THE ESSENTIALS IN 30 SECONDS

  • Advocate General Maciej Szpunar delivered his opinion in Like Company v Google (C-250/25) on November 3, 2026, in a dispute over the Gemini chatbot.
  • He says chatbot answers may infringe the press publishers' right if content is accessed unlawfully or its form of expression is reused recognisably.
  • Training LLMs is, in his view, generally covered by the text and data mining exception; the opinion is not binding and a ruling is expected later.

An adviser to the Court of Justice of the European Union has given the first detailed view on how EU copyright law applies to generative AI chatbots. On November 3, Advocate General Maciej Szpunar delivered his opinion in case C-250/25, Like Company v Google, which commentators describe as the first generative AI case to reach the Luxembourg court.

The dispute

Like Company, a Hungarian operator of online news portals, sued Google Ireland before the Budapest Regional Court. According to CEA Digital Law, the publisher relies on the press publishers' related right in Article 15 of the 2019 copyright directive and claims Google used its articles to train Gemini models, then reproduced them and made them available through the chatbot between June 2023 and May 2024. Agence Europe reports that the Hungarian court referred four questions in March 2025, two on training and two on the chatbot's answers.

What the Advocate General concluded

On answers, Szpunar says the press publishers' right protects the way an article is written, not the facts it reports. Reusing press content in a chatbot response may still infringe that right in certain circumstances. Per the opinion as summarised by RTÉ and CEA Digital Law, the reproduction right comes into play when the chatbot gets around a paywall or other technical barrier, or when the answer reuses the publication's wording in a recognisable way, to the point that users no longer need the original. The provider may be liable, he adds, where such uses are built into its business model or where it has not taken credible and effective countermeasures. Copyright scholar João Pedro Quintais notes that on the example before the court, the Advocate General provisionally found these conditions not met, leaving the final call to the Hungarian judges.

On training, Szpunar recommends declaring both questions inadmissible, since there is no evidence the claimant's articles were used and training is not an "online use" covered by Article 15. He nonetheless indicates that training on protected material is generally permitted under the text and data mining exception, provided its conditions, such as lawful access and respect for opt-outs, are met.

What happens next

The opinion does not bind the court, although RTÉ notes that judges follow most such opinions. A judgment is expected in the coming months and, according to Quintais, could shape other pending Hungarian references and national cases such as GEMA v OpenAI. A Google spokesperson told RTÉ that EU text and data mining exceptions are vital for Europe's AI development and that publishers can opt out.