🔍 Read the full analysis: What Companies Should Watch In The EU’s AI Copyright Debate on ThorstenMeyerAI.com
Get the little things that make your day 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 using online news articles may infringe publishers’ rights if they bypass access restrictions or reproduce recognisable expression. He said the available facts did not appear to meet that test in a dispute over a Gemini summary, and the court has not yet ruled.
Advocate General Maciej Szpunar advised the EU Court of Justice on Nov. 3 that AI chatbots may infringe press publishers’ rights when they retrieve news articles and reuse them in answers, if they bypass access restrictions or reproduce recognisable expression. His non-binding opinion in a dispute involving Google’s Gemini said the available facts did not appear to meet those conditions, leaving the Hungarian court to decide the case after the EU court rules.
The dispute, Case C‑250/25, Like Company v Google Ireland, began with a Hungarian news article published in July 2023 about a singer’s reported plan to bring dolphins to Lake Balaton. Like Company, which runs Hungarian news portals, says Google used its articles to train Gemini and that the chatbot repeatedly reproduced or made them available in answers between June 13, 2023, and May 31, 2024. Those are the publisher’s claims, not findings by the court.
The example before the courts involved Gemini summarising the dolphin article in Hungarian. Szpunar’s proposed test distinguishes the article’s underlying facts from its protected expression: repeating the same facts alone would not be enough. Infringement could arise, he said, if a chatbot obtained an article by getting around a paywall or anti-scraping block, or if its answer carried wording or expression so recognisable that a reader could avoid the original publication.
Szpunar said the conditions did not appear to be present in the example. The source material says the article was accessible without relevant barriers, the sample answer did not reproduce recognisable wording, and Google appeared to have safeguards. The Court of Justice has not issued a judgment, and the Budapest court—not the Advocate General—will decide the underlying dispute in light of the eventual ruling.
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 Compliance Line for AI Search
The opinion gives companies building chatbots and AI search tools a practical issue to monitor: whether their systems access restricted content and whether answers substitute for the articles they draw on. The proposed test focuses not simply on whether an AI system refers to a news report, but on how it accessed the material and how much of its protected expression appears in the response.
For publishers, that approach could support challenges where services bypass technical restrictions or reproduce distinctive expression, while stopping short of giving them control over the underlying facts. For AI companies, the opinion points to the potential value of respecting access barriers, reviewing output for close copying, and linking users to original reporting. These are implications of Szpunar’s proposed reasoning, not final legal requirements established by a court judgment.
The outcome may also affect how online services negotiate access and licensing. The EU press publishers’ right was introduced in part to give publishers control over use of extracts by online services, including news aggregators. Whether that right applies to chatbot answers—and under what conditions—is now before the EU court.
Products Worth Considering
Withings BPM Connect Wi-Fi and Bluetooth Blood Pressure Monitor
Hyperice Hypervolt Go 3 Percussion Massage Gun with Carry Case
Garmin Forerunner 55 GPS Running Smartwatch, Black
As an affiliate, we earn on qualifying purchases.
How the Gemini Dispute Reached Luxembourg
Like Company brought its case before a Hungarian court, which referred four questions to the Court of Justice: two concerning AI training and two concerning chatbot answers. According to the court’s press release, Like Company alleged that Google used its articles to train Gemini and then reproduced or made them available to users through the chatbot. Nine EU governments and the European Commission submitted written observations, and a hearing took place on March 10, 2026.
The questions concern the press publishers’ right created by Article 15(1) of the EU Copyright in the Digital Single Market Directive, adopted in 2019. Szpunar advised the court to treat the two training questions as inadmissible because the case materials did not establish that Like Company’s articles had been used for training. He also reasoned that training on dedicated computers is not the kind of online use covered by that particular publishers’ right.
Separately, and “for the sake of completeness,” the Advocate General discussed the EU text and data mining exception. He said training generally may fall within that exception, subject to the rules that allow rights holders to reserve their rights in a machine-readable form. That discussion was not his operative answer to the case’s training questions; describing the opinion as a ruling that definitively approves AI training would overstate what it says.
The Court Has Yet to Decide
The Advocate General’s opinion is advisory, not binding. The Court of Justice may follow it, adjust its reasoning, or reach a different conclusion. No date for the judgment has been set, and the Hungarian court will still need to resolve the national case after the EU court answers the referral.
The training allegations also remain unproven in the material described by Szpunar. He said the referral contained no evidence that Like Company’s articles were included in Gemini’s training data, and treated the training questions as hypothetical. His additional discussion of text and data mining does not establish whether any particular article was used, whether a rights holder opted out, or whether a specific model’s training complied with copyright law.
Nor does the opinion settle how its proposed test would apply across different products, types of content, or access controls. The available example concerns one summary and one article; the source material does not establish how the court will assess other chatbot outputs or what safeguards will suffice.
Awaiting the EU Court’s Judgment
The next major step is the Court of Justice’s judgment in Case C‑250/25. The court has not announced a decision date. Its 2025 statistics put the average time for preliminary references at 16.9 months, but that is a historical average, not a timetable for this case.
Once the EU court answers the referred questions, the Hungarian court is expected to apply that interpretation when deciding Like Company’s claims. Until then, companies can track the distinction in Szpunar’s opinion between factual reporting and protected expression, while treating his proposed approach as guidance rather than settled law.
Key Questions
Has the EU Court of Justice ruled in the Gemini case?
No. On Nov. 3, 2026, Advocate General Maciej Szpunar issued a non-binding opinion. The court’s judgment is still pending.
When could a chatbot’s use of a news article infringe publishers’ rights?
Under Szpunar’s proposed approach, infringement may arise if a chatbot gets around access restrictions or reproduces recognisable expression to an extent that users can avoid the original publication. Reporting the same facts alone would not be enough.
Did Szpunar find that Google trained Gemini on Like Company’s articles?
No. He said the referral did not establish that the articles were used for training and advised treating those questions as inadmissible. His separate comments on text and data mining did not decide whether Google used any specific article.
What should AI companies monitor while the case is pending?
Companies can review how their systems access online material, whether they respect paywalls and anti-scraping restrictions, and whether generated answers reproduce protected expression. These are practical considerations drawn from the opinion, not final court-ordered requirements.
Source: ThorstenMeyerAI.com
