EU Court Advocate General: Gemini Chatbot Can Infringe Press Publishers' Rights

Advocate General Maciej Szpunar of the Court of Justice of the European Union (CJEU) issued an opinion on 3 November 2026 in Case C-250/25 Like Company v Google. He concluded that an AI chatbot that retrieves a publication from a publisher's website, reuses it and passes an answer to the user can, in certain circumstances, infringe the press publisher's right.

What the dispute is about

Like Company, a Hungarian operator of web portals, sued Google before a court in Budapest. It relies on the neighbouring rights of press publishers. The company claims Google used its articles to train the Gemini language models and that, from June 2023 to the end of May 2024, it repeatedly reproduced them and made them available to the public through the chatbot.

The Hungarian court stayed the proceedings and referred questions to the CJEU for a preliminary ruling. They concern whether model training falls under the publisher's exclusive reproduction right or can benefit from the text and data mining exception, and whether chatbot responses can infringe the right of reproduction or of communication to the public.

Bankier.pl, citing PAP Biznes, reports that Gemini uses a technique called Retrieval-Augmented Generation. It gathers information from the web in real time to improve the accuracy of answers. In the process, the chatbot often retrieves press articles.

Model training left aside

Szpunar proposed that the two questions on model training be declared inadmissible. In his view they are hypothetical, because there is no evidence that Like Company's articles were actually used to train Gemini. According to Bankier.pl, he also held that training in itself is not a form of "online use" of publications.

For completeness, he added that using copyright-protected works to train large language models is in principle permitted under the text and data mining exception. That is a supplementary remark, not the main thesis of the opinion.

When a chatbot infringes

The Advocate General stresses that the press publisher's right does not protect facts themselves or give a monopoly on news. It protects the investment made in creating a publication and its specific form of expression.

According to Prawo.pl, the assessment depends on two factors. The first is whether access to the publication was unlawful, in particular through circumvention of technical measures such as a paywall. The second is whether the answer is limited to facts or recognizably reuses the publication's form of expression to such a degree that the user is completely relieved of the need to consult the original.

The right of communication to the public is in principle infringed when a publication's form of expression is systematically recognizable in chatbot answers. The provider of the AI tool is liable if such conduct is inherent in its business model or if it has not put in place effective measures to credibly counter it.

What comes next, and Poland

The Advocate General's opinion does not bind the judges. According to Bankier.pl and Prawo.pl, the court will take it into account, and the judgment will come at a later date. According to the CJEU, as cited by Bankier.pl, this is the first case concerning generative artificial intelligence to reach the Court of Justice of the EU.

For Polish publishers and companies building products on web content retrieval, the direction of the opinion signals that legal risk may lie more with answers generated live from other people's articles than with model training itself. It is only an advisory position, however, and the final interpretation depends on the CJEU's judgment.