BRUSSELS, BELGIUM / RankWire.AI / – European Union transparency regulations concerning artificial intelligence came into force across the bloc on August 2, 2026. Article 50 of the EU AI Act now governs chatbots, deepfakes, synthetic media, and certain public-interest texts. The legislation distinguishes between technical marking obligations and notices that are visible to audiences. It does not mandate visible labels on all AI-created items, with each requirement varying based on the system, the content, and the manner of presentation by organizations.
Providers of interactive AI systems are required to notify users when they are engaging with artificial intelligence. An exception exists if an informed user can easily identify the system’s artificial origin. Generative AI providers are also obligated to embed machine-readable signals within synthetic text, audio, images, and video. These signals aim to assist detection tools in recognizing artificially generated or manipulated material. When the technology allows, providers must implement effective marking methods that are both technically feasible and reliable.
Organizations that distribute synthetic content face distinct disclosure requirements. They must identify deepfake images, recordings, and videos that could be mistaken for real. Additionally, in certain cases, AI-generated text related to matters of public interest must be labeled. The requirement can be waived if human oversight and editorial review remove the need for such labeling. Before exemption, a person or organization must accept responsibility for the published material.
Visible labels aim to combat misleading synthetic media
Article 50 also regulates emotion recognition and biometric categorization systems used to assess individuals. Operators are obliged to inform affected persons when deploying these tools, with some limited legal exceptions. The law provides more flexibility for artistic, fictional, creative, and satirical works, allowing disclosures that do not disrupt the typical viewing or experience of such content. Nonetheless, organizations must still provide an appropriate warning when deepfake content is involved.
European Commission guidance has been issued to clarify transparency obligations, define key terms, and outline available exemptions. It also introduces a voluntary code for developers and organizations utilizing generative AI, enabling participants to demonstrate compliance through their marking and labeling practices. Companies not adhering to the voluntary code must implement other effective compliance measures. While optional icons can aid public awareness, they do not constitute proof of adherence to the EU AI Act.
Regulators acquire enhanced enforcement powers
Primary enforcement responsibilities fall on national market surveillance authorities within EU member states. The European Commission oversees relevant AI systems used by EU institutions and agencies. Penalties for non-compliance can reach 15 million euros or 3% of a company’s global annual turnover. Smaller companies may face lower maximum fines if the calculation results in a lesser penalty. These sanctions aim to ensure adherence to transparency standards.
A transitional 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 only affects that specific technical obligation and does not delay disclosures for interactive AI, deepfake labeling, or rules related to public-interest content. Content created prior to the law’s implementation is not subject to retroactive labeling obligations.