Google faced a setback on Tuesday when an adviser to Europe's highest court said its Gemini chatbot may breach press publishers' rights when it retrieves news articles and reuses them to answer questions. The opinion draws a clear line that is likely to matter well beyond one case: training a model on published material is usually permitted in Europe, but an answer that echoes an article too closely is a different legal question.
Advocate General Maciej Szpunar gave the opinion in Case C-250/25 before the Luxembourg-based Court of Justice of the European Union. Hungarian website operator Like Company sued Google over Gemini, relying on the press publishers' right in Article 15 of the European Union's 2019 copyright directive. The Budapest court paused its own proceedings and asked the European court how that right applies to a large language model chatbot.
"When an LLM-driven chatbot retrieves a publication from a press publisher's website, reuses that publication to generate an answer, and provides that answer to the user, the press publisher's right may be infringed in certain circumstances," Szpunar said, according to a court statement reported by Reuters.
Szpunar identified two circumstances in particular. An answer can fall under the publishers' right of reproduction when a chatbot gains access by unlawfully circumventing a paywall or other technical barriers, or when it reuses "the form of expression of that publication in a recognisable manner." Copyright does not protect the mere facts reported in an article, he added. It protects only the specific form of expression, a distinction intended to protect publishers' investment without giving them a monopoly over the news itself.
On training, the opinion went the other way. Szpunar said the use of copyright-protected material to train large language models is "freely permitted under the text and data mining exception." He also said there was no evidence that Google had used the Hungarian publisher's articles to train Gemini, and other reporting on the opinion notes that he advised the court to treat the training questions as hypothetical.
Google welcomed that part of the opinion. A Google spokesperson said the text and data mining exceptions are vital for Europe's competitiveness and AI development, and acknowledged the opinion's recognition that European law permits AI training on publicly available information. The spokesperson added that Gemini learns linguistic patterns to create original outputs rather than replicating content, supported by publisher controls that allow rightsholders to opt out.
Publishers argued that the distinction between indexing and answering is the heart of the case. Angela Mills Wade, executive director of the European Publishers Council, said it was important to distinguish between allowing search indexing and authorising AI answers that substitute for original reporting. She said publishers depend on Google to reach readers, while Google increasingly delivers AI-generated answers directly to users, and that publishers should not have to choose between search visibility and protecting their rights.
An advocate general's opinion is not binding, but the court's judges follow the majority of such opinions. A final ruling is expected in the coming months, after which the Budapest court will decide the underlying case. If the court follows Szpunar, a ruling against Google could require fees for certain uses of press content and could offer a lifeline to outlets that say traffic has dwindled as AI summaries replace traditional search results with links. The reasoning would also reach other AI search tools that ground their answers in live news reporting, which would need to summarise facts carefully rather than reproduce an article's wording.