EU Court Adviser Says Google Gemini's Use of Press Content May Infringe Copyright

Google's use of press stories in its Gemini chatbot can infringe European copyright rules in certain circumstances, Advocate General Maciej Szpunar advised the Court of Justice of the European Union in Luxembourg on Tuesday, November 3, 2020 — a non-binding opinion that the court's judges usually follow.

The opinion arrives in a case brought by Hungarian website operator Like Company against Google's artificial intelligence chatbot Gemini. A ruling against Google could force the tech giant to pay fees for using publishers' content, and it could send a lifeline to online media. Publishers such as Like Company have watched traffic dwindle as AI-generated news summaries increasingly replace traditional search results that included hyperlinks to their sites.

What did the Advocate General actually decide?

Szpunar drew a line between two very different uses of press material. He addressed both.

"When an LLM-driven chatbot retrieves a publication from a press publisher's website, reuses that publication to generate an answer, and provides that answer to the user, the press publisher's right may be infringed in certain circumstances," Szpunar said, according to a court statement.

He identified two specific triggers for infringement:

• Chatbots obtain access to publications by unlawfully circumventing paywalls or other technical barriers.
• Chatbots reuse "the form of expression of that publication in a recognisable manner."

But he set a firm limit on the claim's reach. Copyright rules do not protect the "mere facts" reported by press publications, he said, but "only the specific form of expression". Facts remain free for anyone to use. Wording does not.

What about training the AI models themselves?

Here the news is less favourable to publishers. The use of copyright-protected material to train large language models — AI systems that digest vast amounts of text to generate human-like responses and summarize information — is "freely permitted under the text and data mining exception", Szpunar said.

That distinction matters. It means publishers may have a claim over what a chatbot serves back to users, yet no claim over what the model learned from during training. The exception could swallow much of the practical remedy.

What happens next?

Opinions of the advocate general are not binding. The Court of Justice's judges usually follow them, however, and a final ruling in the Like Company case will now follow in due course.

For practitioners, the consequences are concrete. If the court adopts the opinion, publishers gain a reproduction-right claim against chatbot operators whose systems bypass paywalls or reproduce expressive content recognisably in answers. Platforms, in turn, face exposure on the output side while retaining the safety of the text and data mining exception on the training side. Litigation strategy will turn on evidence: how access was obtained, and how closely the answer mirrors the original expression. Facts alone will not sustain a claim.

In plain words: chatbot answers built on stolen access or recognisable copying may cost Google money; training on publishers' archives remains legal.