UMG and Sony vs Suno: 61,000 More Works Added to the AI Lawsuit

Discovery in the major-label case against Suno revealed the AI music generator was trained on millions of copyrighted recordings. UMG and Sony are now trying to add 61,000 more works to the lawsuit. Here's what it means — and how artists can check if their metadata is in the training data. VibeBeats is direct-licensed background music for cafés, restaurants, bars, gyms, retail stores, salons, spas, clinics and hotels. No music licensing/PRO fees (*No separate public-performance music licence needed for VibeBeats music.) Other music sources may still require licensing. The AI DJ reshapes each playlist through the trading day.

TL;DR

Court filings in the major-label lawsuit against Suno — the AI music generator behind viral track factories — have revealed something the company spent a year denying: Suno was trained on millions of copyrighted recordings owned by Universal Music Group, Sony Music and Warner. UMG and Sony are now moving to add 61,000 additional copyrighted works to the complaint. If you record and release music, your catalogue was very likely part of the training set. You can check your metadata against the leaked training-data indexes on our AI training data check page.

What the discovery actually showed

For most of 2024 and 2025, Suno's public line was that its training data was "proprietary" and that discussing it would harm the business. The Music Business Worldwide report confirms what many suspected: once the court forced discovery, Suno's own internal documents showed the model ingested millions of commercial recordings from the three majors, plus a long tail of independent artists.

That's a big legal shift. The plaintiffs no longer have to argue "we believe Suno trained on our music." They can point at Suno's own engineering records.

Why UMG and Sony are adding 61,000 works

Under U.S. copyright law, statutory damages are calculated per infringed work. The original complaint listed a representative sample of songs. Adding 61,000 more works isn't just paperwork — it's a direct multiplier on potential damages, which run up to US$150,000 per willful infringement.

If the court accepts the expanded list and finds Suno's use was willful, we're looking at a theoretical damages ceiling in the multi-billion dollar range. Even a small fraction of that would end Suno as a company.

The independent-artist angle nobody is covering

The majors have lawyers. Independent artists — the ones actually making the long tail of tracks that trained these models — don't.

If your music is on Spotify, Apple Music, YouTube, SoundCloud or Bandcamp, and you're not on one of the three majors, the discovery documents suggest your recordings were probably ingested too. But you're not in the lawsuit. There's no consent, no notice, no payout, and — in international markets — no obvious legal path to force one.

The Atlantic's AI Watchdog team built a searchable index of the datasets these AI companies used. You can check yourself:

Check if your music is in the AI training data →

What international artists should know

international copyright law has no text-and-data-mining exception and no positive right to opt out of AI training. That's the current state confirmed by the composition society and the Creative international markets AI Reference Group (CAIRG).

Practically, that means:

• U.S. lawsuits like the Suno case set the global tone, but they don't directly compensate international rights holders. • The composition society is pushing for a statutory licensing regime — similar to how radio and public performance royalties work today. • Individual artists have almost zero leverage. Collective action through the composition society and CAIRG submissions is the highest-impact path.

Our international-specific action guide walks through the three real levers: sign the composition society petition, submit to CAIRG, and add explicit reservation-of-rights language to your releases.

What this means for businesses using AI-generated music

If you're a café, gym, retailer or hotel running a generic AI-music service, ask one question: where did the training data come from?

Services built on unlicensed scrapes are a legal exposure — not for you as the venue directly, but for the platform you're paying. If Suno-style services lose their U.S. cases, expect a wave of takedowns, catalogue rewrites and price hikes as the industry restructures around licensed data.

VibeBeats sits on the other side of that line. Every track in our catalogue comes from direct licences with the artists we work with. Our music is not sourced from web scrapes and is not shared with third-party AI trainers. That's not a marketing claim — it's the reason we can issue a Certificate of Compliance and skip music licensing entirely.

The three things to do this week

Check your metadata on our AI training data check tool. Sign the composition society AI campaign if you're an international rights holder. Audit your business's music source — if you can't answer "where did the training data come from?", you're carrying someone else's legal risk.

Where this is heading

The Suno case is unlikely to be the last. Udio, Stability, and half a dozen smaller AI music startups face similar exposure. Expect:

• A statutory licensing framework for AI training in at least one major jurisdiction within 12–18 months. • Direct-licensed AI catalogues becoming the only credible offer for commercial venues. • Metadata provenance becoming a standard purchase criterion — the same way "organic" became a food label.

If you're an artist, the leverage is in collective action. If you're a business, the leverage is in choosing a platform that can prove its supply chain. Either way, "we didn't know" stops being a defensible answer very soon.

Ready to switch to a music service with a clean training-data story? Book a 15-minute demo and we'll show you exactly where every track in our catalogue came from.

Related: The rise of AI music artists

Start your 14-day free trial — no credit card required.