EUIPO Will Launch ‘Copyright Knowledge Centre’ to Help Address Emerging AI Issues

Photo credit: Laurie DIEFFEMBACQ

On the same day that the U.S. Copyright Office (USCO) officially released part three of its study into the effects of artificial intelligence (AI) on copyright law, the European Union Intellectual Property Office (EUIPO) announced its own 436-page report on AI and copyright.

The EUIPO report was compiled by a research team of the University of Turin Law School and the Nexa Center for Internet & Society from the Polytechnic of Turin. It was based on “desk research as well as interviews from key stakeholder groups including, copyright holders, AI companies, technology solution providers as well as public organisations,” according to the EUIPO.

The USCO released a pre-publication copy of its latest report, which focused more narrowly on analyzing fair use in the context of AI, on Friday, in between the controversial firings of both the Librarian of Congress and Register of Copyrights. On MondayParts one and two of the reports were published last year and focused on digital replicas and copyrightability, respectively.

The EUIPO’s report is broader and geared to copyright experts, and also organized in three parts:

“1) a technical, legal and economic analysis to further understand the functionality of GenAI and the implications of its development, as well as a detailed examination of copyright-related issues regarding the

(2) use of content in GenAI services development and the

(3) generation of content.”

One of the results of the report is the EUIPO’s announcement that it will create a “Copyright Knowledge Centre” by the end of 2025. The Centre “will equip copyright holders with clear, practical information on how their works may be used in the development of GenAI – and how they can effectively manage and protect their intellectual assets,” said the Executive Brief released with the report.

The new Copyright Knowledge Centre will “help to raise awareness on measures and good practices to mitigate the risks of GenAI output infringing copyright” as well as “support discussions on licensing and mediation mechanisms to facilitate the conclusion of license agreements with AI developers.”

The report also identified two key areas in need of solutions: 1) copyright holders must have the option to opt out of any licensing schemes that emerge to address the problem of generative AI models’ use of copyrighted content and 2) AI developers need to ensure that the content generated through their services is detectable in a machine-readable format.

These solutions are based on existing EU law relevant to copyright, chiefly, the Copyright in the Single Market Directive (CDSM) and the EU Artificial Intelligence Act (AI Act).

The CDSM provides for Text and Data Mining (TDM) exceptions to copyright for scientific research and other purposes, but “allows copyright holders to reserve their exclusive reproduction rights, which is commonly referred to as ‘opting-out’ of the TDM exception,” explained the report.

For publicly available information such as online content, this means the opt-out must be made by the rights holder “in an appropriate manner, including by ‘machine-readable means’. When an opt-out reservation has been expressed, AI developers need an authorisation by the right holder to use their content, for example through licensing agreement. This means that effective solutions for TDM opt-out are needed for a market for content licensing to develop.”

The report also found that AI models need high-quality content to function and that there is no one-size-fits-all solution for protecting rights. “Instead, different approaches and solutions are developing for copyright holders to protect their rights, and for AI developers to respect their regulatory obligations,” said the report. On the input side this means allowing opt-out measures while on the output side transparency is key.

The report was presented to the European Parliament by EUIPO Executive Director, João Negrão, on Monday, May 12.