BRUSSELS, BELGIUM / RankWire.AI / – European Union regulations requiring clear disclosure for certain AI-produced materials became active across member states on August 2, 2026. Under Article 50 of the EU AI Act, obligations are established for both developers and users of artificial intelligence technologies. These rules encompass chatbots, synthetic media, deepfakes, and specific publicly relevant texts. However, they do not mandate a visible warning on every AI-created item. Instead, the legislation applies distinct requirements depending on the type of system, the nature of the content, and the mode of publication.
When AI systems engage directly with individuals, they must inform users that they are interacting with artificial intelligence. This obligation does not apply if the system’s artificial origin is clearly evident to an informed and attentive person. Providers of generative AI are also required to incorporate machine-readable markers into synthetic text, audio, images, and videos. These markers are designed to assist detection tools in identifying artificially generated or manipulated content. The requirement is valid when the technology enables reliable marking without impairing the output’s usability or significantly altering its normal operation.
The legislation stipulates separate disclosure duties for organizations that publish or display synthetic media. They must label deepfake images, recordings, and videos that could be mistaken for authentic material. Additionally, AI-generated text that provides information on matters of public interest must be marked accordingly. However, such labeling is not necessary after a meaningful human review and editorial oversight. Responsibility for publication must be clearly assigned to a person or organization. The required notices should remain visible, accessible, and unambiguous during initial exposure or interaction.
Mandatory disclosure for deepfakes
Article 50 also addresses the use of emotion recognition and biometric categorization tools involving individuals. Operators must alert users when these systems analyze them, subject to limited legal exceptions. The regulation specifies that creative, artistic, satirical, and fictional works are subject to more limited disclosure requirements. Notices for such content must not interfere with the normal viewing or enjoyment experience. Nonetheless, the obligation to disclose deepfake content remains applicable in an appropriate manner. Certain law enforcement applications may qualify for specific exemptions, provided they are authorized by law and safeguarded by appropriate measures.
The European Commission has issued guidelines and a voluntary code aimed at ensuring consistent compliance among member states. These resources can assist providers and users in demonstrating how their marking and labeling strategies meet the legal standards. Adoption of the code is voluntary, although the transparency requirements are mandatory. Companies outside the scope of the code are expected to employ other effective measures and justify their approach to regulators. Although optional icons may assist audiences in recognizing synthetic content, relying solely on an icon does not fulfill full compliance with the AI Act.
National authorities to oversee enforcement
Primary responsibility for enforcing transparency rules rests with national market surveillance authorities. The European Commission also supervises specific cases involving certain systems related to general-purpose AI models and major online platforms. The European Data Protection Supervisor oversees systems utilized by EU institutions and agencies. Violations can result in fines of up to 15 million euros or 3% of global annual turnover. Smaller companies are subject to lower maximum penalties according to the regulation. Member states may also deploy warnings and other corrective measures through their enforcement mechanisms.
A transitional period applies to generative AI systems introduced prior to August 2, 2026. Providers of these systems have until December 2, 2026, to comply with the machine-readable marking requirement. This extension solely covers the technical marking obligation and does not delay other disclosures such as chatbot warnings, deepfake labels, or public-interest content rules. Content created before August 2 is not subject to retroactive labeling. Moving forward, all new content and interactions must adhere to the transparency obligations set out by the regulation from the date it came into force.