Bottom line: Chief Digital Officers must establish processes by August 2026 to transparently label AI-generated media (images, video, audio), texts and chatbot interactions to minimize deception risks.
From 2 August 2026, marketers must label AI-generated or manipulated content as such – as prescribed by Article 50 of the EU AI Act. This applies to deepfakes, synthetic texts without editorial control and chatbots on corporate websites.
The EU AI Act differentiates between risk categories and actor roles. For marketing departments, the “limited risk” category and the role of the deployer are particularly relevant. The focus here is on transparency obligations – especially for synthetic or manipulated media, that is, content created or altered using AI. For AI systems designed for interaction (such as chatbots on websites), the role of the provider may also be relevant for marketers.
Three marketing scenarios are specifically affected: (1) Publication of synthetic or manipulated images, audio or video – particularly photorealistic persons or authentic-sounding voiceovers (deepfakes); (2) AI-generated or manipulated texts without editorial control (LinkedIn posts, blog articles, web content); (3) human-machine interactions such as chatbots or voice assistants on corporate websites. The labeling must occur at the moment of first perception or interaction.
Article 50(4) of the AI Act defines the requirements for deepfakes precisely: Users of AI systems that generate or manipulate image, audio or video content must disclose that the content has been artificially generated or manipulated. An exception applies to artistic, creative, satirical or fictional works – there an appropriate label is sufficient that does not impair the presentation or enjoyment of the work.
Article 3(60) of the AI Act defines deepfakes as AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places or events and would appear authentic or truthful to a person. The standard is whether the content appears realistic and has potential for deception – not technical perfection. The resemblance should be objectively recognizable, and it is sufficient if simulated persons or objects appear realistic and depict something that could exist or could have existed in reality.
Source: www.activemind.legal · Published 6 July 2026
Lumi AI News — AI-assisted curation in accordance with Article 50 EU AI Act. Paraphrase and classification by Lumi News Pipeline v1.7.3.