Hagens Berman — the firm behind the $260 billion tobacco settlement — escalates the indie artist copyright fight against Suno and Udio, ElevenLabs adds loop and ambience generation modes to Music v2, and the Sony v. Suno case builds toward a pivotal July ruling.

The $260 billion tobacco firm is coming for Suno and Udio

Hagens Berman — the law firm that secured the largest litigation settlement in history against the tobacco industry — filed an amended complaint on June 22 against Udio in the US District Court for the Southern District of New York. The firm has joined forces with Delgado Entertainment Law to represent independent artists whose recordings were allegedly copied without permission to train Suno's and Udio's AI music generators. The case was originally filed in June 2025 by country musician Tony Justice and his label, 5th Wheel Records.

The complaint alleges both companies used stream-ripping — software that circumvents protections on YouTube and Spotify to download audio directly — to build their training datasets. That's not just a copyright infringement claim; it's a potential DMCA anti-circumvention violation, which carries its own penalties regardless of how the content is subsequently used. When a firm that took down Big Tobacco decides your industry is worth its time, the stakes have changed. If you're an independent artist who has released music on any streaming platform, this case now has serious legal muscle behind it.

Source: Music Business Worldwide, Digital Music News

ElevenLabs Music v2 adds dedicated generation modes for loops and ambience

ElevenLabs shipped a meaningful API update on June 22: a new MusicGenerationMode parameter with three options — track, loop, and ambience. This is the first time a major AI music API has offered purpose-built modes for different audio contexts rather than treating everything as "generate a song." The update also adds higher-quality MP3 output formats (48kHz at 240 and 320 kbps) and bumps the default quality from 128 to 192 kbps.

If you're building anything that needs background audio — games, apps, podcasts, video — the loop and ambience modes are significant. They signal a shift from AI music generation as a novelty toward AI music as infrastructure. The distinction between "make me a track" and "make me a seamless atmospheric loop" is exactly the kind of control that moves these tools from demo toy to production pipeline.

Source: ElevenLabs Changelog

Eros Innovation launches AI music label with seven virtual artists and the Mohammed Rafi estate

India's Eros Innovation launched an AI-powered music label on June 23, debuting seven AI-native artists drawn from its existing entertainment characters using what it calls a Large Cultural Model. The more notable element: a strategic partnership with the family of legendary playback singer Mohammed Rafi to create new recordings and an ABBA Voyage-style live concert experience, plus the establishment of a Mohammed Rafi Academy.

This is the clearest signal yet that AI-native artist identities are moving from experiment to business model, particularly in markets like India where playback singing has long separated the voice from the visible performer. Whether Western markets follow this path remains an open question, but the Rafi estate deal shows how deceased artists' legacies could be commercially extended through AI — with family consent and compensation built into the arrangement from the start.

Source: Deadline

Sony v. Suno: Discovery bombshell sets up pivotal July ruling

The copyright case that could define the legal framework for AI music training is heading to a summary judgment hearing in July before Chief Judge F. Dennis Saylor IV in Massachusetts. Audio fingerprinting during discovery — conducted by experts who spent two weeks inside Suno's offices using Audible Magic's content-recognition system — revealed millions of copyrighted recordings in the company's training data. The labels have since amended their complaint to add over 61,000 songs, including tracks by Taylor Swift and Bad Bunny.

Sony is now the only major label still in court after UMG settled with Udio (October 2025) and WMG settled with both platforms (November 2025). Meanwhile, in Germany, the GEMA v. Suno verdict has been pushed from its original June 12 date to July 31. Two continents, two rulings, one summer. If fair use holds in the US but training-without-licensing fails in Europe, AI music companies face a fragmented global regime — and practitioners face real uncertainty about which tools are built on legally solid ground.

Source: TechTimes, Music Business Worldwide
Tool of the week
Deezer AI Music Detector

Deezer released a free tool that scans your playlists across 20 streaming platforms — including Spotify, Apple Music, and YouTube Music — and flags tracks its algorithm identifies as AI-generated. The detector uses Deezer's proprietary content analysis technology, which the company claims is over 99% accurate at spotting the specific artifacts left in audio by generative software. No account or subscription required.

Context matters here: Deezer now receives nearly 75,000 fully AI-generated tracks per day, accounting for 44% of all daily uploads to the platform. If you're a playlist curator, sync supervisor, or label A&R trying to separate human-made music from synthetic content, this is the most accessible scanning tool available right now. Paste a playlist link, get results.

Free · Web · newsroom-deezer.com
Worth watching
Who pays the session musicians?

While the headlines focus on whether AI companies can legally train on copyrighted music, a quieter fight is opening behind the labels. The American Federation of Musicians sued UMG and Warner Music earlier this month, alleging that when both majors settled with Suno and Udio, they licensed member recordings to AI platforms without compensating the musicians who actually played on them. The AFM argues these deals triggered a "new use" provision in its collective bargaining agreement — the same mechanism that historically ensured session players got paid when vinyl tracks showed up in commercials or on streaming platforms.

This is a different front entirely. It's not about whether AI training is fair use — it's about who gets paid when labels decide it isn't and cut licensing deals instead. If the AFM prevails, every AI licensing agreement a label signs could carry session musician royalties on top. For AI music companies, that makes licensing more expensive. For session players, it could create an entirely new revenue stream from performances they recorded years or decades ago. Watch this one — it reframes the entire AI music debate from "creators vs. machines" to "creators vs. the intermediaries who own their work."