EUIPO Releases Comprehensive Study on Generative AI and Copyright Implications

Generative Artificial Intelligence (GenAI) operates by leveraging existing content to produce new outputs, marking a shift in the traditional creation paradigm, as not all content is now human-made. This fundamental characteristic of GenAI brings pressing concerns regarding the legal use of copyrighted material for training these systems, and highlights the need for clear methods to differentiate between copyrighted and non-copyrighted content.

To address these intricate challenges, the newly published study by the European Union Intellectual Property Office (EUIPO), aims to enhance overall understanding of how Generative AI operates, while exploring both current and emerging solutions for applying EU copyright and artificial intelligence (AI)regulations. The study provides a comprehensive examination of GenAI advancements through the lens of EU copyright law, encompassing technical, legal, and economic dimensions.

The study is grounded in a robust research methodology that combines desk research, expert interviews, and in-depth analysis of technical practices and solutions, offering a well-rounded basis for informed policy and legal considerations. Developed in close collaboration with relevant services of the European Commission, the EUIPO study focuses on three interrelated areas: 

1. The use of copyright-protected works as training data for Generative AI models;
2. The creation of new content by these systems and the legal challenges this entails; and
3. The broader implications for creators, AI developers, and the overall copyright ecosystem.