On 20 July 2026, the European Commission published its final Guidelines on the transparency obligations under Article 50 of the EU AI Act. Although non-binding, the Guidelines provide important practical clarification ahead of the application of these obligations. Article 50 is not limited to high-risk AI systems and may apply to a broad range of AI solutions, including chatbots, generative AI tools, emotion recognition systems, biometric categorization tools, and deepfake technologies.

Article 50 is intended to reduce transparency risks such as impersonation, deception, manipulation, and misinformation by ensuring that individuals know when they are interacting with AI or exposed to artificially generated or manipulated content. The Guidelines clarify the respective obligations of providers, which develop, place on the market or put AI systems into service, and deployers, which use AI systems under their own authority.

The Guidelines identify four main transparency scenarios. First, providers of interactive AI systems must inform individuals when they interact with AI, unless this is obvious from the context. Second, providers of systems generating synthetic audio, image, video, or text content must ensure outputs are marked in a machine-readable format and detectable as artificially generated or manipulated. Third, deployers using emotion recognition or biometric categorization systems must inform exposed individuals, subject to applicable prohibitions and data protection requirements. Fourth, deployers must disclose deepfake content and, in certain cases, AI-generated text published to inform the public on matters of public interest.

Transparency information must be clear, distinguishable, and provided no later than the first interaction or exposure. Generic references hidden in terms and conditions, website footers, or vague labels are unlikely to be sufficient.

The territorial scope is broad and may extend to non-EU providers and deployers where AI outputs are used in the EU, although incidental, unforeseeable or unauthorized downstream use should not alone trigger obligations for third-country providers.

Businesses should now map AI use cases, determine whether they act as provider, deployer or both, review user-facing disclosures, assess marking requirements for AI-generated content, and document any reliance on exemptions.