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

Signatories are encouraged to support the work and activities of a dedicated task force under the Code aimed at advancing the further development and usability of the EU icon as a minimum state-of-the-art implementation. The task force will, inter alia, aim to:

a) further promote the uptake and development of the EU icon (incl. non-visual/audio-only versions of the icon);

b) update the EU icon to ensure state-of-art disclosure at all times including in line with user experience (UX) standards;

c) assess feasibility and work towards advancing technical solutions, including the development of an interactive second layer linked to the EU icon or other equivalent labels, which provide opportunities for the integration of the EU icon or equivalent label into online content dissemination services to provide end-users with additional information (e.g., type of AI-involvement in a deep fake or published text). This interactive second layer may be implemented to the extent feasible and practical, and in alignment with the technical marking solutions described in Section 1 of the Code, to ensure robustness of the information provided in the icon and the accompanying information;

d) work towards identifying common methods for disclosing the type of AI involvement in the case of AI manipulated deep fakes or published text, with a view to advancing consistency in the implementation among Signatories;

e) provide a forum for Signatories to exchange good practices across sectors and possibly to develop sectoral good practices regarding the disclosure of deep fakes and published text in line with this Section of the Code. Signatories and relevant stakeholders from various sectors, including but not limited to independent researchers, civil society organisations and national competent authorities, are encouraged to participate in the task force.