Introduction
The interplay between fashion and sport has always been commercially vibrant. Coco Chanel’s innovative application of jersey fabric in her 1916 sportswear designs and Virgil Abloh’s significant Off-White collaborations with Nike illustrate the mutual influence of the fashion and sports sectors on each other’s aesthetic lexicons and economic aspirations throughout consecutive generations. However, as that relationship has intensified, characterized by the surge of luxury-athletic collaborations, billion-dollar apparel licensing agreements, and the swift ascent of athlete-established fashion ventures, the legal framework supporting it has not adapted accordingly. The athlete, positioned at the commercial nexus of this confluence, faces legal vulnerabilities that current doctrine is fundamentally unprepared to address.
Two subsequent developments have rendered this vulnerability a matter of significant legal concern. The initial aspect is the restructuring of athlete remuneration in American collegiate athletics resulting from the House v. NCAA settlement, which received final court endorsement from Judge Claudia Wilken of the United States District Court for the Northern District of California on June 6, 2025.[i] The settlement establishes direct permissive institutional revenue-sharing between NCAA universities and student-athletes, while maintaining the existing framework of third-party Name, Image, and Likeness agreements, including fashion endorsement contracts, that have been in effect since the NCAA’s July 2021 suspension of its NIL restrictions and were explicitly endorsed by the settlement’s provisions.[ii] It does not, however, address the disputed issue of athlete job categorization, nor does it create a cohesive federal framework for NIL regulation. Its impact is more accurately described as a rearrangement of the compensation framework rather than a complete legal demolition of the amateurism concept. The second aspect is the rapid incorporation of generative artificial intelligence into the fashion design process, a progression that, as recent academic research and the US Copyright Office’s definitive guidance have increasingly highlighted, poses essential and unresolved inquiries regarding authorship, originality, and the extent of intellectual property protection applicable to algorithmically generated clothing design.[iii]
Each development, when evaluated in isolation, poses a discernible theological difficulty. Together, they create a structural condition of significantly greater severity: one in which the athlete’s identity has become commercially essential across the overlapping domains of fashion and sport, while the legal frameworks designed to safeguard that identity remain reactive, fragmented, and theoretically underdeveloped. This article delineates the precise parameters of that state and advocates for a focused response.
The NIL Contractual Paradigm and Its Structural Inadequacies
The settlement in House v. NCAA, designating roughly $2.8 billion in back-pay damages to reimburse former Division I collegiate athletes who participated from 2016 onwards for forfeited NIL, video game, and broadcast-related opportunities, and permitting direct institutional revenue-sharing of up to $20.5 million annually per institution starting in the 2025 to 2026 academic year- signifies the most significant reconfiguration of athlete compensation rights in American legal history.[iv] Fashion companies face immediate and economically important implications: a new class of viable endorsers, sometimes legally inexperienced and lacking professional legal representation, has emerged in the contractual market at scale.
The structural imbalance inherent in NIL endorsement agreements in the fashion sector is well documented in literature and athlete advocacy discussions; however, it remains conspicuously underexamined in academic legal studies. Fashion firms regularly pursue extensive, long-term licenses for an athlete’s name, image, and likeness, agreements whose duration and geographical reach players sometimes accept without fully understanding the limitations they are relinquishing.[v] Practitioner opinion and athlete advocacy groups have regularly observed that the lack of standardized contractual instruction in collegiate athletic programs renders players more susceptible to clauses that would be readily recognized as excessive by seasoned legal counsel. Exclusivity clauses exacerbate this vulnerability: an athlete who enters into an exclusive apparel contract with Brand A may discover, upon being drafted or recruited by an institution or professional franchise endorsed by Brand B, that their NIL agreement directly conflicts with their new institutional commitments.[vi] The pragmatic resolution of that dilemma, in the absence of explicit contractual exceptions, consistently favours the brand.
Standard IP assignment clauses in NIL contracts often aim to transfer not only the right to utilize an athlete’s existing likeness but also to generate derivative works inspired by or referencing the athlete’s visual identity, such as artworks, illustrations, digital renders, and increasingly, synthetic imagery, without explicit restrictions on the production methods.[vii] Prior to the advent of generative AI, such clauses were of limited practical significance, as their effective range was constrained by the costs of custom creative output. In the age of generative AI picture creation, identical contractual language possesses potentially boundless practical possibilities. An athlete who relinquishes “the right to create imagery inspired by or derived from the Athlete’s appearance and persona across all media now known or hereafter developed” has, upon a straightforward interpretation of that clause, authorized a fashion brand to incorporate their photographic archive into a generative model, thereby generating an unlimited quantity of photorealistic synthetic representations, at minimal marginal cost, without additional consent, and without further remuneration.[viii]
This is a tangible issue. The right of publicity, the primary legal framework for safeguarding personal identities in American business law, functions on a state-by-state basis, without a federal norm and exhibiting considerable variance in extent, duration, and assignability. Koski contends in a comprehensive analysis of deepfakes and the right of publicity that the transformative use test employed by courts is fundamentally insufficient for assessing AI-generated digital replicas, as it was designed for a context where imitation necessitated discernible human creative input.[ix] An AI-generated image of an athlete, photorealistic, commercially viable, and produced in seconds from a generative model trained on thousands of reference images, may adequately fulfil the transformative use criterion, as no individual source image is directly replicated, despite serving the same commercial purpose as an authorized photograph and completely undermining the athlete’s economic interest in managing their own image.
The institutional framework of the House settlement exacerbates this vulnerability rather than alleviating it, and the particular mechanisms of that framework warrant more rigorous examination than they have received so far. The settlement established a specialized clearinghouse mechanism, NIL Go, managed by Deloitte on behalf of the College Sports Commission, to evaluate third-party NIL agreements for adherence to the settlement’s dual criteria of ‘valid business purpose’ and ‘fair market value’.[x] Deloitte’s evaluation system appraises agreements based on twelve evaluation criteria, encompassing the athlete’s social media reach, sports performance metrics, regional market, deal term, and the existence of possible pay-for-play signs. The system is substantive rather than just procedural: reports from early 2026 indicate that the College Sports Commission was rejecting a significant percentage of filed agreements, illustrating that the NIL Go review had authentic enforcement authority.
The enforcement weight is calibrated only to one compliance objective: identifying remuneration arrangements that represent disguised pay-for-play, contravening the settlement’s amateurism-related stipulations. It is not calibrated nor intended to analyze the intellectual property framework of the contracts it examines. A fashion NIL agreement that includes a broadly defined IP assignment provision, valued at authentic fair market value for the athlete’s endorsement services, will successfully pass NIL Go review without alterations. The clause’s ramifications in the age of generative AI, specifically, its provision of a license to utilize the athlete’s likeness for training synthetic image models, are wholly beyond Deloitte’s assessment scope. The disparity in legal sophistication between a first-generation collegiate athlete and a specialist in intellectual property counsel for a fashion business is not a mere accidental aspect within this institutional framework. It is integral to the system’s design.
The Authorship Vaccum: Generative AI And the Deterioration of Design Protection
The integration of generative AI into fashion design is now a reality. According to a study, designers at the 2024 New York Fashion Week, including Collina Strada, showcased garments created with AI image-generation tools for prints and silhouettes, a practice the authors suggest may establish a new standard in fashion practice.[xi] The legal ramifications of this normalization are significant and undervalued. The fundamental premise regulating this domain was delineated by the US Copyright Office in its official guidance that copyright protection necessitates human authorship. Works generated only by AI systems are not eligible for registration, irrespective of their artistic or commercial worth. This is not a disputed fringe stance; it is the established interpretative view of the Office and has been validated by the DC Circuit Court of Appeals.[xii] The normative rationale for this regulation, that copyright stimulates human creative labor, is logically consistent in theory. Its application to the intersection of fashion and athletics, however, results in a significant oddity.
The consequences for sportswear are twofold. A fashion business utilizing generative AI for team kit creation may discover that the design lacks copyright protection, rendering it susceptible to replication by competitors. Secondly, and more critically for the athlete, an AI model trained on an athlete’s appearance, movement, aesthetic, and stylistic associations might generate fashion designs imbued with that athlete’s commercial identity, without those designs invoking any intellectual property protection in favor of the athlete.[xiii]
The Spanish case Vegap v Mango, extensively analyzed by Niyompatama and Lapatoura, serves as a pertinent analogy: it included the digitization of copyrighted artworks and their conversion into NFT fashion wearables without the consent of the original creators.[xiv] The court needed to evaluate derivative authorship within a human-AI creative continuum, a challenge for which copyright theory was clearly ill-equipped.[xv] Apply the reasoning to sports: a fashion company develops a generative model with archive video, advertising images, and branded material that showcases a particular player. The model creates a sportswear line ‘inspired by’ the visual identity of that player. No assignment clause is activated, no likeness is explicitly replicated, and no copyright exists in the result. The athlete possesses no entitlement.
No-Man’s Land: The Juridical Void at the Fashion-AI Nexus
The athlete entering a NIL fashion endorsement agreement in the post-House environment faces:
(a) extensive IP assignment clauses significantly broadened by AI-driven creative tools, and
(b) a structural inability to invoke the primary legal doctrines, copyright, trademark, and the right of publicity, that could otherwise limit a fashion brand’s use of its identity.
The outcome is not only insufficient protection; it is a state of compounded legal invisibility. The athlete is, in the strictest sense, a persona non grata within the business ecosystem that their identity has established.
The structural characteristics of this double exposure are most effectively demonstrated by examining the operational mechanics of a sample case. Consider a collegiate basketball player, a post-House NIL signatory, who enters into a conventional fashion brand endorsement contract that confers upon the brand a “global, non-exclusive license to utilize the Athlete’s name, image, likeness, and persona, including the authority to produce derivative works inspired by or referencing the Athlete’s visual identity, across all media currently known or subsequently developed”. This phrasing, or a similar version, is standard in fashion endorsement practices. Prior to 2023, its practical use was constrained by the costs of custom creative production, generating a derivative image necessitated a photographer, a creative director, and a post-production spend. Currently, the identical clause permits a fashion brand to incorporate an athlete’s photographic archive into a generative model like Midjourney or DALL-E, generate an infinite number of photorealistic synthetic images of the athlete adorned in the brand’s apparel, and disseminate those images through digital and print media, all in accordance with the explicit terms of the license, at minimal marginal cost.[xvi] A subsequent issue lies in the drafting of the contract itself. Most current fashion endorsement contracts lack AI-specific clauses, including explicit permissions to utilize an athlete’s appearance for training generative models, to create new representations from historical materials, or to generate prompt-based variants.[xvii] Instead, they depend on ambiguously phrased historical terminology- ‘derivative works’, ‘all media now known or hereafter devised’, established long before generative AI became a commercially significant issue. The resulting interpretative uncertainty is significant and remains legally unresolved. In practice, the ambiguity is not impartial. Brands with a dominant bargaining position and substantial legal resources have continuously leveraged it to their advantage, transforming contractual silence into a de facto license.
The right of publicity, intended to serve as a primary safeguard against such exploitation, is deficient on several fronts. The existing state-law patchwork results in irreconcilable jurisdictional inconsistencies: California’s Civil Code Section 3344 establishes a strong statutory protection against unauthorized commercial use of an individual’s likeness, whilst several other states give only common law remedies of ambiguous extent.[xviii] Secondly, and more significantly, the transformative use defense, the principal mechanism by which defendants in right of publicity cases evade liability, is inadequately aligned with AI-generated material. In Comedy III Productions, Inc. v. Gary Saderup, Inc., the Supreme Court of California determined that a work is deemed transformative and so protected when it incorporates substantial creative features that alter the original.[xix] An AI-generated image of an athlete, created by a model that has analyzed thousands of reference photographs to generate a credible likeness, can be deemed ‘transformative’ since it does not replicate any single source image, despite being commercially indistinguishable from an authorized photograph and fulfilling the same commercial purpose.[xx] The transformative use test, intended for the age of human imitation, was not constructed for statistical synthesis.
The institutional framework of the House settlement exacerbates, rather than alleviates, this vulnerability. The NIL Go clearinghouse, managed by Deloitte for the College Sports Commission, was established with a singular compliance aim- to evaluate third-party NIL agreements for pay-for-play disguised as endorsement activities, thereby fulfilling the settlement’s criteria of ‘valid business purpose’ and ‘fair market value’.[xxi] It was not intended as an intellectual property protection system, nor does it serve as one. The clearinghouse evaluates the business content of a transaction, whether the remuneration corresponds to the athlete’s authentic market worth as an endorser, rather than being based on the contractual framework. A clause for unrestricted IP assignment inside a contract that values the athlete’s endorsement at a fair market price positively undergoes clearinghouse review without alterations. The disparity in legal sophistication between a first-generation collegiate athlete and a specialist in intellectual property counsel for a fashion business is not accidental to this system; it is fundamental to it.
At the professional level, the situation is only slightly more advantageous. Collective bargaining agreements in the National Football League (NFL), National Basketball Association (NBA), and Major League Baseball (MLB) provide group licensing structures managed by player associations, the National Football League Players Association (NFLPA), National Basketball Players Association (NBPA), and Major League Baseball Players Association (MLBPA) respectively, which provide players with collective power over the commercial utilization of their likenesses.[xxii] These frameworks were negotiated in a pre-generative AI context, and they embody the assumptions of that context. The NBPA’s Group Licensing Programme regulates the utilization of player names, numbers, and photos in legally licensed items, specifically tailored for replica jerseys and trade cards, rather than for AI-generated fashion imagery.[xxiii] No existing major-league collective bargaining agreement includes explicit clauses regulating the utilization of player identity as training data for generative models, the creation of AI-generated fashion designs closely associated with particular players, or the obligations of attribution and compensation related to such usage.
Legislative acknowledgement of the fundamental issue is there, but remains incomplete. At the federal level, two legislative proposals have sought to establish a framework for protecting individual identity against AI-generated reproduction. The No AI FRAUD Act (H.R. 6943, 118th Congress, 2024) provides each individual a federal property right over their voice and likeness, imposing penalties for the unauthorized creation or distribution of digital copies.[xxiv] The NO FAKES Act (S. 4875 / H.R. 9551, 118th Congress), endorsed by prominent media and music industry organizations, aims to establish government safeguards against the unauthorized manufacturing and dissemination of replicas. Both legislative proposals, however, function at a level of abstraction that leaves substantial areas unaddressed. Neither examines the intersection of fashion and sport as a unique legal issue, nor do they address the particular challenge that emerges when AI training is not only allowed but explicitly sanctioned through contractual provisions in NIL endorsement agreements, a deficiency that current legislative drafting has not yet addressed. The Tennessee ELVIS Act (Tenn. Code Ann. §47-25-1101) was created in direct response to AI voice cloning in the music business and expands the state’s current personality rights framework to encompass AI-generated imitations.[xxv] Both instruments indicate authentic legislative cognizance. Neither is customized for the intersection of fashion and sport. It neither tackles the particular issue of AI-assisted fashion design informed by athlete identity, the extent of allowable contractual assignment of likeness rights for AI applications, nor the interplay between NIL endorsement agreements and the right of publicity within the context of generative design.
The economic implications of this legislative stagnation are evident. In Thaler v. Perlmutter, the DC District Court upheld that a work generated independently by an AI system, devoid of human creative contribution, is ineligible for copyright registration, a decision that, in the context of fashion, indicates that AI-generated sportswear designs lack protection from their inception.[xxvi] A rival brand may replicate a kit design generated by artificial intelligence, depriving the original brand of any anticipated uniqueness. For the athlete, the situation is particularly pressing: when a brand engages an AI model to create designs inspired by an athlete’s unique aesthetic, including their tattoos, signature colors, and documented stylistic choices, and these designs are subsequently replicated by third parties. The athlete’s legal standing in this situation is, upon further analysis, notably tenuous. According to Thaler, an AI-generated work lacks copyright protection without human authorship, hence nullifying any copyright claim from the beginning. In the absence of a registered mark, trademark law provides no additional recourse. In most countries, the lack of precise likeness duplication precludes any significant right of publicity claim. The athlete’s tattoos, being unique creative works fixed in a physical medium, provide no protection unless copyright has been explicitly granted in writing; without such assignment, the rights belong to the tattoo artist.
At the 2024 Paris Olympics, NBC utilized AI-generated replicas of presenter Al Michaels’ voice for personalized athlete recaps disseminated to millions of viewers, an initiative well covered yet executed without a definitive legal foundation for such usage beyond a contractual license.[xxvii] The episode serves as a definitive model for the future of athlete identity within the fashion-sport intersection- the methodical, scalable, and commercially profitable extraction of personal identity via AI synthesis, facilitated by broadly formulated contractual agreements and unregulated by doctrinal frameworks established for a previous technological epoch. The tolerance of fashion law for this situation, in relation to the fast commercial deployment of generative AI, is increasingly emerging as its most significant unsolved issue.
Suggestions and Conclusion
The intersection of post-amateurism, NIL commercialization, and the integration of generative AI in fashion design has revealed a structural protection gap that is insufficiently covered by both intellectual property law and publicity rights doctrine. This article contends that the athlete faces dual vulnerabilities- from contractual frameworks that excessively extend in the AI era, and from authorship principles that inadequately safeguard against non-human creative outputs.
Three improvements require immediate attention:
- a) NIL standard-form contracts in the fashion industry must incorporate obligatory IP carve-out stipulations: any provision claiming to license an athlete’s likeness for AI training data, generative model input, or synthetic image production must obtain explicit, separately executed consent, delineating clear restrictions on territorial scope and duration. The NCAA’s clearinghouse architecture is a suitable mechanism for enforcing this rule at the undergraduate level; professional leagues’ collective bargaining processes should have analogous clauses.
- b) The federal right of publicity, as proposed in the No AI FRAUD Act, should be expanded to include a distinct sportswear design right: the authority to prohibit the commercial exploitation of AI-generated fashion designs that are significantly associated with a particular athlete’s visual identity, irrespective of the reproduction of any copyrightable elements. Koski’s idea for a likeness license repository offers a viable framework for managing this right on a large scale.
- c) The requirement for human authorship in AI-assisted fashion design necessitates reevaluation, considering the normative framework establishing a minimum threshold of human creative involvement, instead of a binary classification of human versus non-human, would enable both designers and athletes to secure protection for hybrid creative works while upholding the principle that entirely autonomous AI-generated outputs should not grant monopoly rights.
In 2026, the athlete at the convergence of fashion and sport is both the most economically lucrative and the most legally precarious individual in that domain. The law, in its current form, has rendered them alien to their own narrative. That circumstance is neither unavoidable nor permissible.
References:
[i] House v. NCAA (US District Court for the Northern District of California, 6 June 2025).
[ii] Douglas A. Smith, ‘The Evolution of the NCAA’s Antitrust Challenges: NIL, Revenue Sharing, and the Professionalization of College Sports’ (2025) 50 (1) Journal of Education Finance and Law 70, 95.
[iii] US Copyright Office, Copyright and Artificial Intelligence: Report on the Copyrightability of Outputs Created Using Generative AI (Part 2, January 2025).
[iv] Whitney K. Novak, ‘College Athlete Compensation: Impacts of the House Settlement’ (CRS Legal Sidebar LSB11349, Version 2, 15 August 2025).
[v] Jeffrey F Brown, James Bo Pearl, Jeremy Salinger and Annie Alvarado, ‘A Proposal for Group Licensing of College Athlete NILs’ (2021) 12(1) Harvard Journal of Sports & Entertainment Law, 1, 36.
[vi] ‘Explaining Exclusivity Clauses in Athlete Endorsement Contracts’ (CG Sports Team, 2025) <https://www.cgsportsco.com/cejih-explains/explaining-exclusivity-clauses-in-athlete-endorsement-contracts> accessed 30 April, 2026.
[vii] Ho Keat Leng and James J. Zhang, ‘Emerging Trends in Sport Sponsorship and Branding: An Introduction’: In Sports Sponsorship and Branding: Global Perspectives and Emerging Trends (Routledge, Taylor & Francis Group 2024).
[viii] Jonty Cowan, ‘How Generative AI Is Impacting Athlete Image Rights and Endorsement Agreements’ (LawInSport, 2 April 2025) <https://www.lawinsport.com/topics/item/how-generative-ai-is-impacting-image-rights-practical-tips-for-athlete-endorsement-agreements#:~:text=Whilst%20the%20birth%20of%20artificial,brands%20when%20negotiating%20endorsement%20deals> accessed 30 April 2026.
[ix] Reid M. Koski, ‘Warhol, Drake, and Deepfakes: Monetizing the Right of Publicity in the Generative AI Era’ (2024) 40(4) Georgia University Law Review 981 <https://readingroom.law.gsu.edu/cgi/viewcontent.cgi?article=3277&context=gsulr> accessed 30 April 2026.
[x] David P. Weber, ‘Capping the Market: NIL Income Limits and The Shadow of Antitrust Law’ [2026] Forthcoming in Volume 64 of the Houston Law Review (2027) <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6390018> accessed 30 April 2026.
[xi] Mark Jetsaphon Niyompatama and Ioanna Lapatoura, ‘Generative AI in Fashion Design Creation: A Copyright Analysis of AI-Assisted Designs’ (2025) 20(10) 654, 666 <https://doi.org/10.1093/jiplp/jpaf045> accessed 30 April 2026.
[xii] Runhua Wang, ‘The Copyright Requirement of Human Authorship for Works Containing Artificial Intelligence-Generated Content’ (2024) 13(2) IP Theory <https://www.repository.law.indiana.edu/ipt/vol13/iss2/2> accessed 30 April 2026.
[xiii] Suji Kim, ‘The Impact of Artificial Intelligence on the Sport Industry: The Impact of Artificial Intelligence on the Sport Industry: Trademark Challenges and Legal Issues for Sport Brands’ (Doctoral Thesis, University of South Carolina 2025) <https://scholarcommons.sc.edu/cgi/viewcontent.cgi?article=9250&context=etd> accessed 30 April 2026.
[xiv] Lucia González, ‘VEGAP v MANGO: Transformation of Works of Art into NFTs Constitutes Copyright Infringement’ (World Trademark review, 2025) <https://www.worldtrademarkreview.com/article/vegap-v-mango-transformation-of-works-of-art-nfts-constitutes-copyright-infringement> accessed 30 April 2026.
[xv] ‘Art, Fashion Campaigns and NFTs: EU Orthodoxy Restored for Web3 Uses’ (2025) 74(12) GRUR International 1186 <https://doi.org/10.1093/grurint/ikaf128> accessed 30 April 2026.
[xvi] Jonty Cowan (n 8).
[xvii] Mackenna Dunn, Ariana Benitez Colon and Laura Ganoza, ‘How AI, Digital Doubles and New Laws Are Rewriting Fashion and Beauty’ (The Global Legal Post, 2026) <https://www.globallegalpost.com/news/how-ai-digital-doubles-and-new-laws-are-rewriting-fashion-and-beauty-1113297119#:~:text=New%20York’s%20AI%20Transparency%20in,built%20around%20’resurrected’%20icons.> accessed 30 April 2026.
[xviii] CA Civ Code § 3344.1 (2025).
[xix] Comedy III Prods.v. Saderup, 25 Cal. 4th 387, 391 (2001). 21. Cal. CivilCode§ 3344 (West 2001); Gil Peles, ‘Comedy III Productions v. Saderup’ (2002) 17(1) Berkeley Technology Law Journal, 549.
[xx] ETW Corp. v. Jireh Publishing, Inc., 332 F.3d 915 (6th Cir. 2003).
[xxi] College Sports Commission, NIL Go Portal: Submission and Vetting Requirements (2025); Callan G. Stein and Christopher M. Brolley, ‘“NIL Go”: Deloitte Establishes Basic Framework to Review Third-Party NIL Deals’ (NIL Revolution, 20 May 2025) <https://www.nilrevolution.com/2025/05/nil-go-deloitte-establishes-basic-framework-to-review-third-party-nil-deals/> accessed 30 April 2026.
[xxii] Athletes.org, ‘College Athletics Collective Bargaining Agreement Framework’ (Discussion Draft, 28 January 2026).
[xxiii] Chris Smith, ‘AI Avatar Platform Genies Adds Deal with NBPA’ (Sports Business Journal, 6 March 2026)<https://www.sportsbusinessjournal.com/Articles/2026/03/05/ai-avatar-platform-genies-adds-deal-with-nbpa/> accessed 30 April 2026.
[xxiv] No Artificial Intelligence Fake Replicas and Unauthorized Duplications Act of 2024, HR 6943, 118th Cong (2024); Nurture Originals, Foster Art, and Keep Entertainment Safe Act of 2024, S 4875, 118th Cong (2024).
[xxv] Ensuring Likeness Voice and Image Security Act 2024, Tenn Code Ann §47-25-1101 (effective 1 July 2024); Dennis Crouch, ‘DC District Court: AI-Created Works Ineligible for Copyright’ (Patently-O, 18 August 2023) <https://patentlyo.com/patent/2023/08/district-ineligible-copyright.html> accessed 30 April 2026.
[xxvi] Thaler v. Perlmutter, No. 22-1564 (D.D.C. Aug. 18, 2023).
[xxvii] Benjamin Mullin, ‘Now Narrating the Olympics: A.I.-Al Michaels’ (The New York Times, 26 June 2024) <https://www.nytimes.com/2024/06/26/business/media/nbc-olympics-ai.html> accessed 30 April 2026.
Author: Saumya Verma
