{"slug":"eu-ai-act-art-50-2-synthetic-content-marking","id":"AIPT-OBL-0039","permanent_url":"https://aipolicytracker.org/id/AIPT-OBL-0039","title":"Providers of generative AI must mark synthetic output as artificially generated in a machine-readable way","short_title":"Marking AI-generated content","category":"transparency","summary":"Providers of AI systems, including general-purpose AI systems, that generate synthetic audio, image, video or text must ensure the output is marked in a machine-readable format and detectable as artificially generated or manipulated. The technical solution must be effective, interoperable, robust and reliable as far as the state of the art allows, taking account of content type and cost. Systems that only perform assistive editing or do not substantially alter the input, and law-authorised criminal-detection uses, are outside the duty.\n","practical_action":"Implement watermarking or content credentials on generated output and keep test evidence that the mark survives common transformations.\n","is_binding":true,"applies_from":"2026-08-02","section":"Article 50","source_reference":"Article 50(2)","official_source_url":"https://eur-lex.europa.eu/eli/reg/2024/1689/oj","review_status":"verified","confidence_level":"high","last_verified_at":"2026-09-26","reviewed_by":null,"actors":["provider","gpai_provider"],"sectors":["cross_sector","media_information","consumer_retail"],"use_cases":["generative_ai","content_moderation"],"applicability":[{"description":"Providers of systems that generate synthetic media or text, whatever the risk tier; codes of practice under Article 50(7) will detail acceptable techniques.","actors":["provider","gpai_provider"],"conditions":null}],"evidence_examples":[{"title":"Provenance marking design and robustness test report","description":null,"artifact_type":"report"},{"title":"Output sample with embedded machine-readable mark","description":null,"artifact_type":"record"}],"framework_mappings":[{"framework":"iso_42001","framework_name":"ISO/IEC 42001:2023","reference":"Annex A.8.2, A.6.2.4","note":"System documentation; verification and validation of the marking.","confidence_level":"low","is_original":true},{"framework":"nist_ai_rmf","framework_name":"NIST AI RMF 1.0","reference":"MEASURE 2.7, MANAGE 4.1","note":"Provenance mechanisms as recommended in NIST AI 600-1.","confidence_level":"medium","is_original":true}],"controls":[{"slug":"synthetic-content-labelling-and-provenance","title":"Synthetic content labelling and provenance marking","kind":"technical","relationship":"satisfies","note":"Machine-readable marking with robustness evidence.","confidence_level":"high","url":"https://aipolicytracker.org/controls/synthetic-content-labelling-and-provenance"},{"slug":"accuracy-robustness-security-testing","title":"Accuracy, robustness, fairness and security testing","kind":"technical","relationship":"supports","note":"Tests that the mark is robust and reliable.","confidence_level":"medium","url":"https://aipolicytracker.org/controls/accuracy-robustness-security-testing"}],"url":"https://aipolicytracker.org/obligations/eu-ai-act-art-50-2-synthetic-content-marking","policy":{"slug":"eu-ai-act","title":"Regulation (EU) 2024/1689 — Artificial Intelligence Act","status":"partially_applicable","jurisdiction":"eu","url":"https://aipolicytracker.org/policies/eu-ai-act"}}