🔍 Read the full analysis: AI Chatbot Use Of Publisher Content: What Companies Need To Consider on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
On Nov. 3, 2026, Advocate General Maciej Szpunar advised the EU Court of Justice that chatbots retrieving and reusing news articles may infringe publishers’ rights in specific circumstances. His proposed test focuses on unlawful access and answers that reproduce an article’s distinctive expression; the opinion is nonbinding, and he said the test does not appear to be met in the dispute over a Gemini summary.
Advocate General Maciej Szpunar advised the European Union’s Court of Justice on Nov. 3 that chatbots retrieving online news articles and reusing them in answers may infringe press publishers’ rights in certain circumstances. His nonbinding opinion in Case C‑250/25 says the key questions are whether a chatbot accessed an article unlawfully, such as by bypassing a paywall or anti-scraping measure, or reproduced its distinctive wording so extensively that users could avoid reading the original.
The dispute began with a Hungarian news article about a proposal to bring dolphins to Lake Balaton. Like Company, which operates Hungarian news portals, sued Google in a Budapest court. The company alleges that Google used its articles to train Gemini and repeatedly reproduced and made them available through the chatbot from June 13, 2023, to May 31, 2024. Those allegations have not been established by the opinion.
Szpunar advised that the test for chatbot answers should distinguish facts from the way a report expresses them. Repeating information reported in an article is not, by itself, enough to infringe the publisher’s right, he said. The relevant question is whether protected expression is recognisable in an answer to the extent that a reader can dispense with the original publication.
On the facts described in the opinion, Szpunar said those conditions do not appear to be met: the article was not protected by access barriers, and the example answer did not appear to reproduce recognisable wording. He also referred to safeguards Google appears to have in place. The Hungarian court, not the Advocate General, will decide the dispute.
Infographic 01 · Where the line runs
When a chatbot answer infringes a publisher's right
Advocate General Szpunar's test in Case C‑250/25 for chatbots that retrieve news articles in real time and reuse them in their answers.
Stays free
- Reporting the same facts. Nobody owns the news itself (pts 199–200).
- A short factual summary, presented differently (pt 208).
- An overview built from several sources, restructured, with links back (pts 208, 234).
- Individual words or very short extracts (pt 201).
Can infringe
- Getting around a paywall or anti-scraping block to fetch the article (pt 188).
- An answer that keeps the article’s structure and phrasing, even paraphrased, so readers can skip the original (pts 202, 205, 207).
- Systematically recognisable articles in answers to related questions: making available to the public (pts 213, 240).
The test, step by step
For a chatbot that fetches a news article live (RAG) and answers a user
- Is it an online use by a service in the EU?Yes for RAG answers to EU users, wherever the servers are (pts 165–169).
- Did the bot get in lawfully?No circumvention of effective technical measures (pt 188).
- Is the article’s form of expression recognisable in the answer, to the point of substitution?Side-by-side comparison; facts do not count (pts 196–209).
- Is the provider responsible?If inherent to its business model, or no reasonable measures such as respecting blocks and output filters (pts 221–222).
- Temporary copies during retrieval are then exempt under Art. 5(1) InfoSoc; the TDM exception does not apply to RAG (pts 179–189).
An Advocate General’s opinion, not a judgment. General information, not legal advice.
Sources: Opinion of AG Szpunar, C‑250/25 (provisional text), CURIA press release No 155/26 (3 Nov 2026). Point numbers refer to the provisional text.
Powered by Thorsten Meyer AI
A Narrower Test for AI Answers
The opinion could shape how companies build AI chatbots and search tools that retrieve news in real time. Under Szpunar’s proposed approach, a service’s legal exposure would depend not simply on whether it uses information from a publisher, but on how it accessed the material and how much of its protected expression appears in the generated answer.
For publishers, the opinion preserves a possible route to challenge paywall circumvention and near-copying, while rejecting a general right to control reporting of facts. For AI companies, it points to practical issues around respecting access restrictions and preventing answers from substituting for original articles. These are implications of the proposed test, not a binding ruling or a guarantee of how the court will decide.
Readers may also be affected if services change how they summarize and link to reporting. The opinion does not set product requirements or predict whether companies will use more links or rely on multiple sources. Those decisions will depend on the eventual judgment, other applicable law, and companies’ own policies.
Popular right now
Dyson Ball Animal 3 Upright Vacuum
Dyson V8 Cyclone Cordless Vacuum
Ninja Flip Toaster Oven & Air Fryer, Large Capacity, Stainless, SP151
As an affiliate, we earn on qualifying purchases.
How the Gemini Dispute Reached Luxembourg
EU law has granted press publishers a related right since the 2019 Copyright in the Digital Single Market Directive. The opinion says the measure was adopted in part to let publishers control and license online uses of extracts by news aggregators such as Google News. The current case asks how that right applies when an AI chatbot draws on an article to respond to a user.
The Budapest court referred four questions to the Court of Justice: two about model training and two about chatbot answers. Nine EU governments and the European Commission submitted written observations, and the hearing took place on March 10, 2026, according to the opinion. The example involving Gemini concerned a user asking for a summary of the dolphin article; the source account says the chatbot retrieved the article from the live web.
Szpunar advised the court not to answer the training questions because the record did not show that the disputed articles were included in Gemini’s training data. He considered it implausible that the July 2023 dolphin article was used to train Gemini 1.0, released in December 2023, given the time needed to gather data and train a model. He also reasoned that training on dedicated computers is not the kind of “online use” covered by the press publishers’ right.
Separately, and expressly as a discussion for completeness, he said model training may in principle fall under the EU’s text-and-data-mining exception, subject to rights holders’ ability to opt out in a machine-readable form. That discussion was not his answer to the case’s training questions: he considered those questions inadmissible because the necessary factual basis was missing.
Questions Await the Court’s Judgment
The Advocate General’s opinion is not binding. The Court of Justice may follow it, modify its reasoning, or reach a different conclusion. No date for the judgment has been set in the source material, and the Budapest court will still need to apply the EU court’s eventual interpretation to the case.
The record described by Szpunar does not establish that Like Company’s articles were used in Gemini’s training data. Nor does his view on text-and-data mining settle every question about AI training: he discussed that exception separately after advising that the training questions should not be answered on the available facts. The opinion also does not resolve how its proposed test would apply to other chatbot designs, access controls, or answer formats.
Awaiting the EU Court’s Ruling
The Court of Justice will issue the judgment in Case C‑250/25, after which the Hungarian court can resolve the dispute using the EU court’s interpretation. The source material gives no date for that judgment. Until it is delivered, Szpunar’s proposed limits remain guidance from an independent legal adviser rather than a final rule.
Companies using chatbots to retrieve publisher content can track the ruling for clarification on access restrictions, answer similarity, and publishers’ rights. Publishers and AI providers will also need to distinguish the court’s eventual decision on chatbot answers from the separate, unresolved legal and factual questions around model training.
Key Questions
Has the EU Court of Justice ruled that Gemini infringed publishers’ rights?
No. Advocate General Szpunar issued a nonbinding opinion on Nov. 3, 2026. The Court of Justice has yet to deliver its judgment, and the Hungarian court will decide the underlying dispute.
When might a chatbot answer infringe a publisher’s rights under the opinion?
Szpunar’s proposed test identifies two relevant circumstances: the chatbot accessed an article unlawfully, such as by bypassing a paywall or anti-scraping measure, or its answer reproduced the article’s distinctive expression so recognisably that users could avoid the original. Reporting the same facts alone would not be enough under that approach.
Did the opinion find that Gemini was trained on Like Company’s articles?
No. Szpunar said the record did not establish that the articles were used for training and advised the court to treat the training questions as inadmissible. His separate discussion of text-and-data mining was offered for completeness, not as a factual finding about Gemini.
What should AI companies watch for next?
Companies should watch for the Court of Justice’s final interpretation of lawful access and reproduction of protected wording. The opinion is not binding, so its proposed test may change before the court rules.
Source: ThorstenMeyerAI.com
