Google Gemini Copyright Case: EU Adviser Backs Publishers

In a closely watched Google Gemini copyright case, an adviser to the European Union's top court said on Tuesday that the chatbot's use of press articles may, in some circumstances, infringe publishers' rights, Reuters reported. He added that using copyrighted material to train AI models is allowed under EU rules.

Advocate General Maciej Szpunar gave his opinion in a case brought by Hungarian website operator Like Company against Google Ireland at the Court of Justice of the European Union (CJEU) in Luxembourg.

What the case is about

According to the court's case summary, Like Company published an article about a well-known Hungarian singer. When Gemini, previously called Bard, was asked to summarise it, the chatbot gave a detailed answer that included a summary of information from Like Company's articles.

Like Company argues this amounted to copying and making its protected content available to the public without consent or payment. Google denies infringement. It argues there was no reproduction or communication to the public, that exceptions for temporary copies and text and data mining apply, and that much of the answer was based on the chatbot's "hallucinations". A court in Budapest referred several questions to the CJEU.

Key points of the opinion

• Answers can infringe: when a chatbot fetches an article from a publisher's website, reuses it to generate an answer and gives that answer to a user, the publisher's right may be infringed in certain circumstances.
• Facts are free: copyright does not protect the bare facts reported in news, only the specific way they are expressed.
• Training is allowed: using protected material to train large language models falls under the EU's text and data mining exception.

The opinion is not binding. The CJEU judges will issue their own ruling later, although Reuters noted that judges usually follow the advocate general's view.

Why it matters for AI and news

The opinion separates two stages of how chatbots work: learning from data, and answering users in real time by pulling in fresh web content. Gemini, like other assistants, can search the web to give up-to-date answers, which is especially useful for news summaries. If the court agrees with Szpunar, AI companies may face more pressure to license news content for live answers in Europe, even if training stays protected.

The case adds to a wave of disputes between publishers and tech firms. In the United States, The New York Times this week sued the Brave browser over alleged scraping for AI companies. For users in Thailand and elsewhere, the outcome could affect how freely chatbots quote and summarise news.

Read more about Google's assistant in our report on the Apple and Google Gemini AI deal and our Gemini app integration guide.

Photo: Court of Justice of the European Union, Luxembourg, by FrDr / Wikimedia Commons, CC BY-SA 4.0.