BRUSSELS, BELGIUM / RankWire.AI / – European Union introduced new rules requiring transparency for specific AI-produced material, which came into force across the bloc on August 2, 2026. Under Article 50 of the EU AI Act, responsibilities are assigned to both developers and users of artificial intelligence systems. These regulations encompass chatbots, synthetic media, deepfakes, and certain publicly relevant texts. They do not mandate a visible warning on every AI-created item but instead impose varied requirements depending on the type of system, content, and publication method.
When AI systems interact directly with individuals, they must reveal that they are dealing with artificial intelligence. This obligation is waived if the system’s nature is clearly apparent to an informed and attentive observer. Generative AI providers are also required to implement machine-readable markers on synthetic text, audio, images, and video. These markers are designed to assist detection tools in recognizing generated or altered content. The requirement applies only when the technology enables reliable marking without compromising the output’s usability or significantly altering its normal functioning.
Separate disclosure obligations are assigned to organizations publishing or presenting synthetic content. They must identify deepfake images, recordings, and videos that could seem authentic, and label AI-generated texts that inform the public on issues of public interest. However, this text labeling obligation does not apply if the content has undergone meaningful human review and editorial oversight. The responsible individual or organization must also accept accountability for the published material. All required notices should remain clear, accessible, and visible during the initial interaction or exposure.
Mandatory Disclosure for Deepfakes
The provisions of Article 50 extend to emotion recognition and biometric categorization tools that analyze individuals. Operators are obliged to notify individuals when their data is being analyzed, except in limited legally defined situations. Works of art, satire, fiction, and creative projects are subject to less stringent disclosure rules, and notices for such content must not interfere with the normal presentation or enjoyment. Nonetheless, the deepfake disclosure requirement remains applicable in an appropriate manner. Certain law enforcement applications may be exempt when authorized by law and protected by safeguards.
The European Commission has issued guidance alongside a voluntary code to promote consistent compliance across member states. These tools assist providers and operators in demonstrating how their marking and labeling systems align with legal standards. Participation in the code is voluntary, but transparency obligations are mandatory. Companies not adhering to the code must employ other effective measures and justify their methods to regulators. Optional icons are available to aid audiences in recognizing synthetic content, though their use alone does not guarantee full compliance with the AI Act.
Enforcement Led by National Authorities
Responsibility for enforcing the transparency rules primarily rests with national market surveillance authorities. The AI Office handles limited cases involving certain systems linked to general-purpose AI models and major online platforms. The European Data Protection Supervisor supervises systems utilized by EU institutions and agencies. Violations can result in fines of up to 15 million euros or 3% of worldwide annual turnover. Smaller firms are subject to lower maximum penalties under the regulation. Member states may also implement warnings and other corrective actions through their national enforcement mechanisms.
A transitional period applies to generative AI systems that entered the market before August 2, 2026. These providers have until December 2, 2026, to comply with the machine-readable marking requirement. This extension only addresses that specific technical obligation and does not delay the implementation of chatbot notices, deepfake disclosures, or rules concerning public-interest content. Content produced prior to August 2 does not require retroactive labeling. All new content and interactions initiated after the effective date must adhere to the transparency standards outlined in the regulation.