BRUSSELS, BELGIUM / RankWire.AI / – European Union has implemented new rules aimed at increasing transparency for certain types of AI-produced material, which became effective across member states on August 2, 2026. Under Article 50 of the EU AI Act, obligations now apply to both developers and users of artificial intelligence systems. These regulations encompass chatbots, synthetic media, deepfakes, and specific public-interest texts. They do not mandate a visible warning on each AI-created item, but instead impose varied requirements depending on the system used, the content produced, and the method of dissemination.
When AI systems interact directly with humans, they are required to clearly indicate that the user is engaging with artificial intelligence. This obligation does not extend when the system’s nature is obvious to an informed and attentive person. Generative AI providers must also incorporate machine-readable markers into synthetic texts, audio, images, and videos. These markers assist detection tools in identifying generated or manipulated content. The requirement applies when the technology can reliably mark content without impairing its usability or fundamentally altering its normal operation.
Separate disclosure duties are placed on organizations that publish or present synthetic content. They are responsible for identifying deepfake images, recordings, and videos that could be mistaken for authentic. They must also label AI-generated texts that inform the public about matters of public interest. However, this requirement does not apply once a meaningful human review and editorial control have been established. The responsible person or organization must also accept accountability for the publication. Notices must remain clear, accessible, and visible during the initial interaction or exposure.
Specific disclosure obligations for deepfakes
The regulation also covers emotion recognition and biometric categorization tools used with individuals. Operators are required to notify individuals when these systems analyze them, with some limited legal exceptions. Works that are creative, artistic, satirical, or fictional are subject to narrower disclosure rules, which should not interfere with their normal display or enjoyment. Nevertheless, the deepfake disclosure requirements still apply in a suitable manner. Certain law-enforcement applications may qualify for exemptions when authorized by law and protected by safeguards.
The European Commission has issued guidance and a voluntary code to promote uniform compliance among member states. Providers and users can adopt the code to demonstrate how their marking and labeling systems meet legal standards. While participation is voluntary, the transparency duties themselves are mandatory. Companies outside the code must employ other effective measures and explain these to regulators. Optional icons may assist audiences in recognizing synthetic content, but merely using an icon does not ensure full adherence to the AI Act.
Enforcement led by national authorities
Responsibility for enforcing these transparency regulations primarily falls to national market surveillance authorities. The AI Office oversees limited cases involving certain systems linked to general-purpose AI models and major online platforms. The European Data Protection Supervisor monitors systems used by EU institutions and agencies. Violations can result in fines of up to 15 million euros or 3% of global annual turnover. Smaller companies face lower maximum penalties under the regulation. Member states may also utilize warnings and other corrective actions through their enforcement mechanisms.
A brief transition period applies to generative AI systems introduced before August 2, 2026. Providers of such systems have until December 2, 2026, to comply with the machine-readable marking requirement. This extension addresses only that specific technical obligation and does not delay the implementation of chatbot notices, deepfake disclosures, or rules for public-interest content. Content created before August 2 is not subject to retroactive labeling, but all new content and interactions must adhere to the transparency requirements from the date the rules became effective.