In his Opinion delivered on Tuesday 3 November in Case C-250/25, Maciej Szpunar, the Advocate General of the Court of Justice of the European Union (CJEU), considers that an artificial intelligence-based chatbot’s reuse of press publications to generate responses may, “in certain circumstances, infringe the rights of press publishers”, according to the Court’s press release.
Those circumstances depend on how the article or protected content was obtained, particularly whether the chatbot accessed it unlawfully “by circumventing effective technical measures”, and on the extent to which it was reproduced.
Asked in March 2025 by a Hungarian court to rule on a dispute between publisher Like Company, a Hungarian operator of several online media outlets, and digital giant Google over its Gemini chatbot, the Court is examining a case of this kind for the first time. The publisher accuses Google of using its articles “to train” Gemini, then reproducing them and making them available to users via Gemini.
On training models using protected content, the Advocate General recommends that the Court “declare inadmissible the two questions concerning LLM [large language model] training”. He considers that “the use of works and other subject matter protected under copyright (...) is, as a rule, freely permitted”. The chatbot provider “is liable for such potential acts of use of press publications”, he stresses, particularly if it has not taken “appropriate” measures to “to counter those acts credibly”.
Link to the Advocate General’s Opinion: https://aeur.eu/f/o8c (Original version in French by Louis Dubar)