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

(71) The scope of the marking obligation is limited to implementing marks in a machine- readable format. A machine-readable format means that marks are structured in a way that allows software applications to easily identify, recognise and extract them without human intervention. Perceptible marks and labels are not excluded as a complementary measure, where appropriate, with a view to facilitating the compliance of deployers with their obligation to label deep fakes pursuant to Article 50(4) AI Act.

(72) To comply with the marking obligation, providers may rely on a single marking technique or a combination of techniques, so long as their overall technical solution is machine- readable and meets the requirements for effectiveness, interoperability, robustness and reliability to the legally required degree, as provided by the second sentence of Article 50(2) AI Act (see Section 4.2.3. below).

(73) Recital 133 AI Act provides examples of such techniques to include watermarks, metadata identifications, cryptographic methods for proving provenance and authenticity of content, logging methods, fingerprints or other techniques, and a combination of such techniques. While methods for proving provenance and authenticity are mentioned in Recital 133 AI Act, providers are not required to record or keep a full provenance chain containing information on content origin and modifications or any other relevant assertion concerning the history of the content. However, such provenance methods may also be used for compliance with Article 50(2) AI Act and be conducive in enabling natural persons to distinguish AI-generated or manipulated content from authentic content.

(74) Providers may implement the marking solution for the outputs at different stages of the value chain (e.g. post hoc after content generation at the level of the AI system, at the

(27) Recital 133 AI Act clarifies that it is appropriate to require providers of those systems to embed technical solutions that enable marking in a machine-readable format and detection that the output has been generated or manipulated by an AI system and not a human. Recital 135 AI Act refers to the obligation regarding detection.

level of the underlying AI model or integrated in the AI system’s inference process). They may rely on the marking solution implemented by an upstream model provider or a third party providing the solution (e.g. as an open standard or a specialised service) to the extent that the marking solution is compliant with Article 50(2) AI Act. Such reliance is without prejudice to the responsibility of the provider of the AI system to demonstrate compliance with Article 50(2) AI Act.