Article 50 of Regulation (EU) 2024/1689 (AI Act) entered into application on 2 August 2026, imposing transparency obligations on providers and deployers of AI systems. On the same date, the Commission's enforcement powers over providers of general-purpose AI (GPAI) models became fully effective, covering both transparency documentation and copyright due-diligence obligations under Articles 53 and 55. These provisions are distinct from the high-risk AI system requirements under Chapter III, which the AI Act Omnibus deferred to 2 December 2027 for standalone systems and 2 August 2028 for AI embedded in products.
Article 50(1) requires providers of AI systems designed to interact with natural persons, including chatbots and virtual assistants, to disclose to users that they are interacting with an AI, unless the context makes this obvious. Article 50(4) requires providers and deployers of systems generating synthetic audio, video, text, or images to mark such outputs as artificially generated or manipulated, using machine-readable metadata and, where technically feasible, watermarking. GPAI model providers under Articles 53 and 55 must publish technical documentation, a copyright compliance policy disclosing training data summaries, and, for models classified as presenting systemic risk, adversarial testing results. A limited grace period until 2 December 2026 applies only to the AI-generated content marking requirement for systems placed on the market before 2 August 2026.
AI system deployers across the EU, including legal technology platforms, HR screening tools, content generation services, customer service operators, and synthetic media producers, are now subject to Article 50. GPAI model providers distributing models via APIs to EU users must have published their technical documentation and copyright policy. The Commission may impose fines of up to 3 percent of worldwide annual turnover on GPAI model providers for violations of Articles 53 and 55 under Article 99, and up to 1 percent for supplying incorrect, incomplete, or misleading information.
GPAI models already placed on the market before 2 August 2025 have until 2 August 2027 to achieve full compliance with Articles 53 and 55. No grace period applies to the AI interaction disclosure requirement under Article 50(1). The Commission has not yet published a final list of GPAI models with systemic risk classification under Article 51, meaning providers of high-capability models remain uncertain whether additional adversarial testing obligations apply to them pending that determination.
Licentium advises AI system providers and deployers on EU AI Act transparency compliance, including the preparation of Article 50 disclosure notices, technical documentation for GPAI models, copyright compliance policies, and watermarking assessment. Work we undertake includes EU AI Act transparency advisory, GPAI model compliance, AI-generated content regulation, synthetic media disclosure requirements, and digital product legal compliance across EU member states.