EU adviser says Google’s Gemini may infringe press rights

Google's Alphabet Inc. ↗ 0.08% Gemini chatbot may infringe press publishers' rights when it uses news articles to answer user queries, an adviser to the European Union's top court said on Tuesday, Nov. 3. He also said that training AI models on copyrighted material is generally allowed, Reuters reported.reuters.com

The court says this is the first generative AI case to reach the Court of Justice of the European Union, according to its press release. The opinion from Advocate General Maciej Szpunar sets out a possible test for when chatbot answers cross a legal line. It comes as online media outlets lose traffic to AI-written news summaries that are replacing traditional search results with links, Reuters noted.curia.europa.eu+1

A two-part test

"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 the court's statement. He said the publishers' right protects only an article's "specific form of expression," not the facts it reports. Its purpose is to protect publishers' investment, not to give them "a monopoly over the news itself."curia.europa.eu

He said whether infringement occurs depends on two factors. The first is whether the chatbot got the article unlawfully, for example by getting around a paywall or anti-scraping protections. The second is whether the answer reuses the article's wording so recognisably that the user no longer needs to read the original. Systematic reuse of that kind may also breach the right of making content available to the public. The chatbot provider would be liable if such reuse is built into its business model, or if it has not taken "credible and effective" steps to prevent it, according to Paperjam.en.paperjam.lu

Training questions set aside

Szpunar advised the court to dismiss the two questions about model training as inadmissible. He said there is no evidence that Like Company's articles were used to train Gemini, and that training is not an "online use" of press publications. For completeness, he added that training a model on protected works is, as a rule, "freely permitted under the text and data mining exception," the court said.curia.europa.eu

The Hungarian dispute

Like Company, a Hungarian publisher that runs several news portals, sued Google Ireland at the Budapest District Court. It argues that Gemini reproduced its articles in responses to users between June 2023 and May 2024, Il Sole 24 Ore reported. Google denied that Gemini's answers reach a "new public" and relied on copyright exceptions, including the one for text and data mining, according to Geradin Partners. The case turns on Article 15 of the EU's 2019 copyright directive, EU Law Live noted.en.ilsole24ore.com+2

Advocate general opinions are not binding, but the court's judges usually follow them, Reuters said. The judges have now started deliberating and will rule at a later date. After that, the Hungarian court will decide the underlying case.reuters.com