Code of Practice on Transparency of AI-Generated Content (marcatura ed etichettatura dei contenuti generati o manipolati dall'IA)

Signatories are encouraged to incorporate richer information in the metadata pursuant to Measure 1.1.1, to provide additional context and thereby contribute to increased integrity and trust in the information ecosystem. In particular, Signatories are encouraged to consider relevant standards to provide further information about the origin of AI-generated or manipulated content across workflows, where technically feasible. Signatories are encouraged to add or record relevant content provenance information within content generated or manipulated by their AI systems, in particular the name of the AI system, the company name of the AI provider, and a timestamp indicating when the content was generated or manipulated. Additionally, Signatories may also include the underlying AI model identifier and version number. For AI-manipulated content, Signatories are encouraged to record information about the type of operations performed by their AI system to modify the content (e.g., object removal). Multiple discrete operations carried out by the AI system are recommended to be encoded into a single metadata marker to limit complexity and reduce burden.