Commission publishes second draft code of practice for AI content marking under EU AI Act
The European Commission has published the second draft of the Code of Practice on Transparency of AI-Generated Content to support compliance with the transparency obligations in Article 50(2), (4) and (5) of the EU AI Act. The draft sets out voluntary commitments for providers of generative AI systems and deployers of AI systems to ensure that AI-generated or manipulated content is appropriately marked, detectable and clearly disclosed to natural persons, with the aim of promoting trustworthy and human-centric AI while safeguarding fundamental rights and supporting innovation. For providers, it establishes a multi-layered marking approach requiring at least two machine-readable techniques, notably digitally signed metadata and imperceptible watermarking, supported by detection mechanisms and requirements of effectiveness, reliability, robustness and interoperability. For deployers, it requires clear and distinguishable disclosure of deepfakes and certain AI-generated or manipulated text published on matters of public interest, including through an ‘AI’ icon or equivalent label meeting defined design and placement standards, and sets out proportionate rules for artistic and similar works, internal compliance measures, and conditions for reliance on the exemption where human review and editorial responsibility are exercised.