Music Industry Escalates Fight Against AI Training as Sony Sues Udio Over Thousands of Unauthorised Recordings
Developed over time first seen 2 months ago
The music industry faces mounting pressure as artists and producers increasingly voice concerns about generative AI tools trained on their work without authorisation. Industry professionals argue that these AI models undermine craftsmanship and creative integrity whilst simultaneously extracting value from existing recordings and artistry. Producers working with major artists are joining the chorus of voices demanding accountability and consent mechanisms for AI training datasets.
Legal action is now escalating alongside public criticism. Sony Music Entertainment has filed a new copyright lawsuit against AI music platform Udio, alleging infringement across more than 30,000 sound recordings that a judge previously prevented the label from including in its original case. The legal pressure signals continued resistance from major labels against generative AI music platforms, with Udio remaining a focal point in the industry's broader effort to establish boundaries and protections around AI-generated music.
- Musicians and producers publicly oppose generative AI tools trained on their work without consent
- Sony Music filed a new lawsuit against Udio alleging copyright infringement of 30,117 recordings
- The conflict reflects industry concerns about AI's impact on artistic integrity and fair compensation
New here? Start with this
The lawsuit sits at the intersection of two things that have been building for a while: the rise of generative AI tools that can create music from text prompts, and a growing standoff between record labels and the companies making those tools. Udio is one of the best-known AI music platforms, letting users generate songs in seconds, and it has drawn scrutiny over what data was used to train it. Sony Music is one of the "big three" major record labels, alongside Universal and Warner, and holds rights to recordings by a vast roster of artists.
Labels and artists argue that AI systems are being trained on copyrighted recordings without permission or payment, which they say devalues the work of musicians and producers and raises questions about consent and compensation. AI companies typically respond that training on existing material falls under fair use or similar legal principles, though this varies by jurisdiction and is still being tested in court. This case is part of a wider pattern of litigation and public pressure from the music industry aimed at forcing AI firms to either license material properly or face legal consequences.
The outcome matters beyond this one lawsuit because courts have not yet settled how copyright law applies to AI training on creative works, so rulings here could shape the rules for AI music generation more broadly, affecting artists, producers, tech companies and listeners alike.
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Advocates for the music industry's position argue that recordings represent years of investment, skill and financial risk, and that copyright exists precisely to ensure creators are paid and consulted when their work is used commercially. They contend that training AI models on copyrighted recordings without a licence is not meaningfully different from any other unauthorised commercial use, and that allowing it to proceed unchecked would hollow out the economic basis for professional music-making, harming working musicians and producers far more than large labels. Seeking consent and compensation mechanisms, they say, is not an attack on innovation but a reasonable insistence that new technology operate within long-established legal and ethical norms.
The case against
Those sympathetic to AI music platforms argue that training a model on existing recordings to learn patterns, styles and structures is a transformative act akin to how human musicians learn by listening to and absorbing the work of others, and that copyright law was never intended to give rights holders veto power over such learning processes. They worry that expansive infringement claims, especially ones reaching tens of thousands of recordings, could be used by dominant incumbents to entrench their market position, stifle a promising new creative technology, and limit the tools available to independent artists and smaller developers. On this view, the courts should weigh carefully whether training itself constitutes infringement, rather than assuming that any use of copyrighted material without a licence is automatically unlawful.
Coverage
- Rolling Stone — Have Artists Reached Their Breaking Point With AI?
- Music Business Worldwide — Sony Music sues Udio again, asserting over 30,000 recordings a judge barred the major from adding to its original case