🔍 Read the full analysis: Facts May Be Free, But Publishers’ Writing Has Rights: The AI Question 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 Nov. 3, 2026, Advocate General Maciej Szpunar advised the EU Court of Justice that chatbots using press articles in answers may infringe publishers’ rights if they bypass access barriers or reproduce recognisable expression. In the Gemini case over a Hungarian dolphin story, he said those conditions do not appear to be met; the opinion is not binding, and the court has not set a judgment date.
Advocate General Maciej Szpunar advised the European Union’s Court of Justice on Nov. 3 that chatbots can infringe press publishers’ rights when they retrieve and reuse news articles, but only under specific conditions. In the case involving Google’s Gemini and a Hungarian article about dolphins at Lake Balaton, Szpunar said the available facts do not appear to satisfy his proposed test, leaving the final decision to the court.
The dispute is Case C‑250/25, Like Company v Google Ireland. Like Company, which runs Hungarian news portals, sued Google in a Budapest court. It alleges that Google used its articles to train Gemini and reproduced or made them available through chatbot answers between June 13, 2023, and May 31, 2024. Those are the publisher’s claims, not findings by the Court of Justice.
The case reached Luxembourg after the Hungarian court referred four questions: two on AI training and two on chatbot answers. The Court of Justice describes it as the first generative AI case to reach the court. Szpunar’s opinion says a chatbot’s use may infringe the EU press publishers’ right if it obtained an article by bypassing a paywall or anti-scraping measure, or if its answer reproduces the article’s distinctive expression so closely that users can dispense with the original publication.
On the example answer about the dolphin story, the adviser found no apparent access barrier and no recognisable reproduction of the article’s wording. He also pointed to safeguards that Google appears to use. His assessment is preliminary: the Hungarian court will apply the law to the case, and the EU court’s eventual judgment will provide the binding interpretation.
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
Where Chatbot Summaries Cross the Line
The proposed test separates the reuse of news facts from the reuse of a publisher’s protected expression. A chatbot is not barred, on this reasoning, from reporting the same event or facts found in a news article. The concern arises when it has obtained the article improperly or when its answer preserves enough of the article’s particular wording or form to substitute for reading the original.
That distinction could affect publishers, AI companies and readers. Publishers may have legal grounds to challenge access that evades technical restrictions or answers that closely replicate their work, while the opinion does not give them control over the underlying facts. AI services may need to account for access controls and how much source-specific language appears in generated answers. For users, the standard could leave room for summaries and synthesis, while encouraging answers that point to original reporting rather than replace it.
The opinion is guidance, not a ruling. Its practical effect will depend on the Court of Justice’s judgment and, later, how national courts assess particular systems and outputs.
Products Worth Considering
LG UltraGear 32GS60QC-B 32-inch Curved QHD Gaming Monitor (180Hz, 1ms, 1000R, AMD FreeSync, HDR10)
DJI Osmo Mobile 8 3-Axis Phone Gimbal Stabilizer
DJI Osmo Mobile 8 Gimbal with Mic Mini Transmitter, Infinity Black
As an affiliate, we earn on qualifying purchases.
A Dispute Over Gemini’s Live Retrieval
The example began with a July 2023 Hungarian news report about a popular singer’s reported plan to bring dolphins to a 25-metre pool on the Tihany peninsula at Lake Balaton. A user asked Gemini to summarise the article in Hungarian. Szpunar considered the chatbot’s ability to fetch a live webpage and compose an answer separately from whether the article had been used to train the model.
That distinction matters because the publisher raised both issues. According to the court’s press release, Like Company claimed its articles were used in Gemini’s training data as well as reproduced in responses. The Advocate General said the training questions should be treated as inadmissible because the referral did not establish that the specific articles were in the training data. He noted that the dolphin article appeared in July 2023, while Gemini 1.0 was released in December 2023, and considered it implausible that the article had been included in training given the timing and the months-long process of preparing large models.
Instead, he said the example appeared to involve retrieval-augmented generation: the chatbot searches for relevant web pages and uses retrieved material to form an answer. He also reasoned that training on dedicated computers is not the kind of online use covered by the press publishers’ right. Separately, and expressly for completeness, he said AI training may in principle fall within the EU’s text-and-data-mining exception, subject to rights holders’ ability to opt out in a machine-readable form. Those observations do not decide whether Google trained Gemini on these articles.
“repeatedly reproduced and made them available to the public through the Gemini chatbot”
— Like Company, as summarised in the Court of Justice press release
The Court Has Not Ruled Yet
Szpunar is one of the Court of Justice’s independent legal advisers. His opinion is not binding; the judges will issue the ruling, and no date has been announced. It is not yet known whether the court will adopt his proposed test, modify it or reach a different conclusion.
The underlying factual questions also remain unresolved. The opinion does not establish that Like Company’s articles were used to train Gemini, and the publisher’s allegations about chatbot reproductions have not been finally tested. The Budapest court must decide the dispute after receiving the EU court’s interpretation. The scope of any machine-readable opt-out and how the proposed standard applies to different chatbot designs may also require further decisions.
A Binding Judgment Still to Come
The Court of Justice will now consider the Advocate General’s opinion before issuing its judgment in Case C‑250/25. The court has not set a timetable. Its preliminary rulings took an average of 16.9 months in 2025, according to its statistics, but that figure does not establish when this case will be decided.
After the EU court answers the referred legal questions, the Hungarian court will resume the case and determine the dispute between Like Company and Google. Until then, the opinion offers a proposed framework for assessing chatbot access and reuse, rather than a final judgment on Gemini or a general determination that AI training in this case infringed publishers’ rights.
Key Questions
Did the EU Court of Justice rule that Gemini infringed copyright?
No. Szpunar issued a nonbinding opinion, not a judgment. He said the apparent facts in this case do not seem to meet his proposed conditions for infringement.
When could a chatbot’s use of a news article infringe publishers’ rights?
Under the Advocate General’s proposed test, infringement may arise if a service gets around an access barrier, such as a paywall or anti-scraping measure, or reproduces the article’s distinctive expression so closely that users can forgo the original.
Did the opinion find that Google trained Gemini on Like Company’s articles?
No. The adviser said the referral did not establish that the articles were in Gemini’s training data and recommended that the training questions be treated as inadmissible. The allegation remains a claim by Like Company.
What happens next in the case?
The Court of Justice will issue a binding judgment, with no date set so far. The Budapest court will then apply the EU court’s answers to the dispute.
Source: ThorstenMeyerAI.com
