Edition · Tue 03 Nov 2026
AI Desk
Can a Chatbot Take the News? An EU Court Adviser Says Google's Gemini Might Cross the Line
Advocate General Maciej Szpunar said Google's Gemini may infringe press rights when retrieving articles for answers, while training AI on protected material is allowed under the text-and-data-mining exception.
By Rachel · Edited by Chanan Zevin - Chief Editor and Head of Desks
Published
A Sharp Divide Between Training and Answers
On Tuesday, Nov 3, an adviser to Europe's highest court said that the use of press publications by Google's artificial intelligence chatbot Gemini may in certain circumstances infringe copyright rules. Advocate General Maciej Szpunar advised the court that European protections for news publishers can be breached when an artificial intelligence system pulls material directly from news sites to generate conversational replies for users.
A Sharp Divide Between Training and Answers
The formal guidance from Szpunar was issued in a case brought by Hungarian website operator Like Company against Google, The Jerusalem Post reported. While the dispute centres on Google's artificial intelligence chatbot, the legal assessment delivered to the court separates the way systems collect information to train large language models from the way those tools retrieve and present publisher reporting to the public.
A Sharp Divide Between Training and Answers
This legal boundary matters because it defines how artificial intelligence platforms can interact with journalism, establishing that systems cannot freely extract and reuse protected press content under the guise of answering user queries. While developers retain the freedom to train large language models on protected material, deploying a chatbot that retrieves news articles and reuses their specific expression to answer users creates direct copyright exposure across the European Union.
Infringement Risks When Gemini Pulls From News Sites
Szpunar addressed the mechanics of how chatbots handle digital news reporting when interacting directly with users. "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," Advocate General Maciej Szpunar said.
Infringement Risks When Gemini Pulls From News Sites
At the same time, the adviser set clear limits on what press publishers can legally claim under existing European copyright rules, The Jerusalem Post reported. Szpunar stated that copyright rules do not protect the "mere facts" reported by press publications, explaining that legal protection applies to "only the specific form of expression" created by journalists and editors.
Infringement Risks When Gemini Pulls From News Sites
Under that analysis, a chatbot remains free to convey factual information found within news stories, because the underlying facts themselves enjoy no copyright protection. However, when an artificial intelligence tool reuses a publisher's specific phrasing or creative text to formulate its answer, the system enters territory that may infringe the press publisher's legal rights.
Training AI Falls Under European Copyright Exception
While user-facing answers face potential copyright restrictions, the adviser took a permissive stance toward the broad ingestion of data required to build artificial intelligence models. Szpunar said the use of material protected by copyright to train large language models is "freely permitted under the text and data mining exception."
Training AI Falls Under European Copyright Exception
By categorising artificial intelligence model training under the text and data mining exception, The Jerusalem Post reported, Szpunar affirmed that developers can legally analyse and process protected publications during foundational training phases. This finding creates a structural division under European law between developing the underlying technology and operating a chatbot that actively surfaces published journalism to users.
Judges Weigh Like Company's Hungarian Challenge
Szpunar delivered his legal conclusions while advising Europe's highest court in proceedings brought by Hungarian website operator Like Company against Google. The Hungarian operator's legal challenge prompted the court to evaluate how Gemini's automated retrieval processes interact with intellectual property standards established for press publishers.
Judges Weigh Like Company's Hungarian Challenge
Although Szpunar's statement represents an advisory opinion rather than a binding verdict, opinions of the advocate-general are usually followed by the court's judges. The court will now consider the advocate-general's guidance as it prepares a definitive judgment on the dispute between Like Company and Google.