From the journal

EU AI Act Article 50 Transparency Obligations Apply from 2 August 2026

Article 50 of Regulation (EU) 2024/1689 (the EU AI Act) takes general application on 2 August 2026. Providers of interactive AI systems must notify users they are communicating with an AI and embed machine-readable marks in AI-generated content. Deployers must inform individuals exposed to deep fakes and AI-generated public-interest content published without human review. The European Commission published final implementation guidelines on 20 July 2026.

3 min read

Article 50 of Regulation (EU) 2024/1689 (the EU AI Act) takes general application on 2 August 2026. Competent authorities across all EU Member States may begin enforcement proceedings for non-compliance from that date. The regulation entered into force on 1 August 2024; Article 50's specific application date is 2 August 2026. The article covers providers of interactive AI systems and providers of systems generating synthetic audio, image, video, or text. It also covers deployers of emotion recognition systems, biometric categorisation systems, and AI-generated public-interest content systems.

Article 50(1) requires providers to design interactive AI systems to notify users that they are communicating with an AI, unless context makes this obvious. Article 50(2) requires providers of synthetic content systems to embed machine-readable marks enabling automated detection of AI-generated or manipulated content. Article 50(4) requires deployers of deep-fake systems to disclose that content is artificially generated or manipulated. Article 50(5) requires deployers operating AI-generated text on public-interest matters, published without human review or editorial control, to disclose the AI origin of that content. The European Commission adopted final non-binding guidelines on 20 July 2026 under Article 50 of the AI Act. The guidelines assist competent authorities and operators in applying the obligations uniformly across Member States.

Providers of large language models, multimodal systems, and chatbots must embed machine-readable provenance marks in AI-generated content before placing those systems on the EU market from 2 August 2026. Deployers running customer-facing chatbots must implement user-facing AI interaction disclosures by the same date. Deployers operating synthetic media tools and automated news publishing systems must disclose the AI origin of content. A limited grace period applies only to the machine-readable marking obligation under Article 50(2) for AI systems already on the market before 2 August 2026. Those operators have until 2 December 2026 to embed the required marks. Competent authorities designated under Article 70 may impose administrative fines under Article 99 of the AI Act for violations from 2 August 2026.

AI systems whose AI nature is obvious to a reasonably well-informed person fall outside the disclosure requirement in Article 50(1). Satire and artistic expression are excluded from the synthetic content disclosure requirement under Article 50(4), provided appropriate labelling is in place. The Commission's July 2026 guidelines state that adherence to the Code of Practice on Transparency of AI-generated Content may serve as evidence of conformity with Article 50 obligations. The Code of Practice is voluntary and does not replace the mandatory Article 50 requirements.

Licentium advises providers and deployers on Article 50 compliance, including disclosure design, machine-readable content marking, and exception analysis under the EU AI Act. Contact us to discuss your obligations under the 2 August 2026 application date. Work we undertake includes EU AI Act compliance, AI transparency obligation implementation, AI governance structuring, and technology regulation advisory.

Source: European Commission, IP/26/1653: Guidelines on Transparency Obligations under Article 50 of the EU AI Act, 20 July 2026

AI Regulatory

More from the journal

See all

Hong Kong SFC and FSTB Conclude Consultation on Virtual Asset Advisory and Management Regimes, 26 May 2026

On 26 May 2026, Hong Kong's Securities and Futures Commission and Financial Services and the Treasury Bureau published consultation conclusions on proposed licensing regimes for virtual asset advisory and management service providers. The regimes apply the same business, same risks, same rules principle and align SFC licensing requirements with those for securities advisory and management businesses. A bill implementing the regimes is planned for introduction into the Legislative Council in 2026.

OCC Grants Circle Final Charter for First National Digital Currency Bank N.A., 9 July 2026

The Office of the Comptroller of the Currency granted final approval on 9 July 2026 for Circle Internet Group to establish First National Digital Currency Bank, N.A., operating as Circle National Trust. The bank opened 24 July 2026 under direct OCC oversight and will manage USDC reserves on a directed basis, act as collateral trustee for USDC holders, and provide digital asset custody services to Circle affiliates.

Manitoba Enacts Public Sector AI and Cybersecurity Governance Act June 2026

On 1 June 2026, Bill 51, The Public Sector Artificial Intelligence and Cybersecurity Governance Act (S.M. 2026, c. 43), received Royal Assent in Manitoba, Canada. The Act mandates transparency, accountability structures, and cybersecurity incident reporting for public sector entities using AI systems. Substantive obligations take effect only through regulations yet to be made.