Muscles From Munich? How German Courts Might Stop US Companies From Violating Copyright Through AI Training

Key Takeaways

• Munich court explores international reach of §131 VGG.
• GEMA seeks jurisdiction over US‑based AI training activities.
• Decision could empower EU collecting societies against AI firms.
• US fair‑use defense may be examined under Rome II.
• Outcome may shape cross‑border copyright enforcement for generative AI.

Pulse Analysis

The GEMA‑Suno showdown arrives at a pivotal moment for AI‑driven music creation. After a landmark 2025 judgment that held OpenAI accountable for reproducing song lyrics, German courts are now confronting the more complex question of whether the same legal framework can reach the training data pipelines of foreign AI developers. By invoking §131 of the Verwertungsgesellschaftengesetz, GEMA argues that the act of copying musical arrangements—regardless of where the algorithm processes the data—constitutes an infringement that German courts can adjudicate. This approach leans on the doctrine of "double functionality," which stretches local jurisdiction rules to the international arena, potentially creating a new forum for rights holders.

Legal scholars note that the crux of the case lies in interpreting "place of the infringing act" when the act is a digital process executed on servers abroad. If the Munich judges accept that the training phase qualifies as an infringement occurring in Germany—perhaps because the output is streamed to German users—their decision could set a precedent for applying exclusive jurisdiction beyond national borders. Such a move would force US AI firms to confront German procedural rules and possibly defend their practices under US fair‑use doctrine, as mandated by Article 8 of Rome II, which designates foreign law as the governing substantive rule.

The broader industry impact could be profound. A favorable ruling for GEMA would signal that European collecting societies can pursue cross‑border claims, prompting AI companies to reassess data‑sourcing strategies and licensing models. It may also accelerate negotiations for multinational licensing agreements, as firms seek to mitigate the risk of costly litigation. Conversely, a rejection of international jurisdiction could reinforce the status quo, leaving the onus on legislators to craft clearer trans‑atlantic copyright rules for AI. Either outcome will shape how generative‑AI platforms balance innovation with respect for musical creators’ rights.