The Pulse
EU Court Adviser Says Gemini Answers May Breach Publisher Rights
EU Advocate General Maciej Szpunar says Google Gemini may infringe press publishers’ rights when it reuses articles in answers under certain conditions. He distinguishes those answers from model training, which he says generally falls under
AI.info Team · Written with AI · fact-checked ·
Hungarian publisher Like Company says Google’s Gemini chatbot reused its articles without permission; Advocate General Maciej Szpunar says the evidence does not show those articles trained Gemini, but chatbot answers may still infringe publishers’ rights. In an opinion delivered on November 3, 2026, Szpunar advised the Court of Justice of the European Union to focus on what Gemini retrieves and gives users, rather than decide hypothetical questions about training.
His proposed test turns on how the chatbot accessed an article and how closely its answer reproduces the article’s expression. The opinion is not a judgment, and the court’s judges will decide the case later.
Szpunar draws a line at Gemini’s answers
“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 writes in the court’s account of his opinion.
The adviser says the publisher’s right does not cover the facts reported in an article; it protects the publication’s specific form of expression. His proposed assessment asks whether the chatbot gained access by circumventing effective technical restrictions, such as a paywall or measures against web-scraping. It also asks whether an answer merely reports facts or reuses recognizable expression so extensively that a user no longer needs to read the original article.
For repeated answers, the opinion identifies a separate issue: the right to make a publication available to the public may be infringed when its form of expression is systematically recognizable across responses to related questions. Szpunar also says a chatbot provider may be liable when such uses are inherent to its business model or when it fails to take appropriate measures to counter them credibly and effectively.
Training questions lack evidence in Like Company’s case
Like Company, a Hungarian operator of web portals, alleges that Google used its articles to train Gemini’s large language models and repeatedly reproduced or made them available through the chatbot between June 13, 2023, and May 31, 2024. The Hungarian court referred four questions to the EU court: two about model training and two about chatbot answers.
Szpunar advised the judges to declare the training questions inadmissible because there is no evidence Like Company’s articles were used to train Gemini. He also says training is not an “online use” of press publications, making those questions hypothetical in this dispute. Separately, as a view offered for completeness, he says using copyrighted works to train a large language model is, as a rule, permitted under the EU text-and-data-mining exception.
Google disputes that Gemini replicates articles
A Google spokesperson told Agence France-Presse that Gemini “learns linguistic patterns to create original outputs rather than replicating content,” and pointed to publisher controls that allow rights holders to opt out. The company also welcomed Szpunar’s view on AI training, AFP reported.
That response addresses the distinction at the center of the opinion: training a model and drawing on an article while generating an answer are different uses. Szpunar’s advice does not establish that Gemini infringed Like Company’s rights; the court still has to rule on the EU-law questions, and the Hungarian court must resolve the underlying dispute.
The CJEU’s judges have the final say
Advocates General offer independent legal opinions to the Court of Justice, but their recommendations do not bind its judges. The judges are now deliberating, and the court says its judgment will come at a later date. The Hungarian proceedings will then continue under the EU court’s interpretation of the law.
The case, C-250/25, is the first generative-AI dispute to reach the Court of Justice, according to the court’s announcement. Its eventual ruling will address whether real-time retrieval and reuse of press articles in chatbot answers falls within publishers’ reproduction and public-availability rights—not whether the facts in a news article belong exclusively to its publisher.