🔍 Read the full analysis: AI Chatbots And Publishers’ Rights: What Companies Need To Know on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
On November 3, 2026, Advocate General Maciej Szpunar advised the EU Court of Justice that chatbots reusing press articles may infringe publishers’ rights in certain circumstances. He drew a distinction between reporting the same facts and reproducing an article’s recognisable expression, but said the available facts in the Gemini case do not appear to meet his test. The opinion is not binding; the court has not set a date for its judgment.
Advocate General Maciej Szpunar advised the European Union’s top court on November 3 that AI chatbots which retrieve and reuse online news articles may infringe publishers’ rights if they bypass access restrictions or reproduce an article’s recognisable expression. In the dispute over Google’s Gemini chatbot, however, Szpunar said the conditions for infringement do not appear to be met on the facts presented. His opinion in Case C‑250/25 is not binding, and the Court of Justice of the European Union (CJEU) will issue the final ruling.
The case was brought by Like Company, which operates Hungarian news websites, against Google Ireland. The publisher alleges that Google used its articles to train Gemini and that the chatbot repeatedly reproduced them in answers between June 13, 2023, and May 31, 2024. Those are claims in the litigation, not findings established by the Advocate General’s opinion.
The dispute arose after Gemini produced a Hungarian summary of an article about a singer’s reported plan to bring dolphins to Lake Balaton. Szpunar’s proposed test focuses on how a chatbot accessed and presented the material. A response could infringe, he said, if the system reached an article unlawfully—for example, by getting around a paywall or anti-scraping measure—or if it reproduced the article’s form of expression so recognisably that a user could do without the original publication.
Simply conveying the same facts or information is not enough under the opinion. Applied to the example at issue, Szpunar said there appeared to be no access barrier, no recognisable wording in Gemini’s answer and safeguards on Google’s part. He therefore did not consider the proposed infringement conditions to appear satisfied in this instance. The Hungarian court will assess the case under the CJEU’s eventual interpretation of EU law.
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
Rules for Chatbot News Answers
The opinion offers a potential framework for companies that operate AI chatbots and AI search tools using current web content. It signals that access practices and the wording of generated responses could matter under the EU press publishers’ right, which has applied since 2019. Companies may need to account for paywalls and anti-scraping controls and avoid answers that function as recognisable substitutes for particular articles.
For publishers, the proposed line could support arguments for licensing when a service reuses their material in a way that substitutes for reading the source. But it does not give publishers ownership of the underlying news facts or a general veto over summaries. The opinion instead distinguishes between facts, which can be reported, and the protected expression used to present them.
The practical effect remains uncertain until the court rules and national judges apply its interpretation. If the CJEU adopts the approach, chatbot operators may have to balance useful, source-based answers against access controls and the risk of reproducing distinctive wording. Publishers and technology companies may also use the ruling to guide negotiations over content access and licences.
Popular right now
Hisense 85 U7 Mini-LED ULED 4K UHD Smart Google TV (85U7SG, 2026)
Breville BES870BTR Barista Express Espresso Machine, Black Truffle
Bose QuietComfort True Wireless Earbuds with Active Noise Cancellation, Black
As an affiliate, we earn on qualifying purchases.
How the Gemini Dispute Reached Luxembourg
Like Company sued Google in Hungary, relying on the press publishers’ right set out in Article 15(1) of the EU’s 2019 Copyright in the Digital Single Market Directive. The right was created to let publishers control or license certain online uses of their publications, including extracts used by news aggregation services. The question referred to the CJEU is whether and how that right applies to chatbot answers.
The Budapest regional court sent four questions to Luxembourg: two about alleged use of articles for AI training and two about answers generated by the chatbot. Nine EU member states and the European Commission filed written observations, and the hearing took place on March 10, 2026. The CJEU described the dispute as the first generative AI case to reach it.
On training, Szpunar’s main position was procedural: he advised the court to treat the two questions as inadmissible because the record did not establish that Like Company’s articles were used to train Gemini. He reasoned that the cited article was published in July 2023, after the training data for Gemini 1.0 would likely have been collected, and that the example was consistent with live retrieval of the page. He also said the publishers’ right concerns online use by services, while model training on dedicated computers does not fit that category.
Separately, “for the sake of completeness,” Szpunar discussed the EU text-and-data-mining exception. He considered that training may in principle fall within that exception, subject to rights holders’ ability to reserve their rights in a machine-readable way. This discussion was not his main answer on training and does not establish that the articles in this case were used for training.
“The first generative AI case to reach the Court of Justice.”
— Court of Justice of the European Union, press release No 155/26
The Court’s Final Test Remains Pending
The Advocate General’s opinion is not a binding judgment. The CJEU may follow it, modify its reasoning or reach a different conclusion. No date for the judgment has been announced, and the Hungarian court will still need to decide the underlying dispute after receiving the EU court’s interpretation.
The record described in the opinion does not establish that Like Company’s articles were used to train Gemini. Szpunar’s comments about text-and-data mining were supplementary, not a finding about what Google did in this case. It also remains for the courts to assess the evidence on how the chatbot accessed the article, what its safeguards did and whether any response reproduced protected expression.
The opinion does not set out a detailed technical threshold for when wording becomes recognisable enough to substitute for an article. How the proposed test would apply across different chatbot designs, kinds of summaries and access controls will depend on the CJEU’s ruling and later cases.
CJEU Ruling and Hungarian Proceedings
The next major step is the CJEU’s judgment in Case C‑250/25; the court has not announced when it will issue it. The ruling will answer the questions referred by the Hungarian court and provide guidance on the scope of the EU press publishers’ right in relation to chatbot answers.
After the CJEU rules, the Hungarian court is expected to apply that interpretation to Like Company’s claims and the evidence about Google’s handling of the articles. Until then, the Advocate General’s proposed distinction between reused facts and recognisable expression is guidance, not a final determination of liability.
Key Questions
Has the EU court ruled that Gemini infringed publishers’ rights?
No. The November 3 opinion was issued by Advocate General Maciej Szpunar and is not binding. He said the conditions for infringement did not appear to be met on the facts described, and the CJEU has yet to issue its judgment.
When might a chatbot answer infringe under the opinion?
Szpunar said infringement may arise if a chatbot gets an article by bypassing an access restriction, such as a paywall or anti-scraping measure, or reproduces the article’s expression so recognisably that the user can do without the original. Reporting the same facts alone would not meet that test.
Did the opinion establish that Google trained Gemini on Like Company’s articles?
No. Szpunar advised the court to treat the training questions as inadmissible because the case record did not establish that the articles were used in Gemini’s training data. His separate comments on text-and-data mining were supplementary and did not decide what happened in this case.
What should companies running AI search or chatbots take from the opinion?
The opinion suggests that companies should pay attention to how their systems access news content and whether answers reproduce distinctive expression. It is not yet a final legal rule; companies should monitor the CJEU judgment and seek qualified legal advice about their specific services.
Source: ThorstenMeyerAI.com
