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

In order to facilitate compliance of deployers of generative AI systems with their obligation to disclose deep fakes and AI-generated or manipulated published text pursuant to Article 50(4) AI Act, Signatories who are providers of generative AI systems that are capable of generating or manipulating such content are encouraged to provide an optional functionality in their system's interface and to implement an integrated option that allows deployers and other users to directly apply at their own discretion – upon generation of the output – a perceptible label in consistency with the Commitments and Measures in Section 2 of the Code. Signatories are also encouraged to follow harmonised UX standards, to the extent possible, in an interoperable manner with existing standardised content management systems and workflows of online platforms and media publishers. Signatories are further encouraged to implement other supporting measures for displaying labels and metadata that enable deployers and providers of online platforms, search engines and websites to implement display practices and policies that are appropriate for their use cases. This measure is without prejudice to the responsibility of deployers who remain responsible for the disclosure of deep fakes and AI-generated or manipulated published text in a clear and distinguishable manner in accordance with Article 50(4) and (5) AI Act.