Article 50 of Regulation (EU) 2024/1689 (the AI Act) entered application on 2 August 2026. The obligations apply to four categories of deployment: providers of interactive AI systems must notify users they are communicating with an AI; providers of AI systems generating synthetic audio, image, video, or text must ensure outputs are marked in machine-readable format and detectable as artificially generated; deployers of emotion-recognition or biometric-categorisation systems must inform persons of the operation; and deployers of AI systems producing deep fakes for content of public interest must disclose the artificial origin.
The controlling text is Article 50(1) through (4) of Regulation (EU) 2024/1689, read with Recital 134. Article 50(1) requires providers to ensure that AI systems designed to interact with natural persons are identified as such to users in clear and timely terms, except where the AI nature is obvious from context. Article 50(2) requires providers of generative AI systems to ensure outputs are marked in machine-readable format and detectable as artificially generated or manipulated; content watermarking and provenance metadata are the primary technical means envisaged. Article 50(4) requires deployers of AI systems generating deep fakes to disclose the artificial origin of that content, with limited exceptions for artistic, satirical, or fictional works.
Every business deploying generative AI to produce public-facing content faces disclosure and marking obligations from 2 August 2026. Practical obligations fall on technology companies, media publishers, financial institutions using AI in customer communications, and AI deployers in advertising and marketing. Providers of general-purpose AI systems bear the primary obligation to ensure technical marking capability is built into their systems; downstream deployers bear obligations where they control final output generation or dissemination to the public.
The AI Omnibus provisional agreement adopted in May 2026 grants providers of generative AI systems already on the market before 2 August 2026 until 2 December 2026 to meet the machine-readable marking requirement under Article 50(2), as a transitional measure for legacy systems. The European Commission's Code of Practice on Transparency of AI-generated Content, signed by approximately 190 organisations by late July 2026, provides a compliance pathway. Providers and deployers not adhering to the Code must demonstrate compliance by alternative means of equivalent adequacy, as the AI Office will oversee enforcement.
Licentium advises clients on AI regulatory compliance across EU and national implementing measures, including Article 50 readiness assessments. Contact us to discuss your obligations. Work we undertake includes transparency obligation gap analysis, AI content marking and watermarking policy design, AI governance programme implementation, generative AI product pre-launch review, and Code of Practice alignment mapping.