In the first generative AI case to reach the European Union’s top court, its adviser says chatbots like Google’s Gemini can break publishers’ rights when their answers copy the shape of a news story so closely that nobody needs to read the original. Training AI on articles, he adds, is generally allowed.

By HahuPrime News Desk | November 3, 2026
LUXEMBOURG — It began with a simple request. Someone asked Google’s Gemini chatbot to sum up a news article published by a Hungarian news website. According to the publisher, Like Company, the answer that came back carried over the key content of its piece, closely enough that the reader had little reason to click through.
That small exchange has now become the first generative AI case to reach the Court of Justice of the European Union. On Tuesday, the court’s Advocate General, Maciej Szpunar, gave his opinion, and it cut both ways for the technology industry.
“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 the court’s press release.
At the same time, he said that using copyrighted material to train a large language model is, as a rule, “freely permitted under the text and data mining exception” in EU law.
Facts are free, the wording is not
Like Company runs several Hungarian web portals. It sued Google Ireland before the Budapest Regional Court, arguing that Google used its articles to train Gemini and repeatedly reproduced them through the chatbot between June 13, 2023 and May 31, 2024. The Hungarian judges paused the case and sent four questions to Luxembourg.
Szpunar advised the court to throw out the two questions on training. There was no evidence that Like Company’s articles had been used to train Gemini, he said, and training is not an “online use” of a press publication, so those questions were hypothetical.
The heart of his opinion is about what happens when a chatbot answers a question in real time. Gemini, like many AI assistants, uses a technique called retrieval-augmented generation. In plain terms, it goes out to the web while you wait, pulls in fresh pages, often news articles, and uses them to write its reply.
Szpunar drew a clear line. The press publisher’s right, created by the EU’s 2019 copyright directive, “does not protect the mere facts reported,” he said, but “only the specific form of expression” of an article. Its aim is to protect the investment that goes into journalism, not to give publishers a monopoly on the news.
So two things matter. First, did the chatbot get the article lawfully, or did it get around a paywall or another barrier meant to block scraping? Second, does the answer simply report facts, or does it reuse the article’s wording in a recognisable way, to the point where the reader no longer needs the original? If the second happens systematically, he said, the publisher’s right of making content available to the public is, as a rule, infringed.
He also addressed who pays. The chatbot provider is liable, he said, where such copying is built into its business model, or where it has not put in place measures to counter it “credibly and effectively.”
Google welcomes half of it
Google picked out the part it liked. A company spokesperson told AFP that “Gemini learns linguistic patterns to create original outputs rather than replicating content, supported by extensive publisher controls that give rightsholders clear choice and the ability to opt out.” The spokesperson also welcomed “the advocate general’s opinion recognising that EU law permits AI training on publicly available information.”
The opinion is not binding. Judges at the Court of Justice now begin their deliberations, and AFP reports a judgment is expected in around six months. Reuters notes that the court usually follows its advocates general. If it does here, the ruling will apply across all 27 EU countries, and the Hungarian court will decide the case itself in line with it.
The stakes are commercial and very real. A ruling against Google that forces payment for news content could throw a lifeline to online media like Like Company, Reuters reported. They have watched their traffic fall as AI-written summaries replace the old list of search results with links.
Jörg Hoffmann, a senior research fellow at the Max Planck Institute for Innovation and Competition, wrote earlier this year that the case is not only about copyright. If users get everything they need from an AI summary, he argued, “they may never click through to the original article,” which could “decimate the advertising-based revenue model on which many publishers depend.”
Why newsrooms outside Europe are watching
The court in Luxembourg only rules on EU law. But its answers tend to travel, because the same chatbots serve readers everywhere, and regulators elsewhere are asking the same questions.
Africa has already had its own version of this fight. In November 2025, South Africa’s Competition Commission published the final report of its Media and Digital Platforms Market Inquiry. It required Google to provide R688 million over five years for South African media, including an AI innovation fund for publishers, and to give news outlets controls to opt their content out of its AI products.
For smaller newsrooms in the Horn of Africa and elsewhere, which often rely on search traffic and have no lawyers to take on a tech giant, the Luxembourg case offers a useful test. It separates three things that are often blurred together: training a model, repeating facts, and copying a journalist’s work so closely that the reader never visits the source. Only the last one, Szpunar suggests, needs the publisher’s permission.
Where exactly does a summary end and a copy begin? That is for the judges now. For the people who write the news, and the many more who now read it through a chatbot, the answer will shape who gets paid for the story.
For more on how AI is changing work and information, read our reports on why Sweden blocked ChatGPT and Claude on government computers, Google’s Gemini 4 Carbon model built only for cyber defenders and Google’s AI satellite that wrote a haiku from orbit.
Sources
- Court of Justice of the European Union: Press release No 155/26, Opinion of Advocate General Szpunar in Case C-250/25, Like Company
- Reuters: Google AI chatbot’s use of press items may infringe copyright rules, EU court adviser says
- AFP (via Deccan Chronicle): Google Gemini may break press copyright rules, EU legal adviser
- Paperjam: Gemini may infringe the rights of newspaper publishers
- NUS Law Research Blog: Jörg Hoffmann, AI, Copyright and Media Pluralism: What Like Company v. Google Means for the Future of EU Law
- Competition Commission South Africa: Media and Digital Platforms Market Inquiry final report, November 2025