A second meeting of the Signatory Taskforce under the General-Purpose AI Code of Practice was hel d on 13 March 2026, bringing together nearly all participating signatories to discuss the implementation of copyright-related obligations for AI models.

Moderated by the European AI Office, the discussions focused on measures aimed at mitigating copyright risks in AI systems, including reducing the likelihood of copyright-infringing outputs and establishing effective complaint mechanisms for rights holders.

Participants shared early experiences and best practices, highlighting that copyright risk mitigation requires interventions across the entire AI lifecycle, from data curation and training to pre-deployment testing and output monitoring. Some approaches discussed included the use of licensed training data and similarity-based attribution mechanisms to improve transparency and support fair remuneration for content creators.

The meeting also addressed the establishment of contact points for handling complaints, with signatories exchanging views on how to ensure accessible and timely responses. The AI Office emphasized continued cooperation and invited further input ahead of the next taskforce meeting.