The EU’s highest court could clarify whether rights press publishers have to control the reproduction of their content online are engaged where extracts of their content are displayed to users of generative AI (gen-AI) tools operated by other businesses.
A court in Hungary has asked the Court of Justice of the EU (CJEU) to interpret aspects of EU copyright law in a case relevant to both AI developers and copyright owners.
The CJEU’s ruling on the four questions put to it could provide guidance on whether, and if so the circumstances in which, gen-AI providers might need a licence to display press publishers’ content – as well as whether publishers wishing to prevent their content from being processed by gen-AI tools under an exception provided for in EU copyright law need to take an action in this regard.
The copyright provisions in question are set out in the EU’s Digital Single Market Copyright (DSM) Directive 2019 and in another EU copyright law, the Information Society Copyright Directive, that dates back to 2001.
Together, the laws provide EU-based publishers of press publications with the exclusive right to authorise or prohibit the reproduction and online communication to the public of their copyright works for commercial purposes. Those rights are subject to certain limitations. For example, the rights do not extend to use of individual words or very short extracts and they expire two years after the content is published – with the countdown to expiration starting from the beginning of the next year after the content is published.
The DSM Directive further provides an exception to the rights of press publishers to control reproduction of their works. That exception enables others to reproduce and extract from lawfully accessible works and other subject matter for the purposes of text and data mining (TDM) – including in the commercial context. However, some limitations apply to the TDM exception.
First, such reproductions or extractions can only be retained for as long as is necessary for the purposes of TDM; and second, the exception is conditional on rights holders not expressly reserving that their works cannot be used for the purposes of TDM – those reservations must be made “in an appropriate manner, such as machine-readable means in the case of content made publicly available online”, the law provides.
In the case before the CJEU, the court has first been asked to determine whether press publishers’ rights to control reproduction of their works are engaged when “text partially identical to the content of web pages of press publishers” is displayed in the responses of AI chatbots based on large language models (LLMs). At the heart of the question is whether the display of such text constitutes a ‘communication to the public’ of the publishers’ content as a matter of law – something the CJEU has been asked to determine.
It has been further asked if where the response displayed by a chatbot is derived from merely predicting the next word on the basis of observed patterns, this is relevant to the determination of whether there has been a communication to the public or not.
The EU courts have already considered what is meant by ‘communication to the public’ in the context of copyright claims. It is established that rights holders – in an essential step to successfully asserting copyright infringement based on there being unauthorised ‘communication to the public’ of their works by others – need to demonstrate that their works have been accessed by a ‘new public’ to the one they reached through original publication.
Another question that the CJEU has been asked is whether the process of training an LLM-based chatbot constitutes an instance of reproduction if that LLM is built on the basis of the observation and matching of patterns and this makes it possible for the model to learn to recognise linguistic patterns. If so, the Hungarian court has asked whether that reproduction of publishers’ content falls within the TDM exception.
Finally, the CJEU has been asked whether LLM-based chatbot providers will be said to have reproduced press publishers’ content, engaging their rights to control such reproduction, if the response it generates contains all or part of the publishers’ content but is generated on the basis of a user instruction that matches the text contained in a press publication or refers to that text.
The Hungarian court is seeking answers to its questions to help it resolve a dispute that has arisen between Like Company, a news publisher in Hungary, and Google over content displayed by Google’s Gemini chatbot. It is the first time that questions pertaining to AI and copyright have been put to the CJEU for its interpretation.
The referral comes ahead of hearings scheduled this month before the High Court in London in the AI copyright dispute between Getty Images and Stability AI.
Contact an adviser
