Generative AI is transforming the way knowledge is created, accessed and disseminated while simultaneously challenging fundamental principles of copyright law. This workshop provides a structured overview of the legal issues surrounding generative AI, following the lifecycle of AI systems—from the use of protected works as training data to the legal implications of AI-generated outputs.

The first part focuses on AI inputs, examining the copyright framework governing the training of generative AI models. It presents a comparative legal analysis of the approaches adopted in the European Union and other leading jurisdictions, reviews the most significant court cases and ongoing litigation, and discusses the practical implications of the EU Text and Data Mining (TDM) exceptions, with particular emphasis on Articles 3 and 4 of the DSM Directive. A Slovenian case study illustrates how these legal issues are emerging in practice and what they mean for research organisations and libraries.

The second part addresses AI outputs, exploring questions of authorship, the copyright protection of AI-generated content, and the broader legal challenges arising from the use of generative AI.

The workshop concludes with an overview of the forthcoming European debate on copyright and generative AI. It examines why legislative reform is increasingly considered necessary, what evidence the European Commission is currently gathering, the expectations of different stakeholders, and the possible directions of future policy.

The final discussion highlights the strategic importance of developing sovereign European generative AI and the central role of the mandatory research Text and Data Mining exception under Article 3 of the DSM Directive in enabling innovation, supporting research infrastructures, and strengthening Europe’s digital and technological sovereignty.

Among the countries to be analyzed, Slovenia has already adopted some of the most advanced open science strategies in the EU, supported by legislative measures such as rights retention obligations and the secondary right to publication included in the Scientific Research and Innovation Act (ZZrID). These provisions strengthen the ability of researchers to retain key substantive copyrights in negotiations with publishers, allowing them to publish their research results in open access, thereby accelerating the practical implementation of open science principles. The Slovenian approach can serve as a valuable model for other countries in the region.

The project Mapping of Open Science is supported by the Knowledgerights 21 (KR21) and conducted by Open Data and Intellectual Property Institute with the cooperation of members of the KR21 Regional Alliance.