AI law update from Germany: After OpenAI, GEMA also takes Suno to court
1. GEMA’s lawsuit against Suno
The Gesellschaft für musikalische Aufführungs- und mechanische Vervielfältigungsrechte (GEMA) is a collecting society for authors of musical works. It currently has more than 95,000 members in Germany and represents the rights of more than two million rights holders worldwide.
In a press release dated 21 January 2025 , GEMA announced that it had filed a lawsuit against Suno Inc. with the Munich Regional Court. The lawsuit is part of a comprehensive package of measures taken by GEMA to ensure adequate remuneration for authors.
The case concerns Suno AI, Suno's generative programme for creating music using artificial intelligence, which can generate both instrumental pieces and songs from a mixture of vocals and instruments. According to GEMA, this is the first lawsuit in Europe against one of the leading providers of AI music tools, while a lawsuit against Suno AI is already pending in the US.
As in the case against OpenAI, GEMA accuses Suno of systematically using copyrighted works to train its AI tool, now exploiting it commercially without compensating the authors of the works. Notwithstanding the somewhat doubtful applicability of German law, the legal issue will essentially be the application of the “text and data mining exception” (“TDM exception”) in Section 44b of the German Copyright Act . As in the parallel proceedings against OpenAI , the Munich Regional Court will have to deal with the question of whether the TDM barrier is applicable to the training of generative AI. In its LAION decision, the Hamburg Regional Court answered this question in the affirmative . If the Munich Regional Court follows this opinion, the interesting question will arise as to whether GEMA has declared an effective, in particular “machine-readable” reservation of use for its members pursuant to Section 44b (3) of the German Copyright Act – in this respect, nothing concrete has emerged from GEMA’s statements so far. In its (appealed) LAION decision, the Hamburg Regional Court recently indicated – rather surprisingly – that a reservation of use in natural language could in principle satisfy the requirements of Section 44b (3) of the German Copyright Act .
However, in addition to the input level, the processes “inside” the AI tool are also the subject of the lawsuit, which GEMA considers to be a reproduction requiring a licence. GEMA is also targeting the output level: According to its own statements, GEMA has been able to document that after simple prompts, Suno AI outputs playable content that largely matches the melody, harmony and rhythm of well-known musical works, including “Forever Young” by Alphaville and “Daddy Cool” by Boney M. GEMA has published samples of the allegedly plagiarised music on its website. GEMA also considers the public reproduction of these “plagiarisms” as acts of exploitation requiring a licence.
It will be interesting to see how the Munich Regional Court positions itself on the key legal issues – especially in comparison to the Hamburg courts. What appears certain, however, is that the legal issues will ultimately be clarified by the Federal Court of Justice and the European Court of Justice.
2. Outlook
The free use of copyrighted works for the purpose of AI training has long been a source of displeasure for rights holders. After all, the quality of the training data, often created by human creativity, is the basis for the performance of an AI model, which then competes with the creators. Against this backdrop, “AI lawsuits” such as the Hamburg LAION proceedings and the two lawsuits filed by GEMA in Munich are likely to be just the beginning. From a legal perspective, they are as exciting as they are challenging, as a number of unresolved legal issues need to be resolved, and the individual proceedings often have their own peculiarities.
It is also interesting to note this second GEMA lawsuit is no longer “only” about song lyrics, as in the case against OpenAI , but about playable music titles, in other words GEMA’s core business. In any case, these proceedings will serve as a model for further AI lawsuits and will most likely not be the last.
You might also be interested in this
In its judgment of 9 July 2026 (C-788/24), the ECJ followed the Opinion of Advocate General Rantos in the legal dispute concerning the online publication of “The Diary of Anne Frank”. At its core, the case centred on the question of whether geoblocking measures constitute effective technological protection measures against copyright infringements, even if they can be circumvented by VPN services. The Court has now answered this question with a clear ‘yes’.
The judgment ‘Obligation to Return VI’ (I ZR 123/25) of 3 June 2026 is – as the name suggests – the sixth judgment by the First Civil Senate of the Federal Court of Justice (BGH) on the so-called obligation to return applicable to hire cars. Following the pronouncement of the judgement, it was widely reported that the BGH had confirmed the continued validity of one of the most controversial provisions of passenger transport law. In the meantime, the dust has settled somewhat and it is becoming clear that this interpretation falls far short of the mark. Indeed, it follows from the grounds of the judgement that maintaining the obligation to return contravenes EU law if one takes its traditional objective and justification as a basis.
In November 2025, the EUIPO Boards of Appeal published a case-law research report on the “Thomson Life doctrine” and the concept of an element’s “independent distinctive role” in composite trade marks. The report is not binding, but it is a useful attempt to bring order to an area of trade mark comparison that even the EU Courts have applied inconsistently.
From 2 August 2026, the labelling obligation under Article 50(4) of the EU AI Act applies – and it reaches practically anyone who uses generative AI professionally to create images, videos, or text. On 10 June 2026, the European Commission published the final Code of Practice on the labelling of AI-generated content. We explain when a label is required, what it must look like, and what risks arise if it is missing.
