Written in the stars? Gemini and the future of EU Copyright

The Court of Justice of the European Union (CJEU) is set to consider a reference concerning Google’s generative AI chatbot, Gemini, in what could become a significant decision for the future application of EU copyright law to artificial intelligence systems. The referral stems from proceedings brought by Manchester United Fanatics Klub Magyarorszáq.

The referring court has asked the CJEU to clarify several important issues, including whether Gemini’s generation of summaries or rewritten versions of news articles amounts to a communication of the original copyright-protected works to the public, notwithstanding that the output itself is newly generated.

The court has also sought guidance on whether training a large language model (LLM) on copyright-protected content constitutes an act of reproduction under EU copyright law and, if so, whether such training may benefit from the text and data mining (TDM) exception. The TDM exception permits the automated analysis of digital content to identify and extract information such as patterns, trends and correlations.

In addition, the referral addresses whether copyright-relevant reproductions occur when users provide protected press content in prompts and Gemini processes that material to generate responses. The CJEU has also been asked to examine Gemini’s “grounding” functionality, under which the system accesses Google Search and temporarily uses extracts from indexed webpages to obtain current factual information before those extracts are removed from its intermediate memory.

The case further raises questions regarding rights reservation and consent. In particular, the court will consider whether publishers that have not expressly opted out of AI training or grounding activities should be treated as having consented to such uses and whether a failure to reserve rights under the TDM regime effectively permits those activities.

The referral follows the earlier Like Company reference concerning Gemini but goes further by placing a broader range of AI processes before the CJEU, including grounding and AI-generated summarisation. The decision is therefore expected to provide important guidance on how EU copyright law applies to generative AI systems and the extent to which publishers can control the use of their content in the development and operation of those systems.

If your business creates, publishes or relies on valuable content, the outcome of this case could have significant implications for how that content may be used in the training and operation of AI systems. Please reach out to our Intellectual Property and Technology & Digital teams if you have any queries about protecting your rights, managing AI-related risks or preparing for future developments in this rapidly evolving area.