Round Hill Sues Suno and Anthropic Over AI Training
Round Hill Music filed separate copyright infringement suits against Suno and Anthropic in the Northern District of California on August 17, alleging unauthorized use of hundreds of musical works to train AI systems. Reuters reports that the publisher may expand each case to 10,000 or more compositions and sound recordings, with statutory damages potentially approaching or exceeding $1 billion per case.
Round Hill Music filed separate copyright infringement lawsuits against Anthropic and Suno in the U.S. District Court for the Northern District of California on August 17, alleging that both companies used its musical works without authorization to train AI systems.
According to Reuters, Round Hill alleges that Anthropic used lyrics from at least 500 songs to train its Claude chatbot, while Suno used the same songs to train its AI music-generation system. The publisher's complaints identify works associated with artists including James Brown, the Kinks, and the Goo Goo Dolls.
Round Hill Music LP and five affiliated entities own or control interests in 14,364 musical compositions and 16,873 sound recordings, according to Music Business Worldwide's reporting on the complaints. Each suit includes a 500-work exhibit described in the filings as a prioritized representative bellwether. The complaints state that Round Hill may amend the actions to cover potentially 10,000 or more compositions and recordings.
Claims and potential damages
The suits allege direct copyright infringement under the Copyright Act, as well as circumvention of access controls and removal of copyright-management information under the Digital Millennium Copyright Act, according to Music Business Worldwide and Bloomberg Law. The Suno complaint also names web-data firm Bright Data Ltd. and its U.S. subsidiary as defendants.
Round Hill alleges that Bright Data provided proxy-network and scraping tools used to obtain music and lyrics from licensed platforms, Music Business Worldwide reported. Those allegations have not been adjudicated.
Each complaint seeks statutory damages of up to $150,000 per work for willful infringement. Reuters reported that Round Hill told the court the eventual damages in each action could reach hundreds of millions of dollars and potentially approach or exceed $1 billion, should the litigation expand to the larger catalog of works.
"We intend to take these cases to trial and to hold these companies accountable, and we will not accept a resolution that leaves songwriters and artists deprived of their rightful share of compensation," Round Hill CEO Josh Gruss said in a statement reported by Reuters.
Reuters reported that Anthropic and Suno did not immediately respond to requests for comment.
A growing training-data dispute
The complaints add to a broader series of copyright cases challenging the use of protected material in generative AI training. Reuters notes that Anthropic is already facing litigation from Universal Music Group and other music publishers over alleged lyric use, while Suno faces suits from UMG and Sony Music concerning its training practices.
Reuters also reported that Anthropic agreed to pay $1.5 billion to resolve a class action brought by authors, making it the first major AI company to settle one of the large training-data copyright cases cited in that report.
For ML teams, the Round Hill cases bring attention to a distinct set of inputs: lyrics, recordings, licensed streaming platforms, and alleged methods of bypassing access controls. Comparable disputes often turn not only on whether copyrighted data was used in training, but also on how the material was acquired, copied, and documented. The DMCA allegations against the defendants, and the inclusion of an alleged data-access supplier in the Suno case, place those data-provenance questions directly before the court.
Key Points
- 1Round Hill alleges that Suno and Anthropic trained AI systems on hundreds of protected songs without authorization, expanding copyright pressure on generative AI developers.
- 2The complaints seek up to $150,000 per willfully infringed work and contemplate adding 10,000 or more works, creating billion-dollar stated exposure.
- 3Comparable training-data disputes increasingly examine acquisition methods, access-control circumvention, and provenance documentation alongside the underlying copyright-training question.
Scoring Rationale
The suits target two prominent generative AI providers and frame potential damages at up to $1 billion per case if Round Hill expands the asserted catalog. The allegations also raise practitioner-relevant questions about data provenance, scraping, access controls, and licensed-content ingestion.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
