On 10 June 2026, the European AI Office published the final Code of Practice on Transparency of AI-Generated Content — the practical framework that providers and deployers of generative AI systems can use to demonstrate compliance with the transparency obligations under Article 50 of the EU AI Act. The Code arrives less than two months before those obligations become applicable on 2 August 2026.
Article 50 imposes two distinct sets of duties. Providers of AI systems that generate synthetic audio, image, video, or text must ensure their outputs are marked in a machine-readable format that allows downstream detection that the content was artificially generated or manipulated. Deployers — organisations using these systems professionally — must disclose when content constitutes a deepfake and must clearly label AI-generated text published to inform the public on matters of public interest. The Code translates these requirements into concrete technical guidance, recommending that providers use digitally signed metadata and imperceptible watermarking to embed provenance information, and proposing a standardised EU label for AI-generated content alongside technical standards for watermarking, metadata, and provenance tooling.
Adherence to the Code is voluntary, but it functions in the same way as the AI Act's other codes of practice: signing up provides a presumption of compliance and a defensible basis for demonstrating that an organisation has met its obligations. Companies wishing to be listed among the initial signatories must submit their completed signature forms by 22 July 2026. The Code was published alongside draft Guidelines on Article 50, which the Commission has opened for consultation and which provide interpretive detail on scope, definitions, and exceptions — for example, where AI-assisted editing does not substantially alter the underlying content.
The timeline retains a transitional accommodation. AI systems placed on the market before 2 August 2026 benefit from a transitional period running to 2 December 2026 before full compliance is expected, giving operators of existing systems a short runway to implement marking and labelling mechanisms. New systems placed on the market on or after 2 August must comply from that date.
Acompli perspective: The transparency obligations apply to any organisation that deploys generative AI professionally — not only the providers that build the underlying models. If your organisation uses AI to produce content that could be published, shared externally, or mistaken for human-created material, the deployer obligations under Article 50 are now squarely within scope. The Code gives a concrete compliance pathway, but it does not remove the need to understand where generative AI sits in your operations. Reviewing your risk assessments to cover AI-generated outputs, and ensuring your records of processing reflect where generative AI systems are used and what they produce, are the practical first steps before the August deadline.