List of case law on AI and IP

A. Cases from Germany

1. Hamburg Regional Court – Ref. 310 O 227/23 – A photographer’s copyright in the reproduction of photographs for AI training purposes; The Hamburg Regional Court has ruled that reproduction for subsequent training may be covered by Section 60d of the German Copyright Act (UrhG), at least in the case of scientific models. Read more
2. Federal Court of Justice (BGH) – Ref. X ZB 5/22 – AI models cannot be inventors within the meaning of Section 37(1) of the German Patent Act (PatG). See the decision
3. Higher Regional Court of Cologne – 15 UKl 2/25 – Following a preliminary review in summary proceedings, the plan to train the Meta AI with user data does not give rise to any serious data protection concerns. In particular, the justification under Article 6(1)(f) of the GDPR was examined. To the order
4. The Berlin Regional Court has ruled that even the imitation of a prominent voice using artificial intelligence constitutes an infringement of personality rights. In the specific case, the court awarded the claimant – a well-known voice actor – €4,000 in damages, as his voice had been imitated using AI in commercial YouTube videos without his consent. On the order

B. Proceedings at EU level

1. ECJ – C-806/24 (reference from Bulgaria) – Reference for a preliminary ruling on the interpretation of Article 86 of the AI Act in relation to automated decisions (here: automated invoicing in the event of a telecommunications company’s customer being in arrears) and, in particular, on the links between the AI Act and consumer protection law.
2. ECJ – C-250/25 (reference from Hungary) – Reference for a preliminary ruling on the interpretation of Article 2 of the InfoSoc Directive (reproduction in AI training), Article 4 of the CDSM (text and data mining) and Article 15 of the CDSM (chatbot output as making available to the public) – potential clarification by the highest court of the central issue of AI and IP policy at European level, particularly if the ECJ rules on the question of the applicability of the TDM exception. On the questions referred. Read more at: Hacker, Philipp: Copyright, AI, and the Future of Internet Search before the CJEU: Reflections on Like Company v Google, VerfBlog, 17 July 2025.
   Further details on the proceedings here.

C. Proceedings from EU Member States

1. Městský soud v Praze (Prague Municipal Court) – Case No. 10 C 13/2023 – AI-generated images are not eligible for copyright protection because they do not constitute a creation by a natural person. The prompt used is merely an idea that cannot be protected by copyright.

D. Proceedings from the USA

1. US District Court for the Southern District of New York – No. 1/23 CV 8282 SHS (Authors Guild v OpenAI) – Authors’ copyright rights regarding the use of their works as AI training data, in particular also concerning how to establish proof of the use of works in large language models. Case pending. Read more.
2. US District Court for the Southern District of New York – No. 1/23 CV 11195 (New York Times v Microsoft and OpenAI) – Copyright infringements arising from the use of NYT articles in the training of AI models. In particular, the question of the extent to which LLMs constitute a substitute product for the training data used. In this case, therefore, whether an LLM constitutes a substitute product for newspapers. Furthermore, an in-depth examination of the significance of ‘high-value’ text for the training of AI models (NYT articles constitute the third-largest collection of ‘high-value’ text for AI training, so there are strong arguments that the NYT archive is indispensable for the linguistic sophistication of the models). Regarding the lawsuit.
3. US District Court, Northern District of California – No. 23-cv-03417-VC (RICHARD KADREY, et al., v. META PLATFORMS, INC) – Copyright infringement through the use and output of Meta’s LLM LLaMa, which was trained, amongst other things, using illegally obtained copies of the plaintiffs’ books. The claim has already been largely dismissed; in particular, the court does not consider the LLM as such to be a ‘derivative work’ of the copyright-protected works, nor is the LLM’s output classified as a ‘derivative work’ (in this respect, there was a lack of the necessary concrete evidence for ‘derivative’ output). The proceedings are continuing regarding the use of the training data as a direct copyright infringement; the plaintiffs claim to have evidence that Meta knowingly infringed copyright, with individual employees testifying that they had deliberately used databases of pirated copies, which had also been personally authorised by Mark Zuckerberg (“Meta’s CEO, Mark Zuckerberg, approved Meta’s use of the LibGen dataset notwithstanding concerns within Meta’s AI executive team (and others at Meta) that LibGen is 'a dataset we know to be pirated'").
4. US District Court for the District of Delaware – No. 1/23 CV 00135 (Getty Images (US) Inc. v Stability AI – Copyright infringements arising from the use of the Getty Images stock photo database in training the image-generation AI Stability AI. The focus is on the nature of image generation models as a substitute for stock databases, as well as deliberate attempts to circumvent access barriers such as watermarks during AI development. To the lawsuit.
5. US District Court for the District of Delaware – No. 1/20 CV 613 SB (Thomson Reuters v Ross Intelligence Inc) – First ruling on the justification of AI training using copyright-protected content on the basis of the fair use exception. The District Court of Delaware rejects the application of the fair use exception in this specific case. In particular, the court assumes that complex AI models also constitute direct competing products to the original copyright-protected works, which strongly argues against a fair use exception. Read more.
6. US Copyright Office Review Board – Théâtre D’opéra Spatial – AI-generated images do not meet the ‘human authorship’ requirement if all relevant elements of the image are AI-generated. In the case of “mixed” works, a balancing analysis of the human-created and AI-generated elements of the work is required. The applicant bears the burden of proof in this regard if it is evident that AI was also involved. In doing so, the Review Board adheres to the US Copyright Office’s guidelines for AI-generated works. Read the decision.
7. US Copyright Office Review Board – Zarya of the Dawn – Lack of “human authorship” for individual images created using Midjourney, even with a complex prompting process and minor manual post-processing. However, copyright is possible for the arrangement and selection of the images, as well as for accompanying text (in this case, comic speech bubbles). Read the decision.
8. US Copyright Office Review Board – SURYAST – No “human authorship” either if a self-selected original photo is merely modified using AI applications. Read more.
9. US District Court for the Northern District of California – No. 3/24 CV 05417 WHA (Andrea Bartz et al. v. Anthropic) – Pending proceedings with a preliminary ruling on the application of the fair use rule: It was ruled that the digitisation of lawfully acquired books and the use of the digital database for LL.M. training is covered by the fair use doctrine. The proceedings are continuing with regard to pirated media. To the preliminary ruling.
10. US District Court for the Central District of California – No. 2/25 CV 05275 (Disney v. Midjourney) – Pending proceedings concerning the image generator Midjourney. Disney alleges copyright infringements arising from the AI output and during AI training, in particular through the direct reproduction of copyright-protected characters. To the claim.

E. Proceedings in China

1. Beijing Internet Court (BIC) – (2023) Jing 0491 Min Chu No. 11279 (2023) – AI-generated images are eligible for copyright protection under Chinese copyright law and the ‘prompter’ acquires the copyright in the generated image; in particular, the ‘intellectual achievement’ requirement is met where more than 150 prompts are used to generate the final image.

F. Proceedings from India

1. Delhi High Court – ANI MEDIA PVT LTD v. OPEN AI INC & ANR – Assertion of a copyright injunction to prevent the storage, publication and reproduction of copyright-protected works by OpenAI during the training and provision of ChatGPT.