European Copyright Opinion Challenges Chatbot Practices

A legal opinion delivered to the European Union's highest court has raised the possibility that artificial intelligence chatbots infringe publishers' rights when they reproduce protected news material in answers to users.

Advocate General Maciej Szpunar issued the opinion on Tuesday in proceedings involving Google's Gemini and Hungarian website operator Like Company. The dispute concerns whether retrieving published articles and using them to produce AI-generated responses constitutes unauthorised reproduction or communication of protected material under European copyright rules.

Szpunar distinguished between the facts contained in a news report and the particular wording through which those facts are expressed. He argued that copyright protection generally applies to the latter, meaning a chatbot's response could infringe publishers' rights if it reproduces protected expression. The adviser recommended that the court decline to rule on questions concerning model training because the proceedings did not establish that the relevant publisher's articles were used for that purpose. His analysis nevertheless suggested that certain training activities may fall within European text-and-data-mining exceptions.

The opinion is not binding, and the Court of Justice may reach a different conclusion. Its significance lies in the distinction it draws between gathering information and reproducing the work of publishers. News organisations have argued that AI-generated summaries can reduce traffic to their websites while relying on material they finance. A final ruling could influence how technology companies obtain, summarise and attribute published content, while clarifying the commercial protections available to European publishers.