This study, commissioned by COMMUNIA to Centrum Cyfrowe and Open Future and authored by Maria Draczyk and Alek Tarkowski, identifies the misalignment between the EU’s pursuit of digital sovereignty based on open source AI development, and its existing legal frameworks.
Based on several in-depth interviews, the study finds that legal uncertainty around AI training under the TDM exceptions and sharing of datasets are the biggest legal challenge for the development of open-source LLMs. Compliance with opt-out requirements is only made more complex by a lack of standard, machine-readable information on rights reservations.
