EU Adviser Warns Over AI Summaries

An adviser to the European Union's highest court has warned that artificial intelligence chatbots could infringe publishers' copyright protections when generating answers from news articles retrieved online. The opinion, delivered on 3 November by Advocate General Maciej Szpunar, concerns Google's Gemini chatbot and could influence how AI companies use journalism in responses to users.

The case was brought by Hungarian website operator Like Company, which argued that Gemini used material from its publications without authorisation. The dispute concerns whether retrieving and reusing protected news content constitutes reproduction or communication to the public under EU copyright law. Szpunar concluded that infringement may arise when an AI system retrieves an article, processes its contents and generates a response reproducing protected expression.

However, the opinion draws an important distinction between copyrighted expression and the underlying facts contained in journalism. Facts themselves do not receive copyright protection, meaning that summarising information does not automatically constitute infringement. The adviser also distinguished the use of material to train language models from its subsequent retrieval and reproduction in generated answers, identifying different legal considerations for those activities.

The opinion is not binding, and the Court of Justice must still deliver its judgment. Nevertheless, it addresses a significant question in the relationship between publishers and AI developers: when does providing an answer based on journalism become an unauthorised use of protected content? The eventual ruling could influence licensing negotiations, AI search services and the commercial rights of European news organisations.