLC, 488 F.3d 1102 (9th Cir. 2007) limits it to federal intellectual property, and Hepp v. Facebook, 14 F.4th 204 (3d Cir. 2021) applied it to a Pennsylvania right of publicity claim.

The FTC's parallel regime

False endorsement is a private suit. The FTC comes at the same conduct from the consumer's side, and its rules apply whether or not the person depicted is recognizable.

The Endorsement Guides, 16 C.F.R. Part 255, are interpretive guidance under Section 5 of the FTC Act, 15 U.S.C. § 45. They define an endorsement to include depictions of a person's name, signature, likeness, or other identifying personal characteristics. Endorsements must reflect the endorser's honest opinions. An advertiser may use a celebrity's endorsement only so long as it has good reason to believe the celebrity still holds the views. If the ad says the endorser uses the product, the endorser must have been a bona fide user. Section 255.1(g), added in the 2023 revision, says that using an endorsement with the image or likeness of someone other than the actual endorser is deceptive if it misrepresents a material attribute of the endorser. Intermediaries such as agencies can be liable as well.

The Rule on the Use of Consumer Reviews and Testimonials, 16 C.F.R. Part 465, has more teeth because it is a trade regulation rule. Section 465.2 prohibits a business from writing, creating, or selling a consumer or celebrity testimonial that materially misrepresents that the testimonialist exists, used the product, or had the stated experience, and prohibits disseminating one about its own products if it knew or should have known of the misrepresentation. "Celebrity testimonial" is defined by reference to depictions of a well-known individual's name, likeness, or identifying characteristics. A cloned celebrity voice saying "I take this every morning" fits the definition comfortably.

The Rule on Impersonation of Government and Businesses, 16 C.F.R. Part 461, does what the title says and no more. In 2024 the FTC proposed extending it to the impersonation of individuals, later dropped a proposed provision on means and instrumentalities, and scheduled a hearing. No final rule extending it to individuals has issued. A Federal Register notice published October 1, 2026 sought comment on platform issues related to impersonation scams. These rules generally do not give a private plaintiff a cause of action, but they are excellent context for a demand letter and strong evidence of how regulators read consumer perception. For the compliance side, use Sponsorship, Endorsement, and Influencer Agreements: A Practical Guide and Consumer Marketing Compliance Toolkit: TCPA, CAN-SPAM, Auto-Renewal, and Endorsements.

Remedies: what winning is worth

Section 35 of the Lanham Act, 15 U.S.C. § 1117(a), allows recovery of the defendant's profits, damages sustained, and costs, subject to the principles of equity. The plaintiff need only prove sales, and the defendant must prove deductions. The court may adjust the award, but it must be compensation and not a penalty. Fees are available in exceptional cases to the prevailing party, which can be either side.

Profits. Romag Fasteners, Inc. v. Fossil, Inc., 590 U.S. 212 (2020) held that willfulness is not a precondition to profits for a Section 43(a) violation, though the defendant's mental state remains "a highly important consideration." Fifty-Six Hope Road affirmed the district court's profits awards against three defendants, held that no jury right attached to the calculation of profits, and found willfulness in a licensing executive who was told before sales began that the defendants had no rights. The difficulty in advertising cases is causation. How much of a club's revenue is attributable to an Instagram post? In Electra, the record showed no additional profits.

Damages. District courts in the Second Circuit have treated intent to deceive as a prerequisite to money damages in a false endorsement case, at least in decisions that predate Romag. In Beastie Boys v. Monster Energy Co., 66 F. Supp. 3d 424 (S.D.N.Y. 2014), the jury awarded $500,000 on the Lanham Act claim, and the court upheld it on circumstantial evidence. The Monster employee who made the video knew the group had not endorsed Monster, yet produced four minutes of promotion suffused with its songs. Fair market value of the lost endorsement is the natural measure, but it must be proved. In Electra, the district court struck a damages expert as speculative and methodologically unreliable, and later courts describe that report as having overstated the plaintiffs' payment history and multiplied damages without justification. Beware duplication as well. Waits lost $100,000 that way.

Injunctions. A plaintiff who establishes a violation, or likely success, gets a rebuttable presumption of irreparable harm under 15 U.S.C. § 1116(a), as restored by the Trademark Modernization Act. Delay can rebut it, which is a problem for people who find a post two years late. See Trademark Preliminary Injunctions After the Trademark Modernization Act: The Presumption of Irreparable Harm and the Delay Trap.

Fees. Waits tied the exceptional case standard to malice, and many circuits now use the more flexible test from Octane Fitness, LLC v. ICON Health & Fitness, Inc., 572 U.S. 545 (2014). See Lanham Act Attorney's Fees Under 15 U.S.C. § 1117(a): A Practical Guide.

Who pays. Liability can reach beyond the entity. Edmondson v. Velvet Lifestyles, LLC, 43 F.4th 1153 (11th Cir. 2022) states that a corporate officer who directs, controls, ratifies, participates in, or is the moving force behind the infringing activity is personally liable, though the court also reversed summary judgment against an individual and an affiliate where the plaintiffs had simply lumped defendants together. Blaming a marketing vendor may help on bad faith, as it did in Electra, without eliminating the claim. In Souza the club brought its social media vendor in as a third-party defendant.

Insurance: the surprise exclusion

A business sued for one of these claims should look at Coverage B of its commercial general liability policy immediately. Standard "personal and advertising injury" definitions can reach publication of material that violates a person's right of privacy, use of another's advertising idea, and infringing a copyright, trade dress, or slogan in an advertisement. Whether a model's photograph fits is contested, and insurers have been adding exclusions aimed directly at this business.

In Princeton Excess & Surplus Lines Insurance Co. v. A.H.D. Houston, Inc., 84 F.4th 274 (5th Cir. 2023), applying Texas law, the Fifth Circuit reversed a judgment requiring an insurer to cover clubs facing a $1,405,000 state court judgment. It disagreed that the models' images were an "advertising idea" and held that an entertainment exclusion left enough coverage in place to avoid being illusory. Judge Ho dissented in favor of certifying the question. Nobriga v. Clear Blue Specialty Insurance Co., No. 3:24-cv-01980 (D. Conn. July 10, 2026) enforced a similar "exhibitions and related marketing" exclusion under Connecticut law and expressly declined to follow Princeton Excess & Surplus Lines Insurance Co. v. R.I. Cranston Entertainment Inc., 725 F. Supp. 3d 184 (D.R.I. 2024), which had come out the other way. A Florida federal court reached a contrary result as well in Princeton Express v. DM Ventures USA LLC, 209 F. Supp. 3d 1252 (S.D. Fla. 2016). The result is a split that depends on state law and policy language, so read the actual form and tender quickly. See Business Insurance and Coverage Disputes: CGL, E&O, Cyber, and D&O.

Three worked hypotheticals

Hypothetical one: the Instagram post in Spokane

Suppose a Spokane nightclub called the Lantern Room posts a photo of Noor Halvorsen, a swimwear model from Miami with 40,000 followers, captioned "Saturday: bottle service with our girls." The photo came from a Google search by the club's part-time social media manager. Noor has never been to Washington.

Under Section 43(a), her problem is plain. Does anyone in the Lantern Room's audience know who she is? If she can show only follower counts and a modest modeling career, a Ninth Circuit court applying White and Downing would likely find a weak mark, though the Ninth Circuit's friendliness to juries gives her somewhat more room than Souza would. The caption "our girls" looks like a claim that she works there, which Souza calls a nature-of-the-product deception and not a goodwill problem. She should look to Washington's personality rights statute, Wash. Rev. Code ch. 63.60, and to the photographer, who may hold a copyright claim of his own. The club's owner should take the post down, check its Coverage B, and find out how the manager picks images. See also Image Copyright Demand Letters: How Photo Licensing Enforcement Works and What Claims Are Really Worth.

Change one fact. Noor is the current face of a national energy drink, and the Lantern Room targets the same crowd. Now the first factor is strong, the ad is unprotected, and the case looks a lot like Carmen Electra's, with a better damages expert.

Hypothetical two: the sound-alike in a truck ad

Suppose Cascadia Auto Group hires an agency to promote its truck sale in the Pacific Northwest. The creative director wants the sound of Marlo Venn, a regionally beloved country singer who has declined every commercial offer for twenty years. After Venn's manager says no, the agency hires an impersonator and writes a jingle echoing her best-known song. The song's melody is licensed from the publisher.

This is Midler and Waits. The refusal is the most damaging document in the file, because it establishes knowledge and shows the imitation was a substitute for a declined endorsement. Regional recognition is enough where the ad runs regionally. The publisher's license answers the copyright question and nothing else. Cascadia may well face a state law claim that dwarfs the Lanham Act number, so the parties should think about duplication early. Would a line at the end, "Not an actual Marlo Venn performance," help? It might reduce confusion among those who hear it, in a medium where few people listen to the end. It would not cure the state claim. Allen suggests a court would want something much bolder, spelled out, and hard to miss.

Hypothetical three: the AI doctor who sells supplements

Suppose a Facebook ad shows a local Tulsa television physician, Dr. Priya Castellane, apparently recommending a sleep supplement called Quellwell. The video is a real clip of her newscast with her lips and voice regenerated by AI. The seller is a shell company using a virtual address. She sells nothing but has a paid speaking business and a television contract.

Her Lanham Act claim is strong on the merits. She is recognized in the audience the ad reaches, the statement is an express false endorsement, and she has a commercial interest under any reading of Lexmark. Hauf suggests that a community-level reputation can carry the claim. Lovo does not get in her way, because the voice is being used to invoke her and not as a generic narration product. The FTC testimonial rule is implicated too, if she counts as well-known and because she never used the product.

Her real problem is finding someone to sue. She should capture the ad library entry and landing page, use a platform complaint process, consider early discovery against the platform and payment processor, and notify her station, which may own the source footage and hold a copyright claim suited to a fast takedown. A judgment against an empty shell has limited value. The most valuable outcome may be a fast removal and a paper trail. Scam-style ads like this also connect to the domain tools covered in Lookalike Domains and Typosquatting: Phishing, Invoice Fraud, and the Legal Tools That Take Them Down.

Unsettled questions worth watching

Several issues remain truly open. Whether Lexmark governs subsection (A) has not been resolved at the Supreme Court. Whether Lovo's product-versus-source distinction will spread beyond voice actors, or whether appellate courts will treat it as too demanding, is unknown, and it has real tension with cases in which a performer's identity and work are inseparable. How far Hara extends into marketing materials will be tested. How much recognition is enough, measured by what kind of survey, remains stubbornly case by case. And if NO FAKES passes, courts will have to decide how much of the persona branch of Section 43(a) is still needed.

Secondary reading that repays the time includes J. Thomas McCarthy's McCarthy on Trademarks and Unfair Competition and The Rights of Publicity and Privacy, the Restatement (Third) of Unfair Competition §§ 46 to 49, Jennifer E. Rothman's The Right of Publicity: Privacy Reimagined for a Public World and her 2022 article on trademark's lost theory of personality, 135 Harv. L. Rev. 1271, which Lovo relies on, and Robert G. Bone's Hunting Goodwill, 86 B.U. L. Rev. 547 (2006), which Souza cites. To run these authorities down yourself, UpLaw (uplaw.us) carries most of the opinions discussed here.

This article explains general legal principles and cannot tell you how a court would treat your particular photo, voice, or ad. If you are looking at a post with your face on it, or a demand letter about someone else's, a short conversation with a lawyer who handles these claims early, before limitations periods run or evidence disappears, is usually worth more than one held later.

Related Documents

Related Guides

Related Checklists

Related Toolkits

Related Articles