BMG Sues Anthropic Over AI Use of Bruno Mars, Rolling Stones 2026

Music publishing giant BMG Rights Management has filed a major copyright lawsuit against AI company Anthropic, alleging unauthorized use of copyrighted song lyrics — including hits by Bruno Mars, The Rolling Stones, and Ariana Grande — to train its AI models.

The complaint, lodged March 18 in U.S. District Court in California, thrusts the intersection of generative artificial intelligence and intellectual property into the forefront of legal battles shaping the tech industry’s future. 

This case marks the latest escalation in a series of lawsuits by creators and publishers seeking accountability from AI developers for training large language models on copyrighted content without permission.

What BMG Claims in the Lawsuit

According to the complaint, BMG alleges that Anthropic knowingly copied and reproduced hundreds of copyrighted lyrics as part of its training data for Claude, the company’s flagship AI chatbot.

Core Allegations:

• Unauthorized use: Anthropic allegedly used copyrighted lyrics from major artists to train its large language models.

• Reproduction of works: Some of these lyrics were allegedly capable of being reproduced by Claude when users prompted the AI.

  ADVERTISEMENT

• Wide scope: BMG claims 493 examples of infringement spanning rock and pop catalogues.

BMG’s filing argues that sourcing data from unauthorized sites or torrents and using it without a license or consent violates U.S. copyright law and harms rights holders

Anthropic’s Broader Legal History

This lawsuit is not Anthropic’s first encounter with copyright litigation. Over the past few years, the company has faced multiple legal fights relating to how it trains and deploys artificial intelligence models:

Past court exposure has already shown that Anthropic’s training practices are legally contentious, especially where pirated or unlicensed works are concerned.

Why This Suit Matters

The BMG lawsuit underscores legal and ethical questions at the heart of generative AI:

1. Copyright versus Innovation

AI developers argue that training on large datasets containing copyrighted materials can qualify as fair use if the resulting models are “transformative.” However, rights holders counter that unlicensed ingestion and reproduction of lyrics undermines their economic rights and control over artistic works.

2. Financial Stakes

Under U.S. copyright law, statutory damages for willful infringement can reach $150,000 per work infringed. With nearly 500 examples cited, potential liabilities could reach into the tens of millions or more — on top of litigation costs and damages.

3. Industry Precedent

This lawsuit follows another ongoing case brought by Universal Music Group and other publishers against Anthropic in 2023, indicating sustained industry pressure on AI companies.

A major settlement for AI training practices could influence future licensing negotiations between technology developers and content owners, potentially reshaping business models across the AI ecosystem.

Arguments From Both Sides

BMG’s Perspective

BMG’s lawsuit frames Anthropic’s actions as:

• Systematic infringement without consent or compensation.

  ADVERTISEMENT

• Harmful to rights holders who traditionally license their works for use in other media.

• Conduct that undermines the core value of creative property.

The company also faulted Anthropic for allegedly sourcing copyrighted content from unauthorized torrent sites.

Anthropic’s Defense (Historical Context)

Anthropic has historically maintained that:

• Use of copyrighted materials in training is fair use under U.S. law.

• Any instances of reproduced copyrighted content are technology errors or “bugs”.

• The company has enacted internal safeguards to minimize infringement risk.

No official public response specific to the BMG complaint has been issued as of the latest filings.

Legal and Technological Implications

Impact on AI Training Practices

If courts reject the fair use defense for using copyrighted curricula in AI training, companies may be forced to:

• Negotiate licensing deals with rights holders.

• Purchase curated, rights‑cleared datasets.

  ADVERTISEMENT

• Build proprietary data collection systems.

Such outcomes could significantly raise the cost and complexity of training advanced generative models.

Content Industry Pressure

The music industry — alongside authors, news publishers, and creators — is increasingly mobilizing in courts to hold AI companies accountable for unlicensed use of intellectual property. Victory for rights holders could embolden other industries to pursue similar claims.

What Happens Next

Litigation timelines in federal court can stretch for months or years. Potential developments include:

• Preliminary injunctions to halt certain activities.

• Discovery and evidence exchange about training datasets.

• Settlement negotiations, which could set benchmarks for future AI licensing deals.

• Appeals if either party seeks to challenge lower court decisions.

  ADVERTISEMENT

Given the scope of alleged infringements and the precedent implications, BMG v. Anthropic is poised to become a landmark case in the generative AI era.

Conclusion

The lawsuit filed by BMG Rights Management against Anthropic highlights a critical flashpoint in the ongoing battle between creative rights and technological innovation. With major artists, publishers, and technology companies at stake, the outcome will resonate beyond the courtroom — influencing how generative AI systems are built, trained, and commercialized in years to come.

Read next

Mystery AI model | Nvidia H200 approval | Google AI opt-out | BMG sues Anthropic | Amazon AI revenue | Alibaba AI strategy | UK AI content labels | Tencent AI investment | Softcat profit forecast | Samsung-AMD memory deal