CJEU adviser says AI chatbot summaries can infringe publisher rights | Coral Wire
Daily news for authors · with CoralGet the brief
Legal · Thu, Nov 5 2026 · Coral Index 35/100

CJEU adviser says AI chatbot summaries can infringe publisher rights

Advocate General Szpunar's nonbinding opinion proposes a side-by-side test for when a chatbot answer counts as reproduction

CJEU adviser says AI chatbot summaries can infringe publisher rights
Image generated for this story · Coral Wire

Advocate General Maciej Szpunar of the Court of Justice of the European Union issued an opinion on Tuesday in the first generative AI case to reach the Court, according to IPWatchdog. The opinion found that AI chatbots retrieving and reusing online press publications in real time to answer user queries "may, in certain circumstances, infringe the rights of press publishers." The opinion is nonbinding on the CJEU.

The case was brought by Like Company, a Hungarian web portals operator, against Google over its Gemini chatbot, in the Budapest Regional Court. Like Company alleged Google used its articles without permission from June 13, 2023 through May 31, 2024, first via Bard and then Gemini, both to train the underlying models and through real-time retrieval during user queries. In one cited example, a user asked Gemini to summarize a Like Company article about a plan to bring freshwater dolphins to Lake Balaton; Gemini's summary included details, such as proposed pool dimensions and claims about public reaction, that the original article did not contain.

Szpunar proposed that a chatbot's reproduction of an article infringes if the article was accessed unlawfully, or if the chatbot's answer retains the article's protected expression recognizably enough, determined through "a side-by-side comparison," that users could read the answer instead of visiting the original. On the separate question of AI training, which he considered hypothetical here since no record showed Like Company's articles trained Gemini, Szpunar offered the provisional view that the EU's text and data mining exception can in principle cover training reproductions, provided developers take steps to reduce the risk of outputs closely resembling specific training works.

Coral’s read

Here's the plain English version of a very dense opinion: if a chatbot's answer still lets you recognize the original article's expression, that's not a neutral summary, that's a copy. The Advocate General calls it the side-by-side test, and he applied it in a case about a Hungarian news site and Google's Gemini, where the chatbot's "summary" invented details the article never had. That alone should make you pay attention to how any AI tool handles your work, not just whether it copies word for word, but whether a reader could skip you entirely and get the gist from the bot. This is one opinion in one case, nonbinding on the court, so there's no ruling to act on yet, just a test worth remembering.

From the Coral Wire edition of Thu, Nov 5 2026 · Coral Index 35/100 (Headwinds)

Spot something wrong?

Every story here links its sources so you can check them. If we have got a fact, a figure or a date wrong, tell us and we will correct the page and say so on it.

Prefer email? info@coralwire.ai

◆ The Daily Brief

Get the wire in your inbox

One short email each morning with the day’s news that actually moves your author business. Free, from the same desk.

From the Coral Hart group

Write it, plan it, launch it