Music rights holders are escalating legal action against AI companies over unauthorized use of copyrighted songs in training data, with one publisher claiming potential damages exceeding $1 billion. Independent music publisher Round Hill Music filed suit against Anthropic and Suno in California federal court, alleging the companies misused hundreds of songs to train their AI systems without permission or compensation. The lawsuit represents the latest flashpoint in a widening global dispute over how AI models are trained and who bears responsibility when copyrighted material is used at scale.

What Exactly Are AI Companies Accused of Doing?

Round Hill Music alleges that Anthropic used lyrics from at least 500 songs in which the publisher holds rights to train Claude, the company's large language model (LLM), which is an AI system trained on vast amounts of text data to understand and generate human language. Suno, an AI music generation platform, allegedly used the same songs to train its music-creation system. The songs cited include works associated with artists such as James Brown, The Kinks, and the Goo Goo Dolls. Round Hill said it could eventually add "ten thousand or more" compositions to the lawsuits, with potential damages that could approach or exceed $1 billion, according to reporting by Reuters.

James Brown, The Kinks, and the Goo Goo Dolls

This case is not isolated. BMG Rights Management sued Anthropic in March, alleging that the company used copyrighted lyrics from artists including the Rolling Stones, Bruno Mars, and Ariana Grande to train Claude, citing 493 alleged examples of copyright infringement. Anthropic is also facing a separate ongoing lawsuit brought by Universal Music Group and other music publishers, while Suno has faced claims from major music companies over the use of copyrighted recordings in AI training.

How Are Courts Around the World Responding?

Legal scrutiny is not confined to the United States. A Munich regional court ruled on July 31 that Suno violated copyrights involving songs represented by German licensing agency Gema. The lawsuit concerned the alleged use of music available on Alphabet's YouTube, including songs associated with Alphaville. The court said Suno did not have the right to process the works and ordered the company to disclose illicit revenue, while damages are yet to be quantified. Gema CEO Tobias Holzmueller stated that the "verdict is of global significance," according to Reuters. Suno disagreed with the ruling and said it would evaluate its options, including an appeal.

Tobias Holzmueller

However, the industry is also exploring licensed alternatives. In November 2025, Warner Music Group settled its copyright dispute with Suno, paving the way for new licensed AI music models. Under that arrangement, Suno said its new models would replace existing versions in 2026, illustrating how the industry is simultaneously fighting unauthorized use and exploring licensing-based commercial models.

Why Is This Battle So Complicated?

The core issue centers on a fundamental question: can AI companies train their systems on copyrighted material without obtaining licenses or permission? This question remains unresolved in India and many other jurisdictions. The challenge is especially significant for music because AI can analyze enormous quantities of lyrics, compositions, recordings, and vocal characteristics before generating new material. When AI-generated works incorporate protected material without attribution, questions arise over whether the original creator loses revenue and whether existing royalty systems can identify the use before payments are distributed.

The monetary impact is central to the dispute. Rights holders earn from multiple streams, including streaming, publishing, and public performance royalties. If AI-generated works incorporate protected material without attribution, the original artist may lose revenue. The recognition and attribution of original artists associated with the music becomes critical to ensuring proper compensation flows through the system.

How Do Royalties Actually Get Tracked in an AI World?

Royalties could depend heavily on the credits and metadata attached to a release. Hari Nair, Strategic Advisor at ContentLens and former CEO of Tips Music, explained the challenge: "It depends entirely on what the artist or the company that uploads the song declares at the point of release in their metadata," he said. "Where no original rights holders are credited, royalties could flow entirely to the uploader, making monitoring and identification of unauthorised use increasingly important." This is one of the biggest unresolved questions in India as well as globally.

Hari Nair, Strategic Advisor at ContentLens and former CEO of Tips Music

Steps to Address AI Copyright Challenges in Music

• Licensing and Consent: AI companies should obtain explicit licenses from rights holders before using copyrighted material in training data, following the model established by Suno's settlement with Warner Music Group.
• Metadata and Provenance Tracking: Clear attribution and metadata standards must be established at the point of release to ensure royalties flow to original creators when AI-generated works incorporate protected material.
• Clearer Legal Frameworks: Jurisdictions need to establish definitive rules about whether AI training on copyrighted material constitutes fair use, and who bears responsibility when unauthorized use occurs.

What Does India's Copyright Law Say About AI-Generated Music?

The Indian copyright law separately protects sound recordings, musical compositions, and lyrics, subject to applicable legal requirements. For computer-generated literary and musical works, the Copyright Act of 1957 identifies the author as the person who causes the work to be created. However, determining who that person is becomes more difficult when AI plays a substantial or autonomous role.

JV Abhay, Partner at Shardul Amarchand Mangaldas, noted the complexity: "If the lyrics and musical composition of a song are generated entirely through the AI tool, with minimal human intervention or input, whether or not there would be any copyright in such lyrics and composition would need to be clarified first," he said. Abhay pointed out the Delhi High Court's recent interim-stage decision in Tarun Chaudhary & Anr. v. Kuldeep Meena & Ors., where the court refused interim relief in a dispute involving an AI-generated song because it was not satisfied that copyright could subsist in the work.

Jasmine Damkewala, Advocate on Record at the Supreme Court, similarly noted that Section 2(d)(vi) of the Copyright Act does not conclusively identify whether the developer, AI company, commissioning label, or prompting user should be treated as the author. She noted that a bare prompt with little further human creative contribution could face difficulty in meeting India's originality threshold, while selection, editing, arranging, and other traceable creative decisions could strengthen an ownership claim.

Why Are Companies and Rights Holders Taking Action Now?

Experts said that the increase in disputes reflects both rising awareness of intellectual property rights (IPR) and the growing importance of intangible assets in the digital economy. Nirupam Lodha, Partner at Khaitan & Co., explained the broader context: "The increase in copyright and personality rights disputes reflects both a greater awareness of IPR and the unprecedented scale at which content is being created, distributed, and monetised in the digital economy," he said. "At the same time, advancements in AI, social media, and digital platforms have created new forms of use and potential misuse of content, making disputes more visible and more frequent. What we are witnessing is the natural evolution of IP law in response to technological change."

Nirupam Lodha, Partner at Khaitan & Co

As AI-generated music moves from experimental technology into a commercial part of the music ecosystem, the stakes continue to rise. The lawsuits against Anthropic and Suno, combined with court rulings in Germany and ongoing disputes globally, signal that rights holders are increasingly willing to enforce their legal rights. The outcome of these cases could reshape how AI companies approach training data, licensing agreements, and compensation models across the music industry and beyond.