Read, Rethink, Reform: Why a Strong TDM Exception for Research Matters Now More Than Ever
What’s it About
A well-functioning text and data mining (TDM) exception is essential to AI-driven scientific research and education in Europe. However, as Dr. João Pedro Quintais 🟥 (Institute for Information Law (IViR), University of Amsterdam) highlights in his latest study, legal uncertainty around the definition of a “research organisation” in EU law creates fragmentation across Member States. This inconsistency affects (cross-border) AI research, public-private partnerships, and the ability of institutions to leverage TDM without facing legal risks.
Why It Matters
Article 3 of the EU Copyright Directive (CDSMD) allows research organisations to conduct TDM without needing rightsholder consent. However, because each country interprets “research organisation” differently, legal uncertainty hampers AI development. Meanwhile, the EU AI Act imposes additional compliance obligations, but research organisations could benefit from reduced copyright constraints—if they qualify under the current patchwork of definitions.
A case in Germany (Kneschke v. LAION) confirmed that non-commercial AI research is protected under the TDM exception. Yet, public-private partnerships remain in a grey zone, with no clear legal path to industry-academia cooperation. This uncertainty slows down innovation and puts Europe at a disadvantage in the global AI race.
Rethink – The Bigger Picture
The UK Consultation: A Chance to Lead on AI & Copyright
The UK government has just closed its consultation on AI and copyright policy. While the emphasis seems mostly on TDM outside of research, any rethinking of TDM presents an opportunity to establish a forward-thinking legal framework that supports all AI-driven research and education.
Switzerland’s Approach: A Model for the Europe?
Switzerland offers an alternative approach: rather than focusing on who conducts the research, it considers why the research is being conducted. This purpose-driven model ensures that both commercial and non-commercial research can thrive without unnecessary legal roadblocks—a key lesson for EU and UK policymakers.
Reform – The InfoLabs Take: Possible Path Forward
To keep Europe competitive and place it at the forefront of responsible AI innovation, policymakers should:
- Clarify the definition of “research organisation” at the EU level to prevent national fragmentation.
- Ensure legal certainty for public-private partnerships in AI research.
- Adopt whenever possible a purpose-driven approach to TDM, similar to Switzerland’s model, to foster innovation while upholding copyright.
Written by Caroline De Cock, LL.M. , Head of Research.
February 26, 2025
