CJEU to rule on landmark AI and copyright case from Hungary
On 3 April 2025, the Court of Justice of the EU (CJEU) received its first case addressing the interplay of artificial intelligence (AI) and copyright law. The preliminary ruling request was submitted by the Budapest District Court of Hungary. The case is currently pending, but observers are closely watching developments since the CJEU’s decision could set a precedent for how AI-generated content is treated under European copyright law.
The dispute: Like Company vs Google Ireland
The case concerns a dispute between the Claimant, the Hungarian press publisher and operator of various news portals Like Company and the Defendant, Google Ireland Limited, the developer and operator of the generative AI chatbot Gemini. The matter centres on the Claimant’s assertion that Google’s AI system infringed on its related rights as a publisher of press publications after Gemini allegedly reproduced and made its content publicly available without authorisation.
Background and functionality of Gemini
The Claimant Like Company operates several news portals and earns advertising revenue based on user traffic. It alleges that Gemini, in response to user queries, generated summaries of its articles, which is content protected under copyright law.
Gemini is a large language model (LLM) designed to generate content such as text, images, music, and code. It does not store data in a traditional database but instead processes information using tokenisation and pattern recognition (i.e. string searching). While Gemini can quote and summarise content from web pages, it typically directs users to the original sources via hyperlinks. Importantly, Gemini relies on Google Search to gather data and does not maintain a fixed repository of information. When it is asked a specific question directly, the chatbot provides a response that displays the content of a protected press publication.
Claimant’s position
The Claimant argues that Gemini’s use of its content goes beyond the permissible scope of “individual words or very short extracts of a press publication” and constitutes unauthorised reproduction and communication to the public. Like Company emphasises that consenting to search-engine indexing does not equate to a broader dissemination of its articles via AI-generated responses. The Defendant did not pay compensation and, therefore, no basis for consent exists. Furthermore, the Claimant contends that the chatbot’s interface is distinct from traditional search engine result pages and qualifies as public communication of protected content.
Defendant’s defence
The Defendant, Google Ireland, counters that Hungarian law does not apply since the chatbot was not trained in or by infrastructure located in Hungary. It also denies that Gemini’s responses amount to reproduction or making material available to the public. It asserts that the content does not reach a “new public”, but can be accessed by the same public that can view the original articles online. Moreover, the responses displayed do not reach the threshold for the application of “individual words or very short extracts”.
Google further argues that even if the responses are deemed infringing, they fall under exceptions for temporary reproduction and text and data mining. Additionally, it claims that the chatbot’s outputs are not identical to the Claimant’s articles and often include “hallucinated” content: information generated by the AI that does not originate from the protected publications.
What’s next
The case is currently pending before the CJEU, and a final judgement may take years. Legal observers are closely watching this development since it could set a precedent for how AI-generated content is treated under European copyright law. The procedure before the CJEU can be followed here. Law-Now will publish an update when the final judgement is issued.
For more information and guidance on AI and copyright compliance, contact your CMS client partner or these CMS experts.
The article was co-authored by János Bálint.
