Guidelines on the implementation of the transparency obligations for certain AI systems under Article 50 of the AI Act — C(2026) 5054 final, Annex

(117) Article 50(4), first subparagraph, AI Act obliges deployers of AI systems that generate or manipulate image, audio or video content constituting a deep fake to ‘disclose that the deep fake content has been artificially generated or manipulated’. Recital 134 AI Act clarifies that deployers using AI systems to generate or manipulate deep fake content should clearly and distinguishably disclose that such content has been artificially created or manipulated by labelling the AI output accordingly and disclosing its artificial origin. Labelling or disclosure methods applied in accordance with Article 50(4), first subparagraph, AI Act should be understandable and perceivable by natural persons (e.g. with visible or audible labels), without them needing to rely on any specific technical tools or performing dedicated actions. Therefore, deployers cannot rely on the machine- readable marking embedded in the content by the provider under Article 50(2) AI Act, since those markings are not immediately clear and distinguishable for the natural persons exposed to the deep fake content.

(118) In light of Article 50(7) AI Act, deployers may adhere to a code of practice, assessed as adequate pursuant to Article 50(7) AI Act to demonstrate compliance with this disclosure obligation. For further information regarding the effects of adherence to such a code and the supervision of other deployers not adhering to a code, see Section 8.1 below.