Key takeaways
- In Case C-250/25 Like Company, Advocate General Maciej Szpunar advised the EU Court of Justice on Nov. 3, 2026, that an AI chatbot that retrieves a press article in real time, reuses it to generate an answer and gives that answer to a user may, in certain circumstances, infringe the press publisher’s rights [1].
- For the reproduction right, he said this depends on whether the chatbot accessed the article unlawfully, for example by circumventing a paywall or anti-scraping measures, and whether the answer reuses the article’s form of expression so recognisably that the user no longer needs to read the original [1].
- The opinion is not binding. The judges are now deliberating, and the Court will give judgment at a later date [1].
Direct answer: The case, which the Court’s press release calls the first generative AI case to reach the Court of Justice, concerns the real-time retrieval behind AI answers rather than only model training [1]. The Advocate General said the press publisher’s right protects the specific form of expression of a publication, not the facts it reports [1].
What happened: Like Company, a Hungarian operator of web portals, sued Google Ireland before the Budapest Regional Court over Google Gemini [1]. It claims protection for its articles under the EU right for press publishers over the online use of their press publications. It argues that Google used the articles to train the Gemini models, and that between June 13, 2023, and May 31, 2024, Google repeatedly reproduced them and made them available to the public through the Gemini chatbot [1]. The Hungarian court asked the Court of Justice four questions: two on model training and two on the chatbot’s answers [1].
The press release describes Gemini as using retrieval-augmented generation (RAG), or “grounding,” to collect information from the web in real time. It says the chatbot often retrieves online press articles to answer questions about the news [1].
What did the Advocate General advise?
- Training questions. He advised the Court to declare the two training questions inadmissible as hypothetical, because there is no evidence that Like Company’s articles were used to train the Gemini models and because training a model is not a form of online use of press publications [1]. For completeness, he suggested that using protected works to train a model is, as a rule, freely permitted under the text and data mining exception [1].
- Reproduction right. In his view, whether a chatbot’s use of a press publication falls under the publisher’s reproduction right depends on two factors: whether the chatbot accessed the publication unlawfully, in particular by circumventing effective technical measures such as a paywall designed to restrict access to, or web scraping of, the publication; and whether the answer merely reports facts or reuses the form of expression “in a recognisable manner, to such an extent that the user, by reading that answer, is entirely dispensed from accessing the original publication” [1].
- Making available to the public. He said this right is, as a rule, infringed where the form of expression of the press publication is systematically recognisable, to that extent, in the chatbot’s answers to related queries [1].
- Provider liability. He said the chatbot provider is liable for such uses where they are inherent to its business model, or where it has failed to put in place appropriate measures to counter them credibly and effectively [1].
Advocates General propose a legal solution to the Court in complete independence; the opinion does not bind the judges [1]. In a preliminary ruling the Court does not decide the national dispute: the Budapest court will resolve the case in line with the Court’s ruling [1].
Why does this matter for AI visibility?
Our view: the Advocate General’s proposed reproduction test looks at how the chatbot accessed the article and whether its answer reuses the article’s form of expression so recognisably that the user need not read the original [1]. Both concern the real-time retrieval the press release describes as RAG or grounding, not model training [1]. Competition authorities are also examining how AI answers use publishers’ content, as in the interim measures on Google AI Overviews in Türkiye we covered on Nov. 3.
What we don’t know yet
- Whether the Court will follow the opinion and when it will rule [1].
- How the Budapest Regional Court will apply the ruling to the facts of Like Company’s claim against Google [1].
- Google’s response; the press release does not include a statement from Google or Like Company [1].
- The detailed reasoning in the full opinion, which is published on the CURIA website; this article is based on the Court’s press release and case page [1][2].
What to do now
This is not legal advice.
- If you publish news in the EU, note which technical measures, such as a paywall, restrict access to or scraping of your articles: the proposed reproduction test asks whether a chatbot circumvented effective technical measures [1].
- When you monitor AI answers that draw on your articles, note whether they report facts or recognisably reuse your form of expression, the distinction the opinion draws [1].
- Treat the opinion as a non-binding proposal until the Court of Justice rules in Case C-250/25 [1].
Update log: Nov. 4, 2026 — First published.
Sources
[1] Court of Justice of the European Union. (2026, Nov. 3). Press release No 155/26, “Opinion of the Advocate General in Case C-250/25 Like Company.” Accessed Nov. 4, 2026. https://curia.europa.eu/site/upload/docs/application/pdf/2026-11/cp260155en.pdf
[2] Court of Justice of the European Union. (n.d.). “Opinion: Case C-250/25 Like Company.” CURIA. Accessed Nov. 4, 2026. https://curia.europa.eu/site/jcms/p1_1000088866/en/opinion-case-c-250/25-like-company
How this was made
Prepared with AI assistance from the sources cited above. Editorial review by Nhi Phan (editor) and fact-checking by Jackies continue after publication; any change will be recorded in the update log and in our Corrections Policy and Log.