AI Copyright: Training-Data Lawsuits and Licences — 2026-09-23
This week's headlines center on the collision of litigation and licensing: record labels that sued AI music companies are now licensing to them via Suno's new major-label-backed model, while a runoff of fresh US filings pushes total AI copyright suits to 140. In Europe, Munich courts keep issuing landmark rulings on AI memorisation, and Japanese media flag a mysterious bulk purchase of tens of tonnes of Japanese books likely destined for AI training abroad.
AI Copyright: Training-Data Lawsuits and Licences — 2026-09-23
Top developments
Suno's new label-backed model shifts music labels from suing to licensing
A new Suno AI music model built with major record labels is now live — the same labels that sued the AI music companies are licensing them, though payouts remain undisclosed, the lawsuits continue, and musicians are questioning who actually benefits. The Los Angeles Times reports the deals mark a turn from pure litigation toward monetisation of training data in music.
German court rules GEMA's case against Suno: memorised output is infringement
The Regional Court of Munich I (LG München I) held that memorised music works and similar AI outputs produced by Suno constitute copyright infringement — a follow-on to its landmark November 2025 ruling against OpenAI over unlicensed song lyrics (including Grönemeyer's). The GEMA ruling is not yet final (nicht rechtskräftig).
Microsoft executive concedes licensing obligation in German AI copyright fight
Unsealed documents in the German litigation show internal warnings and significant click losses for media sites, and the Microsoft chief acknowledged an obligation to license — with a judgment in the case not expected before 2027.
Meanwhile in the US: Google suit, NYT fair-use test, and Anthropic's $1.5B payout
Fresh US filings and rulings frame the week: Google faces a new copyright suit over Gemini's training data; the NYT v. OpenAI fair-use decision is pending before a Manhattan federal judge; and the Bartz v. Anthropic $1.5 billion settlement received final approval on July 20, 2026, remaining the largest payout in this space.
Local view
Germany: The BGH held a hearing on 3 September 2026 (Case I ZR 281/25) on whether AI may train on protected photographs — a case with major implications for photographers under German text-and-data-mining law.
Japan: Nippon TV (日テレNEWS) reports a wave of "mysterious bulk orders" at Japanese second-hand bookstores, with export records to the US showing "50 tonnes of Japanese books" — the article links it to Anthropic's identified practice of destructively scanning books for AI training. Yomiuri separately covered Sony Group's music publisher subsidiary among 35 companies suing Anthropic over unlicensed use of tens of thousands of song lyrics.
Context & numbers
- US copyright lawsuits against AI companies reached 140 cases by September 2026, up sharply from 45 in June 2025 (Edward Lee's case map).
- The Anthropic (Bartz) copyright settlement of $1.5 billion received final approval on July 20, 2026.
- The largest disclosed licensing deal to date remains OpenAI's $250+ million over five years (announced May 2024).
- News/journalism dominates the licensing economy with 48 deals, ahead of music/audio (16) and images/video (12).
On the radar
- Manhattan fair-use ruling in NYT v. OpenAI — the judge must decide whether training on copyrighted news and books is fair use; a decision could reshape the economics of generative AI.
- German Suno ruling on appeal — the LG München I decision is not final; appeal status worth tracking.
- Judgment in the German case against Microsoft over Copilot-style media use expected in 2027.
- Suno's undisclosed label payouts — watch whether per-track or per-stream splits become public, and whether musicians file separate claims. (Rumored areas of tension flagged by reporting.)
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
