EU court adviser says Gemini may infringe news publishers' rights

Advocate General Maciej Szpunar advised the Court of Justice of the European Union on November 3, 2026 that when a chatbot like Google's Gemini retrieves a press article in real time and reuses it to answer a user's query, the publisher's rights may be infringed, according to the Court's press release No 155/26. The Opinion, in Case C-250/25, Like Company v Google Ireland, turns on the right that Article 15(1) of the CDSM Directive (EU) 2019/790 gives press publishers over the online use of their publications. The Court calls the reference the first generative AI case to reach it.

The dispute comes to Luxembourg from the Budapest Regional Court, where Like Company, a Hungarian operator of web portals, sued Google. Like Company claims Google used its articles to train the Gemini models and reproduced and made them available through the Gemini chatbot between June 13, 2023 and May 31, 2024, the release states. The Hungarian court stayed the proceedings and referred four questions.

Two of the four asked whether training a model on press articles falls under the publisher's reproduction right or is allowed under the text and data mining exception. The other two asked whether the chatbot's answers fall under the reproduction right or the right of making available to the public. Szpunar advises the Court to declare the two training questions inadmissible. He finds no evidence that Like Company's articles were used to train the Gemini models, and reasons that training a model is not a form of "online use" of a press publication, which makes the questions hypothetical. For completeness, he adds that using protected works to train a model is "as a rule, freely permitted under the text and data mining exception." That exception is Article 4 of the same directive.

The answers carry the risk. Szpunar suggests that when an LLM-driven chatbot retrieves a publication from a publisher's site, reuses it to generate an answer, and gives that answer to the user, the publisher's right "may be infringed in certain circumstances." Two factors decide whether the reproduction right is engaged. First, whether the chatbot "accessed the press publication at issue unlawfully," including by circumventing technical measures such as a paywall meant to restrict access or scraping. Second, whether the answer merely reports the facts or reuses "the form of expression of that publication in a recognisable manner," to the point that a reader is "entirely dispensed from accessing the original publication."

The right of making available to the public is infringed, as a rule, where an article's form of expression is "systematically recognisable" across the answers a chatbot gives to related queries. The release also names who answers for it: the provider of the chatbot is liable for such acts where they are "inherent to the business model" or where the provider failed "to put in place appropriate measures in order to counter those acts credibly and effectively."

The advice rests on a boundary the release states plainly. The press publisher's right does not cover the facts a publication reports, only the specific form in which it expresses them. It exists to protect the investment in making a publication, "not to give publishers a monopoly over the news itself."

An Advocate General's Opinion is not binding on the Court. The Judges are beginning their deliberations, the release says, and judgment will be given at a later date. Reuters, in coverage carried by CNA, reported that Advocates General's opinions are not binding but are usually followed by the Court's judges.

The test reaches past this one chatbot. Any AI product that grounds an answer in a publisher's page and returns enough of that page to stand in for it faces the same two questions if the Court follows the advice: how the page was reached, and how much of its expression came back. Two limits matter for anyone building with this. Szpunar's opinion is built on a text chatbot's answers, and the release discusses LLM-driven chatbots only. It says nothing about video, so any claim that it governs an AI video tool that reproduces a script or a scene is an implication, not part of the case.

The Court's judgment is pending. The Opinion proposes a solution the judges are free to reject, and the Court has announced no date for the ruling.