Offit Kurman Blogs
Intellectual Property
Longclaw Loses Its Edge: Court Dismisses Game of Thrones Sword Copyright Claim
Jalic, Inc., the officially licensed maker of Game of Thrones prop replicas sold under its Valyrian Steel and Jalic Blades brands, sued more than twenty toy, costume, and novelty companies, accusing them of selling swords that copy “Longclaw,” the wolf-pommel blade (i.e., a sword with a decorative wolf’s head at the end of the handle) carried by Jon Snow in HBO's Game of Thrones series. One defendant, Windy Woman, Inc., doing business as Childish Tendencies, moved to dismiss. In September 2026, a federal court in the Eastern District of Tennessee granted that motion, finding that Jalic had not sufficiently asserted the elements of copyright infringement. The court's reasoning turned on the “Substantial Similarity” filtering test courts use to separate protectable expression from the unprotectable underlying idea. After taking judicial notice of the show's wide cultural reach to satisfy the access element, the court set aside every generic, functionally necessary sword feature (a hilt, a blade, a crossguard) as unprotectable “scenes a faire,” or elements so common to forms of expression that they are ineligible for copyright protection. It also set aside the concept of putting a wolf's head on a pommel; an idea which, however distinctive, sits outside copyright's reach, and, thus, protection attaches only to Jalic's particular form of expression of the wolf’s head. Comparing the exact contours, proportions, and detailing of Longclaw's specific wolf-head design with the accused sword, the court found the overlap too general for a reasonable observer to call the works substantially similar. For businesses that manufacture, license, or sell merchandise inspired by well-known franchises, or that compete in a crowded design category like fantasy weapons and collectibles, the ruling is a reminder that copyright protection reaches the specific way an idea gets rendered and generally not the broad concept behind it. A client accused of copying should ask what is simply generic to the category before conceding similarity, and a client protecting its own designs needs a record of the concrete, distinctive choices that separate its product from the idea behind it. The takeaway is: if you are careful to document the specific expressive detail that makes your product unique, that record, not the shared premise, is what may win the day.
October 8, 2026
Intellectual Property
Lost in 1957, Litigated in 2026: The Copyright Risks Hidden in Older Creative Works
Puerto Rican singer-songwriter Roy Brown spent decades setting the poems of celebrated nationalist poet Juan Antonio Corretjer to music. After Corretjer died in 1985, one of his heirs assigned the poems to Latin American Music Company (LAMCO), a music publisher that registered the works with the Copyright Office in 2000 and began claiming royalties on Brown's recordings. Brown sued, LAMCO countersued, and the parties have been in federal court in Puerto Rico for over twenty years across multiple rounds of litigation. The most recent chapter closed on August 12, when Judge María Antongiorgi-Jordán of the District of Puerto Rico issued a ruling on cross-motions for summary judgment – one which should be on the radar of anyone whose work involves older creative material. The dispute over twelve of the thirteen poems turned on the Copyright Act of 1909, which governed all creative works published before January 1, 1978. Under that statute, publishing a work without a proper copyright notice was fatal to copyright protection, as the work entered the public domain immediately, with no opportunity to fix the mistake. LAMCO and Corretjer's daughter, Consuelo, tried to avoid this result by invoking the "limited publication" doctrine, which provides that sharing copies only with a specifically select group, for a limited purpose, and without allowing further distribution does not trigger the notice requirement. Corretjer was a politically persecuted dissident under FBI surveillance, and the claimants argued he kept his manuscripts within only a trusted circle of allies. They supported that theory with an affidavit from an administrator who had no personal knowledge of any of the events at issue, but the administrator’s description of Corretjer, a political poet determined to reach as many sympathizers and spread his political ideas, undercut the claim that he was carefully restricting who could receive his work. The court granted summary judgment for Brown on all twelve poems. The thirteenth, "Boricua en la Luna" (a piece that functions as something close to a second national anthem for Puerto Rico), survived as a question for the jury. Brown claims Corretjer personally handed him the manuscript at a 1976 concert in Ponce and told him to set it to music. Under First Circuit law, this kind of exchange can establish an oral implied license. Even so, Brown is not out of the woods: issues regarding scope of permission remain, and an informal instruction to "set it to music" may not necessarily authorize decades of commercial recordings, printed lyrics, and album packaging. For clients in the music and arts space, this case illustrates two risks. Pre-1978 creative works are governed by a legal framework that offered no safety net for a missing copyright notice. If the notice was absent when the work was first published, the copyright was lost, and no subsequent registration could claw back the rights. Anyone acquiring or asserting rights in older material needs to examine the original publication history before assuming protection exists. The case also highlights the danger of relying on informal permissions. A verbal authorization granted in a concert hall fifty years ago, with no surviving witness and a principal who has since passed away, may end up before a jury rather than settling the question cleanly, and even a favorable verdict may leave the scope of those rights unresolved.
September 14, 2026
Intellectual Property
Pop Art Time Bomb: The Second Circuit's Ruling in Hayden v. Koons
In the late 1980s, American artist Michael Hayden created a Styrofoam serpent sculpture for Ilona Staller, the Italian adult film actress and parliament member better known as Cicciolina, to use as a prop during her live erotic performances. Hayden sold the work to Staller's production company in 1988 for approximately $900. A year later, Staller’s husband, American artist Jeff Koons, posed with Staller for a series of erotic photographs that would become Koons’ Made in Heaven series. Three of those works depicted Koons and Staller atop Hayden's sculpture, and they debuted at the 1990 Venice Biennale to what Hayden himself described in his complaint as a "media sensation and scandal" that "launched Koons into the art world's stratosphere." Hayden claims he did not discover any of this until 2019, when a news article about an unrelated Staller lawsuit caught his attention. He registered his copyright and sued Koons in December 2021. The case never reached the merits. The Copyright Act requires that infringement claims be filed within three years of when the copyright owner discovers, or reasonably should have discovered, the infringement. The Second Circuit affirmed dismissal on statute of limitations grounds, rejecting Hayden's argument that constructive discovery requires a plaintiff to have actual knowledge of specific triggering facts before the clock starts running. The court clarified that constructive discovery turns on a fact-intensive, objective inquiry into whether a reasonably diligent copyright holder, given all the surrounding circumstances, should have uncovered the infringement. Applying that standard, the panel found the answer here was obvious: Hayden lived in Italy for nearly three decades, was fluent in Italian, consumed Italian news daily, had a direct professional relationship with Staller, and was present in Italy during the very Biennale that made Koons internationally famous, with Staller prominently featured. The court was careful to note that its ruling does not create a "celebrity privilege" that automatically starts the limitations clock whenever a famous artist is involved. Fame is one factor among many, not a categorical rule. The practical lesson for copyright owners is sobering. A rights holder who ignores widespread, international coverage of allegedly infringing work does so at significant legal peril, regardless of whether they actually saw that coverage. Hayden's claim failed not because he sat on a known injury, but because the court concluded a reasonably diligent person in his position could not plausibly have missed it.
August 11, 2026
Intellectual Property
Danger Zone: Top Gun Heirs Ask Supreme Court to Rewrite the Rules on Copyright Similarity
When Ehud Yonay wrote "Top Guns" for California Magazine in 1983, an 11-page account of the Navy's elite fighter pilot training program, Paramount Pictures came calling almost immediately, licensing the rights, and using the article as the springboard for the 1986 blockbuster Top Gun. After Yonay's death in 2012, his heirs attempted to exercise a powerful but often-overlooked copyright tool: statutory termination rights under 17 U.S.C. § 203, which allow an author's heirs to reclaim copyright grants made during the author's lifetime after a statutory period has elapsed. When Paramount released Top Gun: Maverick in 2022 without compensating or crediting the Yonay estate, the heirs promptly sued for copyright infringement and breach of contract. A district court disposed of the case in 2024, granting Paramount’s motion for summary judgment. The Ninth Circuit affirmed in January 2026, holding that Maverick did not infringe the article because every meaningful similarity between the two works, the shared setting, the fighter-pilot culture, aerial training sequences, etc., reflected unprotectable facts about the real Top Gun program rather than any protected original expression from Yunay’s work. While courts ultimately resolved the § 203 claim forming the basis of the suit in the Yonay’s favor, confirming that the rights to the initial article reverted to them, the central legal battleground became "substantial similarity,” the established standard that copyright plaintiffs must satisfy to show that a defendant copied their original creative expression, as opposed to the underlying facts or ideas, which lie in the public domain. The Ninth Circuit applies a two-step approach to determine substantial similarity. First, an "extrinsic test" that dissects a work into its component parts, filters out unprotectable elements like facts and generic ideas, and compares what remains piece-by-piece. Only works that clear this analytical gauntlet proceed to an "intrinsic test,” a holistic, ordinary-observer assessment of overall similarity. The Yonays argued that the article's vivid language, innovative structure, and the distinctive way Yonay wove those elements into a coherent portrait of a fighter pilot’s life were all protectable, including under a "selection-and-arrangement" theory, which recognizes that an original combination of otherwise unprotectable elements can itself merit copyright protection. The Court disagreed; every similarity, the Court stated, either dissolved into uncopyrightable facts about the real program or evaporated into abstract narrative ideas too general to be owned (the "redemption of a hero," beauty and terror juxtaposed in aerial sequences). Thus, the Yonays' selection-and-arrangement theory failed because the narrative patterns they identified were storytelling conventions, not Yonay's original contribution. The Yonays are appealing the Ninth Circuit’s decision and are applying for certiorari to make their case before the Supreme Court, basing their appeal on a new theory highlighting a significant circuit split. While the Ninth Circuit requires plaintiffs to clear the extrinsic-dissection test before any holistic similarity assessment, the Second, Third, Fifth, Seventh, and D.C. Circuits consider overall similarity without the threshold hurdle. Should SCOTUS grant certiorari and find for the Yonays, the result may be a sea change in the way copyright analysis is handled in one of the busiest jurisdictions for such cases. The Second and Ninth Circuits handle the heaviest copyright dockets in the country, covering the publishing and entertainment capitals. Yet, rights holders face materially different legal standards depending on which coast their lawsuit lands. The Yonays’ petition frames this as a $2 trillion problem, invoking copyright-intensive industries' contribution to the national economy, and argues that the Ninth Circuit's approach effectively strips protection from works like Yonay's, where original expression is inseparable from journalism about real events. Whether the Court grants certiorari remains to be seen, but the petition puts a consequential question about how courts measure creative similarity directly on its radar.
August 3, 2026
Intellectual Property
AI Back in Court: MiniMax Studio’s “In your Pocket” Faces Hollywood Studios Copyright Infringement Claims
When an AI company markets its product as a "Hollywood studio in your pocket," it probably shouldn't be surprised when Hollywood lawyers come knocking. Such is the lot of MiniMax, a Shanghai-based tech company whose video and image generation platform, Hailuo AI, became the target of a joint copyright lawsuit filed by Disney, Universal, and Warner Bros. Discovery in a California federal court last fall. The studios' complaint alleges that Hailuo AI was built on a foundation of stolen intellectual property: that MiniMax scraped and trained its model on the studios' copyrighted films without permission, and that the resulting platform can generate eerily accurate, downloadable images and videos of characters like Darth Vader, Wonder Woman, and the Minions, all with MiniMax's own branding slapped on them, at the push of a button. The lawsuit raises two distinct types of copyright infringement claims. The first involves the AI's training: the argument that feeding a model copyrighted films without a license is itself an unauthorized reproduction of those works, regardless of what the model later produces. The second involves the AI's outputs: the finished videos and images that directly replicate protected characters. The studios also pursued a theory of contributory infringement, arguing that MiniMax didn't just passively enable infringement, but actively encouraged it. The company's own promotional materials featured generated clips of the studios' characters, and it sponsored tutorial videos walking users through how to produce content like "Spider-Man and Supergirl kissing in the park." On May 22, 2026, a federal judge denied MiniMax's motion to dismiss, rejecting both the company's claim that a U.S. court lacked authority over a Chinese defendant and its argument that the studios hadn't stated a viable legal claim. On the contributory infringement theory in particular, the court found the studios' allegations sufficient to proceed. The studios have framed the stakes in stark terms, warning that as generative AI advances, it's only a matter of time before these tools can produce full-length unauthorized films. While that outcome remains speculative, the core legal question the case will force courts to answer is concrete: can an AI company build a commercial product on copyrighted works it never licensed, and then profit from an output that reproduces those works on demand? If the studios prevail, the answer will reshape how AI developers approach content licensing and what rights holders can expect in return. For now, MiniMax's motion to dismiss has been denied, the parties are headed toward discovery, and the "Hollywood studio in your pocket" is facing the real Hollywood in court.
June 22, 2026
Trademark and Copyright
“Alright, Alright, Alright,” — Taylor’s Version. Taylor Swift follows Matthew McConnaughey’s Novel Approach to Using Trademark Rights to Enforce Against AI Impersonation
Ever eager to retain control over her masters and ensure that she “never goes out of style,” Taylor Swift is the latest public figure looking toward registration of sensory trademarks to protect her name and likeness in a roundabout way. On April 24, 2026, Taylor Swift's company, TAS Rights Management, filed three trademark applications with the U.S. Patent and Trademark Office: two "sound marks" capturing her spoken voice (which include the language "Hey, it's Taylor" and "Hey, it's Taylor Swift") and one design mark consisting of a photograph of Swift performing onstage during The Eras Tour. This echoes our prior writing regarding similar applications filed by the actor, Matthew McConnaughey, as Swift’s applications represent the latest in a growing movement of public figures attempting to use trademark rights to protect their names and likenesses — most likely due to the increasing accessibility of AI technology, which can impersonate such figures. While sound marks have historically been used to protect iconic brand audio cues, like Netflix's "tu-dum", the MGM lion roar, or NBC's chimes, these public figures have attempted to apply the same framework for their spoken voices and image. This genuinely novel use of trademark law is as-of-yet untested, and Swift's motivation here is not hard to read, as her likeness has been used without permission in numerous AI-generated fakes, including by Meta's AI chatbots, in non-consensual pornographic images, and in false political endorsements shared during the 2024 presidential election. The legal theory underlying these filings is novel and creative precisely because existing law was never designed for this purpose. Under current U.S. law, a musician's recorded performances are protected by copyright law, while the unauthorized commercial exploitation of a person's name or likeness is addressed by state-level right-of-publicity statutes. Individual states, including New York and California, have right-of-publicity laws that prevent unauthorized commercial use of a person's name, image, and likeness (“NIL”), but trademark infringement claims can be filed in federal court, making them a potentially more powerful deterrent as those cases apply nationwide and are not dependent on individual states’ differing enforceability limitations. Most importantly, trademark enforcement doesn't just stop identical uses as copyright enforcement does. Rather, trademark enforcement is designed to protect a rights owner against anything "confusingly similar" to a registered mark. This is a meaningfully broader standard that could reach AI-generated content that approximates, but doesn't exactly replicate Swift's voice or appearance. Trademark claims also enhance the ability to obtain emergency injunctive relief and to recover damages against AI platforms themselves. However, these applications face an unsure road to registration. Trademark protection traditionally requires that the mark function as a source identifier (i.e., signaling to consumers the origin of a product or service) and it is far from settled whether a person's voice or image satisfies this standard. Historically, trademarks are not designed to protect an individual's general likeness, voice, or persona. Swift's filings may be best understood as a deliberate effort to layer additional federal remedies on top of existing right-of-publicity and copyright protections rather than a “cure-all” to the elusive offense of AI impersonation, the scale and sophistication of which is not subject to a single body of law. Whether these applications ultimately succeed, they reflect a broader and accelerating trend: public figures and their counsel actively searching for any available legal structure to fill the enforcement gap that generative AI has created. It is clear that Swift believes that she will continue to Party Like It's 1989™. image credit SockaGPhoto - stock.adobe.com
May 11, 2026
Trademark and Copyright
AI Copyright Litigation Continues as NVIDIA Training Data Case Moves Forward
A ruling earlier this month by Judge Jon S. Tigar in Nazemian et al. v. NVIDIA Corp., No. 4:24 cv 01454 JST (N.D. Cal. filed Mar. 8, 2024), declining to dismiss key claims in the case following NVIDIA’s motion to throw out portions of the complaint, signals that courts continue to be reluctant to resolve copyright disputes concerning AI training and outputs at the pleading stage. The ongoing class action against NVIDIA demonstrates why disputes over AI training data sourcing will continue to shape copyright doctrine well beyond the first wave of generative AI cases. In Nazemian a class of authors, including Abdi Nazemian, Brian Keene, Stewart O’Nan, Susan Orlean, and Andre Dubus III, allege that Nvidia violated the Copyright Act by copying and storing unauthorized digital copies of their books to train its NeMo Megatron large language models, asserting claims for direct infringement, contributory and vicarious infringement, statutory damages, and injunctive relief. They also make claims under the Digital Millennium Copyright Act, alleging removal of copyright management information. Central to the case are the plaintiffs’ allegations that NVIDIA’s training datasets incorporated pirated works sourced from “shadow libraries,” including Books3 (derived from Bibliotik), The Pile, SlimPajama, and Anna’s Archive, each of which allegedly contain massive numbers of unauthorized copies of copyrighted books. Unlike earlier AI disputes that focused on whether model outputs were substantially similar to copyrighted works, the Nazemian action frames infringement as complete at the point of copying of the inputs into the model when works were allegedly downloaded and retained, regardless of whether subsequent model training is transformative. In allowing the direct infringement and related claims to proceed, the court made clear that fair use presents a mixed question of law and fact not suited for resolution on a Rule 12(b)(6) motion, particularly where the provenance, scope, and scale of the copied materials remain disputed. The ruling ensured that NVIDIA would not obtain an early exit from the litigation and underscored that allegations of unlawful data acquisition alone can carry a complaint past the pleading stage. The Nazemian litigation sits within an expanding ecosystem of AI copyright cases, which at present comprises more than 50 such actions pending in U.S. federal courts, including actions involving Meta Platforms, Anthropic, and OpenAI. While recent fair‑use rulings have not stemmed the AI litigation tide, the legal discussion has shifted from abstract debates about innovation policy to examinations of data sourcing, internal decision‑making, and statutory compliance. Even as courts acknowledge that AI training may satisfy the “transformative use” inquiry, they continue to treat market harm, licensing markets, and unlawful acquisition as fact‑dependent questions. It appears that so long as AI developers rely on massive training data sets and courts remain skeptical of practices involving pirated or unlicensed sources, copyright litigation over AI training models will continue to pervade.
May 4, 2026
Intellectual Property
Knowing Isn’t Enough: The Supreme Court Redefines ISP Liability for Piracy
When users pirate music, movies, or other creative works online, the internet service provider (“ISP”) supplying their connection may know more than you might think. Companies like Cox Communications receive thousands of automated notices identifying exactly which subscriber accounts are associated with illegal downloading — in Cox’s case, such notices accrued over a period of two years. In Cox Communications v. Sony Music Entertainment, decided March 25, 2026, the Supreme Court confronted a deceptively simple question: if an ISP knows a subscriber is using its service to steal copyrighted content and keeps providing that service anyway, is the ISP itself liable? A jury of the lower court said “yes,” issuing relief to the tune of roughly $1 billion. The Supreme Court has now unanimously reversed the jury’s decision, although the Justices aren’t in agreement with respect to their rationale and extent. Writing for the majority, Justice Thomas held that an ISP can only be liable for contributing to its users' infringement if it intended that the provided service be used for infringement, particularly in two narrow circumstances: 1) if the ISP actively encouraged the illegal activity, or 2) if the service itself was essentially designed for piracy. The Court found that Cox never promoted piracy and, in fact, issued warnings to and suspended infringing accounts. The majority made clear that simply knowing about infringement and failing to cut off service to every potential infringing account (and, indeed, the record suggests that Cox did not know with total particularity which accounts engaged in infringement) is not enough. Justice Sotomayor, concurring, agreed Cox was not liable but warned that the majority had gone too far in strictly defining only two theories of “intent.” She argued that the ruling diminishes the DMCA safe harbor, which was specifically designed to give ISPs an incentive to crack down on repeat infringers in exchange for legal protection. If ISPs can't be held liable regardless of the very strictly defined theories of intent, that no longer has material effect. Justice Jackson joined Justice Sotomayor in her concurrence. For technology providers, implementing procedures to warn against infringement, and even taking action such as suspending service, may successfully ward off secondary liability. For copyright holders, particularly in the music, film, and entertainment industries, this decision has the potential to present a significant setback for IP enforcement, as avenues for pressuring ISPs to police their networks have been substantially narrowed. Going forward, rights holders may need to focus enforcement efforts more directly on individual infringers or on platforms that actively facilitate piracy, rather than on the companies providing the underlying internet connections. While the decision is a major win for ISPs, the Sotomayor concurrence reasoning could signal that future litigation (or future legislation) may set new standards.
April 27, 2026
Intellectual Property
The Lion King Chant Roars Into Federal Court
Most people know the opening chant of Disney's "The Lion King," even if they can't quite place the words. That chant, "Nants'ingonyama bagithi Baba," was composed and originally performed in 1994 by Grammy-winning South African artist Lebohang Morake, known professionally as Lebo M. Its meaning, as published in the 2019 soundtrack liner notes, is a royal Xhosa proclamation: "All hail the king, we all bow in the presence of the king." On March 16, 2026, Lebo M filed a federal lawsuit in the Central District of California against Zimbabwean-born comedian Learnmore Jonasi, who told millions of podcast and social media viewers that the chant actually translates to "Look, there's a lion. Oh my god." The complaint, seeking over $20 million in damages, raises an unusual and provocative question: when does a comedian's viral joke cross the line into actionable harm to an artist's livelihood and legacy? The lawsuit brings four distinct claims. The first invokes Section 43(a) of the Lanham Act, arguing that Jonasi's false "translation" amounts to a misleading description of Lebo M's commercially significant creative work. The remaining three claims concern state law: 1) defamation per se, alleging the joke implies that a celebrated, award-winning composition is meaningless gibberish; 2) trade libel, targeting the disparagement of the composition itself as a commercial product; and 3) tortious interference with prospective economic advantage, based on Lebo M's concern that the viral mockery could jeopardize his decades-long working relationship with Disney. Notably, the complaint does not include a copyright infringement claim, as Jonasi never reproduced the composition itself. Therefore, the theory of harm here is reputational and commercial, not about unauthorized copying of protected works. Ultimately, the case will hinge on the tension between intellectual property protections and the First Amendment. Jonasi's legal team is likely to argue that a humorous riff on a song lyric's meaning is protected speech, comprising an opinion or parody, not a provably false "statement of fact" as defamation law requires. But the complaint lays groundwork to counter that defense: it alleges that Jonasi presented his translation in an informational podcast setting (as opposed to, for example, a stand-up special where such humor is expected as a matter of course), that Lebo M personally contacted Jonasi with the correct translation, and that Jonasi explicitly refused to retract. If the court finds those facts credible, the case for actual malice (and potentially significant damages) becomes much more compelling. For creators, brand owners, and anyone whose professional reputation is tied to a specific body of work, Morake v. Mwanyenyeka is a case worth watching as it develops.
April 6, 2026
