Perspective Chapter: Fit for Purpose? Creative Commons Licensing for Research Data in the Age of Artificial Intelligence

Licensing is an important component of the re-usability of research data, itself part of the FAIR principles: without clear, machine-readable licensing, datasets risk becoming technically interoperable but legally unusable, undermining reproducibility and trust. In academic contexts, Creative Commons (CC) licenses provide the standardized framework that makes reuse predictable and legally secure. Yet their design predates large-scale AI training and thus raised new questions in the age of (generative) AI with its large language models. This article first examines the challenges to CC licenses for used in AI training and compares and contrasts US and EU approaches on this issue. In the United States, fair use may override CC conditions in AI training. In the European Union, structured text-and-data mining (TDM) exceptions and the AI Act impose stricter obligations: developers of general-purpose AI models must respect opt-outs and implement copyright compliance policies, even extraterritorially. Secondly, the article outlines the challenge to CC licenses for data under both U.S. and EU law, works created without sufficient human input currently fall outside copyright protection and CC licensing, effectively entering the public domain. This creates new tensions for creators, researchers, and open science communities. Finally, the article explores emerging responses: machine-readable CC preference signals, codes of practice under the EU AI Act, and proposals for compensation mechanisms or reciprocity obligations.