EU Court Adviser Says Google's Gemini May Infringe Press Publishers' Rights

Advocate General Maciej Szpunar gave the opinion in a case brought against Google by Like Company, a Hungarian company that runs online news portals, Reuters reported. A Hungarian court referred the dispute, registered as Case C-250/25, to the Court of Justice of the European Union. It asked the Luxembourg court to clarify what protection publishers have when generative AI systems use their content.

According to Il Sole 24 Ore, Like Company alleges that Google used its articles to train Gemini's language models. It also says the chatbot reproduced its articles in answers to users on several occasions between 13 June 2023 and 31 May 2024.

What The Adviser Said

"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 Reuters.

The opinion separates facts from expression. Copyright rules do not protect the "mere facts" reported in press publications, Szpunar said, "only the specific form of expression". Il Sole 24 Ore reported that, in his view, infringement depends on two things. The first is whether the chatbot reached the content unlawfully, for example by getting around a paywall. The second is whether an answer reproduces the article's form so recognisably that the user no longer needs the original.

The Italian newspaper also reported that Szpunar considered whether such answers amount to making the work available to the public. He said systematic reproduction of an article's form in chatbot answers could infringe that right. He added that the chatbot provider could be liable if the reuse is part of its commercial strategy, or if it fails to take credible and effective steps to prevent it.

On training, Szpunar said using copyright-protected material to train large language models is "freely permitted under the text and data mining exception". He nonetheless recommended that the court declare Like Company's training claims inadmissible, because the case file contains no evidence that the company's articles were used to train Gemini.

The claim rests on the press publishers' right in Article 15 of the EU's 2019 Copyright in the Digital Single Market Directive. That right gives news publishers control over certain online uses of their publications.

Why It Matters

The opinion is not binding, and the court has not yet ruled. However, the court's judges usually follow the advocate general's opinion. If they do so here, the ruling would shape how AI companies operating in the EU can use news content.

For publishers, the opinion points to the moment a chatbot fetches and reuses an article to answer a question, rather than to model training, as the stronger basis for a claim. That matters because AI assistants increasingly answer questions by pulling in live web content. For AI developers, the view that training falls within the text and data mining exception would remove one area of legal uncertainty in Europe. But it would leave providers exposed when their products reproduce the wording of paywalled or protected journalism in answers.

The case is one of the first generative AI copyright disputes to reach the EU's top court. Its outcome is likely to be watched by publishers, AI companies and regulators well beyond Hungary. Disputes between news organisations and AI developers over the use of their content continue in several jurisdictions.