The Legal Beat: AI Tech Faces Off with the RIAA on Intellectual Property
- TULJ
- 5 minutes ago
- 12 min read
Hannah Reilly
Edited by Keerthi Chalamalasetty, Emily Mandel, Judge Baskin, and Sahith Mocharla
The rapid and sudden implementation of generative AI technology has drawn significant attention in recent years, presenting one of the most critical legal debates regarding intellectual property law, intended to protect the products of human ingenuity and creativity. At the center of this evolving conflict are the cases of UMG Recordings, Inc. v. Suno, Inc. (2024) and UMG Recordings, Inc. v. Uncharted Labs, Inc. (2024), involving AI music-generation services. Both of these lawsuits had major record labels as plaintiffs and will establish the legal constraints going forward for generative AI usage in music, possibly requiring services such as these to comply with strict copyright laws in areas such as software training and website operation. While these large industries spearhead this legal battle, independent artists face identical threats without institutional backing, emphasizing the importance of equitable protection across the entire music industry. Overall, these cases, particularly the settlement of UMG Recordings, Inc. v. Uncharted Labs, Inc. (2024), will serve as a definitive test of AI’s legal bandwidth and its ability to utilize patented creations and copyrighted records, ultimately asserting whether such a usage for the purpose of training generative models is a violation of basic intellectual property rights for the future of these technologies.
Filed in the United States District Court for the Southern District of New York in June of 2024, the case of UMG Recordings, Inc. v. Uncharted Labs, Inc. (2024) brought together a coalition of music labels and record companies. The case included big names like Sony Music Entertainment (SME), Warner Records, Inc., and UMG Recordings, Inc. which, collectively, fall under the larger umbrella of the Recording Industry Association of America (RIAA). The RIAA assured these record labels that stakeholders, including musicians and songwriters, would maintain control over their works, and that they would put an end to their music’s unlicensed use [1]. According to the RIAA, they are seeking two primary outcomes in the lawsuit [2]. The first is the confirmation that the services participated in copyright infringement and are subject to injunctions that would obstruct them from infringing on such rights, particularly from these record labels, in the future. The RIAA’s second stated goal is that the companies address damages that have already occurred. Addressing these damages could include compensating shareholders for the unauthorized use of their recordings, identifying and immediately removing these infringements from their training datasets, disclosing future training sources and materials to the RIAA, or even retroactively licensing the use of copyrighted material by paying royalties to artists [3]. The plaintiffs are primarily targeting Uncharted Labs, Inc., the company that the AI-music creation service, Udio operates under. In this case, the record labels, who are a “great majority of the most commercially valuable sound recordings in the world,” allege that Udio has engaged in “willful” copyright infringement on a tremendous scale [4].
To understand the significance of these lawsuits, it is necessary to examine the vast scope of data ingestion, the training process, and general operation of models like Udio’s. Both Udio and Suno have established subscription-based business models that monetize AI-generated music; a consequential fact given that the commercial nature of their technology is a significant factor in the alleged acts of copyright infringement. For example, Udio charges users between $10 and $30 per month depending on the subscription tier, with the highest pricing tier allowing for the maximum amount of music files and/or subscription amenities. For a time, the platform had launched a free tier but quickly eliminated it upon the increase of user demand, which is likely clear evidence of commercial exploitation of copyrighted works without compensation. AI music platforms such as these software programs often operate on an industrial scale. This industrial output is made possible by generative modeling techniques trained on a massive “corpus of communications,” allowing the platforms to recombine patterns of melody, harmony, rhythm, and vocal style [5]. Studies of these text-to-music platforms such as Udio show that they have brought in hundreds of thousands of users and collectively generated over 100,000 songs within the span of just a few months [6]. These technologies rely on a “textual prompting interface,” the same system implemented by ChatGPT and other Large Language Models (LLMs), allowing these users to create brand new songs through everyday language descriptions such as “in the punk rock style,” or “a smooth jazz sound” [7]. Analyses of certain datasets from these platforms reveal an ongoing pattern of similar prompts and lyrics, frequently including references to real artists, genres, and protected, copyrighted works. This evidence demonstrates how users employ certain creation strategies in order to reproduce stylistically recognizable signatures, which blurs the line between inspired and replicatory. The complaint from UMG Recordings observes these findings, documenting music outputs that are almost identical to copyrighted recordings [8]. These reproductions capture not only genre but also specific melodic, harmonic, and vocal styles. These patterns are all consistent with the observation that AI systems embed stylistic elements from human-created works that are used in their training. Given the software’s capacity to generate millions of tracks at an almost instantaneous speed, they pose a significant threat to most licensing companies for synchronization, sampling, and derived works. Without legal oversight, the influx of AI-generated music will likely devalue human creativity and ingenuity as well as undermine the economic basis of copyright protection.
One of the many parameters of creating generative technology like Udio or Suno is the ingestion of great amounts of data in order to “train” new software models. In the case of UMG Recordings, Inc. v. Uncharted Labs, Inc. (2024), the plaintiffs argue that Udio copied and ingested “decades worth of the world’s most popular sound recordings,” music described by Udio’s executives as “the best quality music that’s out there” [5]. These recordings allegedly originated from existing online repositories and public platforms that currently host copyrighted works, rather than from licensed datasets. As such, the plaintiffs are filing for both compensatory and statutory damages to address both the implicit damages caused as well as the pre-determined amount awarded in any case of copyright infringement. Additionally, UMG is seeking injunctive relief. This would constitute a court order requiring an immediate stop to all activities and training programs that violate copyright laws to prevent future infringements from the defendants. The plaintiffs allege that the magnitude of the case calls for rather extreme measures to be taken, given the extensive range of protected works that were infringed. For example, the claims address the full historical scope of recorded music protected under these record labels and under federal law, spanning from as early as 1792 to the present. In turn, the plaintiffs have taken the measures deemed necessary by them to address the enormity of these violations. They explicitly requested a jury trial in order to determine questions of liability and assess appropriate damages. This choice in procedure likely reflects the plaintiffs confidence in presenting compelling evidence of infringement to a jury of peers, rather than relying solely on judicial determination [6]. The presence of a jury also plays a significant role in how this case will be used to determine the outcome of similar cases, given that a verdict given by a jury of peers may be favored to a judge’s verdict in shaping long-term norms regarding ethical AI usage in the public sphere.
The plaintiffs in the Udio case represent the most powerful entities in the recorded music industry: the aforementioned Universal Music Group, Sony Music Entertainment, and Warner Music Inc., as well as Capitol Records, Arista Music, Atlantic Recording Corporation, and Rhino Entertainment Company. These plaintiffs own or exclusively control copyrights in what the complaint describes as “a great majority of the most commercially valuable sound recordings in the world” [7]. The coordination among these normally competitive industrial giants emphasizes the considerable threat that they all perceive from the training of this AI music software. Their unified legal strategy aims to establish a clear precedent that copyright law fully applies to AI development, disregarding any justifications of technological advancement or innovation. The case of UMG Recordings, Inc. v. Uncharted Labs, Inc. (2024) is positioned to be a determinant of whether generative AI companies can claim exemption from basic intellectual property protections that have governed creative industries in the U.S. for years. The plaintiff’s choice of venue is likely strategically significant because the case is being adjudicated in the United States District Court for the Southern District of New York, where Uncharted Labs, Inc. is based. This ensures that the court has personal jurisdiction over Uncharted Labs, Inc. when the verdict is given.
The complaint demands similar forms of relief designed to establish strict operational constraints on AI companies throughout all industries. The plaintiffs seek both preliminary and permanent injunctions that would constrain the defendants from continuing their, allegedly, violating conduct. Such injunctions would essentially require these and similar AI companies to always obtain proper authorization before training on any copyrighted material. This would likely establish a licensing regime or system that extends from AI companies’ initial data collection phase of training all the way through the entire model development process as well as the ongoing operation of the platform [9]. This would ensure that no single part of the process infringes on the intellectual rights of creators. From a monetary front, the plaintiffs are requesting statutory damages for willful infringement or, alternatively, the actual damages combined with the disgorgement or the releasing of all profits that the companies gained from training or operation that involved copyright infringement. The complaint emphasizes that the statutory framework permits damages up to $150,000 per work that was willfully infringed. This could result in a considerable liability to these companies given the enormous number of alleged recordings copied. This level of damages not only would serve as compensation in a settlement but as a strong deterrent in the future, making unauthorized training financially out of reach for any generative AI software company considering similar practices.
This current AI music litigation bears instructive parallels to earlier landmark copyright cases that arose during other feats of technological developments. Most notably, the case of A&M Records Inc. v. Napster, Inc. (2001), provides an applicable precedent for rejecting technological innovation as a justification for copyright infringement. In Napster, the Ninth Circuit confronted a peer-to-peer file-sharing technology that enabled users to freely exchange copyrighted music without authorization, a development that coincided with a period in which the “majority of music consumption was [became] unauthorized” and ultimately enabled “a pattern of blatant copyright infringement” that the music industry struggled to regulate through licensing [10]. Napster argued, similar to Udio, that the service qualified for certain protections under various legal theories including (but not limited to) fair use, substantial non-infringing uses, and the limitations of secondary liability. The industry’s aggressive litigation strategy proved largely ineffective at stopping this unauthorized distribution, as it failed to find a concrete system for regulating or tracking peer-to-peer file sharing [11]. The court systematically rejected these defenses, pointing out that Napster facilitated direct infringement by users and that they were liable for contributory and vicarious copyright infringement [12]. The court held that the introduction of some new, innovative technology did not exempt companies from compliance with copyright law.
Several key principles from Napster apply to the current AI music disputes and support the conclusion that courts can and should impose strict copyright compliance requirements around emerging technologies. First, the Ninth Court emphasized that commercial use of copyrighted material starkly contrasts with the doctrine of fair use, particularly because the use serves the same function as the original work, such as Napster creating music to be listened to. Suno and Udio operate on essentially the same model, commercializing services that generate music (for listening) which is identical to the purpose of any copyrighted recording with a record label [14]. Second, the court in Napster found that the availability of a substantial market for licensing the copyrighted works negates their first defense of fair use, as the unauthorized use interfered with the copyright holder’s ability to exploit the market [15]. Suno and Udio mirror this in an even stronger light. The major record labels involved in the case have actively and clearly licensed their catalogues of music to numerous technology platforms including streaming services, social media platforms, and fitness apps. The existence of these licensing markets demonstrates both the willingness of rights holders to authorize the use of their music, and also negates any kind of accessibility defense from Udio. These AI platforms’ decision to bypass these markets, rather than going through the legal process, reflects a conscious choice to risk copyright infringement under the assumption that they would not be eligible to gain the rights to the copyrighted material under fair use [16].
Applying Napster’s reasoning to AI music platforms reveals that these cases should result in clear victories for the rights holders in the cases of UMG Recordings, Inc. v. Suno, Inc. (2024) and UMG Recordings, Inc. v. Uncharted Labs, Inc. (2024), and establish comprehensive legal constraints on AI training in the future. Like with Napster, Suno, and Udio produce outputs that clearly violate copyrighted recordings, serve fundamentally the same purpose as the original works, and operate in a context where licensing markets do exist. The cases of Suno and Udio actually present a stronger case for liability than Napster. Where Napster only facilitated the distribution of existing copyrighted recording, these AI platforms blatantly utilize these recordings and meld them into commercial products for their own monetary gain––critically, without any authorization from the record labels. The layered nature of their infringement as well as the faultiness of their defense only makes the explicit violation of copyright more obvious, not less.
These cases may establish a precedent for copyright infringement on these record companies but it still brings into question the legitimacy of such a precedent in regards to small and independent artists. Krystle Delgado, known for her youtube page, “Top Music Attorney,” is an independent artist and lawyer at Delgado Entertainment Law. Delgado is pursuing an additional lawsuit against Suno and Udio on behalf of independent artists who have been “left without a seat at the table” and lack support from major record labels like Warner Records and UMG Recordings [17]. Degalo argues that a great majority of the “tens of millions” of copyrighted songs belong to independent musicians and creators. When the doctrine of fair use was debunked as a legitimate defense in regards to Suno and Udio’s usage of copyrighted material, Degalo Entertainment Law joined forces with Ellzey, Kherkher, Sanford & Montgomery, LLP (EKSM) to engage in a class-action lawsuit. The case is similar to UMG Recordings, Inc. v. Uncharted Labs, Inc (2024), with a focus on getting independent artists the compensation that they equally deserve. The plaintiffs in this case are pursuing a similar outcome of demanding damages and compensation for the creators that have been affected as well as requesting a permanent injunction against Suno, Udio, and their respective companies engaging in copyright violation [18]. The only difference is these injunctions would ensure that independent artists are protected from copyright infringement on their works as well, not just those backed by the RIAA.
Looking forward, these lawsuits will establish binding legal precedent that shapes how generative AI software can be integrated and developed across all creative industries. Major record labels are currently litigating against companies much like Suno and Udio, who defend their training on millions of songs by arguing that the resulting outputs are “transformative,” and therefore fundamentally new works [19]. The case will answer the key legal questions of whether AI training requires authorization, whether moderate copying constitutes infringement and whether the fair use doctrine applies to AI platforms, particularly monetized AI. These questions contribute to determining if generative AI is a lawful or unlawful extension of human creativity or a fundamental exception to copyright law. District courts must now decide whether AI training for models qualifies as a “non-expressive” fair use, as the defendants argue, or whether it unlawfully reproduces protected works. This has strong potential to reach the Supreme Court because they breach topics of fair use and unresolved questions about AI’s legal and creative bandwidth. A broad interpretation of fair use, as the defendants propose, would allow for largely unrestricted text and data sifting. By contrast, a narrow interpretation would require licensing in order to better protect rights holders. This approach aligns with both the U.S. and the European Union’s rights-holder-friendly frameworks which are consistent under the EU AI Act, permitting opt-outs and requires transparency from AI platforms about training data usel [20]. The UMG Recordings-Udio settlement, reportedly involving licensing, suggests that AI companies relying on copyrighted material will ultimately need to compensate creators, independent and label-holding alike, in one way or another. Meanwhile, the ongoing suit against Suno reflects these AI platforms' strategy of claiming the ambiguity of laws and attempting to establish clear legal boundaries for AI training. With more international lawsuits emerging, such as Denmark and the E.U., global coordination is likely to establish a middle-ground in AI regulation, balancing innovation with protection of human authorship [21]. If courts hold that AI training requires authorization, companies will face obligations to license training data, maintain records, pay royalties, and disclose training sources publicly. These standards would be similar to licensing in film, publishing, and other non-AI softwares. Such outcomes will not harm innovation but simply integrate AI into existing legal and ethical structures. Ultimately, these cases will determine whether copyright and intellectual property law remains a meaningful safety net for human creativity or if that becomes obsolete in the era of generative AI technologies.
[1] Record Companies Bring Landmark Cases for Responsible AI Against Suno and Udio in Boston and New York Federal Courts, Respectively, RIAA (Jun. 24, 2024), https://www.riaa.com/record-companies-bring-landmark-cases-for-responsible-ai-againstsuno-and-udio-in-boston-and-new-york-federal-courts-respectively/.
[2] See [1]
[3] See [1]
[4] UMG Recordings, Inc. v. Uncharted Labs, Inc., 1:24-cv-04777, S.D.N.Y., 2025
[5] C. Roads, Artificial Intelligence and Music, 4 Computer Music Journal 13 (1980) https://www.jstor.org/stable/3680079?seq=1
[6] Luca Casini, et al., Data-Driven Analysis of Text-Conditioned AI-Generated Music: A Case Study with Suno and Udio, (2025).
[7] See [6]
[8] See [6]
[9] See [1]
[10] Charles Goldstuck, Past Precedent, Future Proof: Toward a New Legal and Commercial Framework for AI-Generated Music (SSRN, 2025). https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5393574
[11] See [10]
[12] A&M Records, Inc. v. Napster, Inc., 239 F.3d 1004 (9th Cir. 2001) https://jipel.law.nyu.edu/andersen-v-stability-ai-the-landmark-case-unpacking-the-copyright-risks-of-ai-image-generators/
[13] See [12]
[14] See [12]
[15] See [4]
[16] See [4]
[17] I’m Suing Suno & Udio..., (2025), https://www.youtube.com/watch?v=VlKBpyjZU4I.
[18] The AI Copyright Problem... | Lawyer Reacts, (2025), https://www.youtube.com/watch?v=wy08U3gEU8w.
[19] Tristan Williams, Composing the Future: AI’s Role in Transforming Music, Center for AI Policy (Nov. 17, 2024). https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5959675
[20] See [18]
[21] Matt Blaszczyk, et al., Artificial Intelligence Impacts on Copyright Law, (2024).