EU Court Adviser: AI Chatbot Answers Can Infringe News Publishers' Rights, but Training Is Generally Allowed

An adviser to the Court of Justice of the European Union said Tuesday that AI chatbots that pull news articles from the web in real time to answer users' questions may, in certain circumstances, infringe the rights of press publishers, in the first generative AI case to reach the Luxembourg court, according to the court's press release.

Advocate General Maciej Szpunar's opinion in Case C-250/25, Like Company v. Google Ireland, is not binding. The judges have begun deliberating and will rule at a later date, the court said. The court's role for advocates general is to propose legal solutions "in complete independence," according to the release.

The case

Like Company, a Hungarian operator of web portals, sued Google in the Budapest Regional Court over Google Gemini, the AI chatbot made by Alphabet's Google. The company claims protection under the EU's press publishers' right, created by the 2019 Copyright in the Digital Single Market Directive, which covers the "online use" of "press publications," according to the press release.

Like Company argues that Google used its articles to train the large language models behind Gemini and repeatedly reproduced them and made them available to the public through the chatbot between June 13, 2023, and May 31, 2024. The court said Gemini uses a technique called retrieval-augmented generation, or "grounding," which collects information from the web in real time to improve the accuracy of answers, and that to answer news questions, the chatbot often retrieves, or "scrapes," press articles.

The Hungarian court paused its case and asked the EU court four questions: two about training and two about the answers the chatbot gives.

Training: inadmissible, and generally permitted

Szpunar advised the court to declare the training questions inadmissible. Because there was no evidence Like Company's articles were used to train Gemini's models, and because training is not a form of "online use" of press publications, the questions were hypothetical, he said.

"For the sake of completeness," the press release said, he suggested that using protected works to train a language model is, as a rule, freely permitted under the EU's text and data mining exception, Article 4 of the 2019 directive.

Answers: two tests

On the answers, Szpunar said that when a chatbot retrieves a publication from a publisher's website, reuses it to generate an answer and provides that answer to a user, the publisher's right may be infringed in certain circumstances.

He stressed that the right does not protect facts reported in the news, only the specific form of expression, because its aim "is to protect the investment that goes into making such a publication, not to give publishers a monopoly over the news itself," according to the court's summary.

Whether a chatbot's use falls under the publisher's reproduction right depends on two factors, the adviser said. The first is whether the chatbot accessed the article unlawfully, "in particular by circumventing effective technical measures (such as a paywall)" designed to restrict access or scraping. The second is whether the answer merely reports facts or "reuses the form of expression of that publication in a recognisable manner, to such an extent that the user, by reading that answer, is entirely dispensed from accessing the original publication."

The right of making available to the public is, as a rule, infringed where the article's form of expression is "systematically recognisable" in the chatbot's answers to related queries, he said.

Liability

The chatbot provider is liable where those uses "are inherent to the business model of that provider" or where the provider has failed to put in place appropriate measures to counter them "credibly and effectively," according to the opinion summary.

Why it matters beyond Europe

The opinion arrived one day after The New York Times and three other U.S. newspaper publishers sued Brave Software in federal court in Manhattan, alleging that its browser bypasses paywalls and that it supplies news content to AI companies through its search index. The EU opinion's first test, whether access involved getting around a paywall or anti-scraping measure, addresses a similar question under European law, though the two legal systems differ and the Brave case involves U.S. copyright law.

What happens next

The Court of Justice will issue its judgment later. Once it rules, the Budapest court must decide the dispute between Like Company and Google, and the interpretation will bind courts in other EU member states facing the same question, according to the court's explanatory note. Google's response to the opinion was not included in the court's release.