🔍 Read the full analysis: What Is At Stake In Europe’s AI Chatbot And Copyright Debate? on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
On 3 November 2026, Advocate General Maciej Szpunar advised the EU Court of Justice that AI chatbots which retrieve and reuse press articles may infringe publishers’ rights in specific circumstances. He said the test turns on unlawful access or answers that reproduce recognisable expression, not simply on reporting the same facts. The opinion is non-binding, and the underlying case concerns Google’s Gemini chatbot and a Hungarian publisher.
Advocate General Maciej Szpunar advised the European Union’s top court on 3 November that an AI chatbot may infringe press publishers’ rights if it obtains an article by bypassing access restrictions or reproduces its distinctive expression so closely that users can skip the original. The non-binding opinion in Like Company v Google Ireland (Case C‑250/25) addresses how EU press-publisher protections apply to chatbots such as Google’s Gemini, but it is not the court’s final ruling.
The dispute began after Hungary-based publisher Like Company challenged Google over Gemini’s response to a request to summarise a Hungarian news article about a singer’s reported plan to bring dolphins to Lake Balaton. The Advocate General’s opinion says the example answer did not appear to reproduce recognisable wording from the article. It also says there was no indication that Google had bypassed a paywall or other access barrier. On those facts, Szpunar considered that the conditions for infringement did not appear to be met.
The proposed test distinguishes an article’s facts from its form of expression. A chatbot’s account of the same events does not, by itself, infringe the publishers’ related right, according to the opinion. The result could differ if a system accessed content unlawfully or reproduced expression so recognisably that a reader could dispense with the original publication. Szpunar also referred to safeguards that Google appeared to have in place in the example before him.
Like Company’s wider case alleges that Google used its articles to train Gemini and repeatedly reproduced or made them available through the chatbot between 13 June 2023 and 31 May 2024, according to the EU court’s press release. But Szpunar advised the Court of Justice to treat the two questions about training as inadmissible: the case material did not establish that the articles were used in training. The Budapest court referred four questions to Luxembourg, covering both training and chatbot answers.
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
The Line Between Facts and Copying
The opinion addresses how EU press-publisher protections may apply when AI services retrieve and present news in generated answers. If the Court of Justice adopts the proposed approach, the relevant factors would include whether a service bypassed access restrictions or reproduced protected wording so closely that users could dispense with the original publication. The opinion does not extend publishers’ rights to the underlying news facts or say that every AI-generated summary requires a licence.
For chatbot operators, the proposed test focuses on how content is retrieved and presented. Whether a system circumvents access restrictions and how closely its answer resembles a source may be relevant in disputes. The opinion does not prescribe specific product changes or determine how other services should handle news content.
The case concerns the EU right created for online uses of press publications, including news aggregation. The Court of Justice’s interpretation will address whether that right also applies to chatbot answers that retrieve articles in response to a user’s request.
Popular right now
TCL 65-Inch T7 Series 4K QLED Smart Google TV, 144Hz (65T7)
Roborock Robot Vacuum and Mop Combo with Self-Empty Dock
Dyson Big Ball Multi Floor Canister Vacuum, Yellow
As an affiliate, we earn on qualifying purchases.
From News Aggregators to Gemini
The dispute concerns the press publishers’ related right introduced under Article 15(1) of the EU Copyright in the Digital Single Market Directive, adopted in 2019. The opinion recalls that lawmakers designed the right to let publishers control or license certain online uses of extracts, including by services such as Google News. The present reference asks whether that protection also reaches answers generated by a chatbot that retrieves articles in response to a user’s request.
Like Company operates Hungarian news websites and brought its case before a court in Budapest. That court sent four questions to the Court of Justice of the European Union, which interprets EU law for national courts. Nine EU governments and the European Commission submitted written observations, and the hearing took place on 10 March 2026, according to the Advocate General’s opinion.
On training, Szpunar’s main conclusion was procedural: he said the questions should not be answered because the record did not show that the disputed articles had been used to train Gemini. The dolphin article appeared in July 2023, while Gemini 1.0 was released in December 2023; the opinion considered training on that article implausible in light of the timing and instead described the chatbot as retrieving the article live. Separately, and expressly for completeness, he said that AI training is generally covered in principle by the EU’s text-and-data-mining exception, subject to rights holders’ ability to opt out in a machine-readable form. That discussion is not a ruling that the alleged training occurred.
The Court Has Yet to Rule
An Advocate General’s opinion is advice, not a binding judgment. The Court of Justice may follow it, adapt its reasoning or reach a different conclusion. No date for the final judgment has been set in the source material, and the Budapest court will still need to resolve the underlying dispute under the EU court’s interpretation.
The record described in the opinion does not establish that Like Company’s articles were used to train Gemini. The training discussion therefore does not settle whether any particular material was included in a model’s training data, nor does it decide a specific publisher’s claim under the text-and-data-mining exception. The facts about the chatbot’s retrieval and answer also come from the particular example considered in this case; the opinion is not a finding about every version of Gemini or every chatbot.
It remains uncertain how the Court of Justice will define the practical threshold for a recognisable reproduction, how the reasoning may apply to other AI search and answering products, and what evidence national courts will require to show that a system bypassed access restrictions or reproduced protected expression.
Judgment and Case Return
The next major step is the Court of Justice’s judgment in Case C‑250/25. The court has not announced a decision date. The source notes that preliminary references took an average of 16.9 months in 2025, but that historical average does not establish when this case will be decided.
After the EU court answers the referred questions, the Hungarian court will apply its interpretation to Like Company’s claims against Google. That national proceeding will address the dispute on its facts, including the allegations about Gemini’s use of the publisher’s material. Until then, the Advocate General’s proposed limits offer guidance on the legal issues, but do not determine liability.
Key Questions
Has the EU Court of Justice ruled that Gemini infringed copyright?
No. The 3 November opinion is non-binding advice from Advocate General Maciej Szpunar. The Court of Justice has yet to issue its judgment, and the opinion says the infringement conditions did not appear to be met in the example described.
When might a chatbot’s answer infringe publishers’ rights?
Under the Advocate General’s proposed approach, infringement may arise if a chatbot obtains an article by bypassing access restrictions or reproduces its expression so recognisably that users can dispense with the original. Repeating the same facts alone would not be enough under that test.
Did the opinion find that Google trained Gemini on Like Company’s articles?
No. Szpunar advised that the training questions were inadmissible because the case material did not establish that the articles were in Gemini’s training data. His separate discussion of training and text-and-data mining was offered for completeness, not as a finding that the alleged use occurred.
What happens after the Advocate General’s opinion?
The Court of Justice will issue a judgment interpreting the EU law questions; no date has been set. The Hungarian court will then use that interpretation to decide the publisher’s case against Google.
Source: ThorstenMeyerAI.com
