Written by: Aidan Persinger
The rich history of musical copyright law can be traced back to the 1700s. Shortly after the U.S. gained independence, Congress, using its enumerated powers, enacted the Copyright Act of 1790. Under the 1790 Act, creators were granted exclusive rights over their work for an initial period of 14 years following registration.[2] Music was not mentioned specifically until the year 1831, where musical works were finally recognized as copyrighted content.[3]
Technological progression has enabled the public to create music digitally. Computer scientists first explored the idea of using algorithms to compose music in the 1950s.[4] A landmark moment in this era was the creation of the Illiac Suite in 1957.[5] This piece, generated using the ILLIAC I computer, followed rules derived from traditional music theory, creating the first instance of AI-generated music.[6] Technology has progressed even further since then, and AI can now learn patterns and create music autonomously.[7]
The rise of technologically-created music has caused friction between musicians and AI companies. This friction materialized in the ongoing case, Concord Music Group, Inc. v. Anthropic PBC, (N.D. Cal. Mar. 25, 2025). Plaintiffs in the case allege that Anthropic’s AI program, named “Claude,” used their copyrighted song lyrics to train Claude, who can output said lyrics to any user of the AI program.[9] Anthropic contests that its use of the lyrics fall under the “fair use” exception of 17 U.S.C. § 107.[10]
Courts analyze fair use using a four part test.[11] First, the purpose and character of the use, including whether such use is commercial or for educational purposes.[12] Anthropic argues that the use of copyrighted material in training Claude serves a distinct purpose from the original intent of the works.[13] By employing the material to teach its AI model, Anthropic contends that this application adds new meaning and message, aligning with fair use principles.[14] Second, the nature of the copyrighted work.[15] Anthropic acknowledges that the creative nature of song lyrics places them at the core of copyright protection, but the company emphasizes that the training process does not exploit the expressive aspects of these works but rather focuses on analyzing patterns and structures, which should weigh in favor of fair use.[16] Third, the amount and substantiality of the portion used in relation to the copyrighted work as a whole.[17] Anthropic maintains that any reproduction of copyrighted material during training is incidental and justified by the transformative purpose.[18] The company argues that the extent of use is limited to what is necessary for the AI to learn language patterns, supporting a fair use claim.[19] Fourth, the effect of the use upon the potential market for or value of the copyrighted work.[20] Anthropic contends that using copyrighted works to train Claude does not harm the market for the original works.[21] The company argues that the AI’s training process does not substitute or diminish the value of the original content, thereby not adversely affecting the potential market.[22]
The ongoing legal feud between Thomson Reuters and Ross Intelligence could be a persuasive factor in future AI cases.[23] If Concord, like Thomson Reuters, can prove Claude’s use of musician-owned lyrics was commercial instead of transformative, it might give them an edge in regards to the first factor of the fair-use analysis.[24] Concord could argue that Anthropic “stands to profit from exploitation of the copyrighted material without paying the customary price,” by training Claude with copyrighted music.[25] If an original work and a secondary work share a similar purpose, and the second use is of a commercial nature, the first factor is “likely to weigh against fair use.”[26] The second fair-use factor, nature of the original work, could potentially weigh in Concord’s favor as well, since a musician is essentially an “artist drafting a work from scratch.”[27] In regards to the third factor, the “amount and substantiality” made accessible as a competing substitute would be perceptually similar for an average Claude user looking for lyrics.[28] Finally, Claude’s “effect on a potential market” for musicians could be substantial, as “restrictive licensing conditions,” can foreclose rivals’ access to copyrighted material. [29] The Thomson Reuters Court maintained that Ross had the burden to prove that said potential markets do not exist and would not be affected, so Anthropic could have to do the same if the Court considers this non-generative AI ruling to be relevant.[30]
This case will be one to watch closely, as its holdings are extremely likely to affect the future of AI in the musical marketplace.
[1] Copyright Alliance, History of Music Copyright, COPYRIGHT ALL (Mar. 29, 2025), https://copyrightalliance.org/history-music-copyright/. [https://perma.cc/26HN-QPKR].
[2] Id.
[3] Id.
[4] The Origins of AI Music, STACCATO.AI (Mar. 29, 2025), https://blog.staccato.ai/en-US/The-Origins-of-AI-Music. [https://perma.cc/HV6G-XAHP].
[5] Id.
[6] Id.
[7] Id.
[8] Concord Music Grp., Inc. v. Anthropic PBC, No. 24-cv-03811-EKL, 2025 U.S. Dist. LEXIS 55639, at *5 (N.D. Cal. Mar. 25, 2025).
[9] Id.
[10]17 U.S.C. § 107 (2025).
[11] Id.
[12] Id.
[13] Concord Music Group et al. v. Anthropic PBC, MUSIC COPYRIGHT INFRINGEMENT RES. (Mar. 29, 2025), https://blogs.law.gwu.edu/mcir/case/concord-music-group-et-al-v-anthropic-pbc/. [https://perma.cc/PB5V-MF3D].
[14] Id.
[15] Id.
[16] Id.
[17] Id.
[18] Dani Deahl, Anthropic Reaches Deal with Music Publishers over AI-Generated Lyrics, THE VERGE (Jan. 3, 2025), https://www.theverge.com/2025/1/3/24334866/anthropic-claude-music-publishers-lyric-copyright-lawsuit-deal. [https://perma.cc/FXM5-ZPJX].
[19] Id.
[20] Id.
[21] Id.
[22] Thomson Reuters Enter. Ctr. GmbH v. Ross Intel. Inc., No. 1:20-cv-613-SB, 2025 U.S. Dist. LEXIS 24296, *21 (D. Del. Feb. 11, 2025).
[23] Id.
[24] Id.
[25] Id.
[26] Id.
[27] Id.
[28] Id.
[29] Thomson Reuters Enter. Ctr. GmbH v. Ross Intel. Inc., No. 20-613-LPS, 2022 U.S. Dist. LEXIS 75493, at *17 (D. Del. Apr. 26, 2022).
[30] Id.
