EU court adviser says Gemini’s use of press content may infringe copyright in some cases

An adviser to Europe’s highest court said on November 3 that Google’s AI chatbot Gemini may, in certain circumstances, infringe copyright rules when it uses press publications. The opinion from Advocate General Maciej Szpunar comes in a case brought by Hungarian website operator Like Company against Google.

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 wrote.

He also set the outer edge of protection: copyright does not cover the mere facts reported by press publications, but only the specific form of expression. Reproducing facts and reproducing expression are therefore treated differently.

On training, his reading is friendlier to AI developers. Using copyright-protected material to train large language models is freely permitted under the text and data mining exception, he said, separating the training stage from the output stage rather than treating the whole pipeline as infringement.

The case pits Like Company, a Hungarian website operator, against Google. Opinions of the advocate general are not binding, but the court’s judges usually follow them; a final judgment is expected later.

The practical value of the opinion is that it draws a line inside the EU’s press publishers’ right rather than upholding or rejecting it wholesale. Training on published material sits on the permitted side, while the retrieve, reuse and answer chain has to be assessed on its own facts.

For Gemini and similar assistants, much will depend on how the Court defines those circumstances: how much is retrieved, how much is reused, and whether the answer substitutes for a visit to the original article. Those variables will shape licensing talks between AI companies and European publishers.

What to watch next is the Court’s judgment and how national courts apply the distinction. For AI companies negotiating with publishers, the training side now looks more settled; the output side is where compliance work is likely to concentrate.