BRUSSELS, BELGIUM / RankWire.AI / – In a move towards increased transparency in artificial intelligence, new regulations from the European Union became effective on August 2, 2026. These rules, outlined in Article 50 of the EU AI Act, specify how companies must identify certain AI-driven interactions and artificially created content. The legislation encompasses chatbots, deepfakes, generated media, and some public-interest texts. However, it does not mandate visible labels on all AI-produced materials. Instead, responsibilities are tailored based on the system, its output, and how an organization presents it.
Entities operating interactive AI systems are required to disclose when they are engaging with artificial intelligence. This requirement is waived if the average user can easily recognize the system’s artificial nature. Additionally, providers of generative AI must embed machine-readable information within synthetic texts, images, audio, and video. The purpose of this data is to facilitate automated detection of manipulated or generated content. These measures must be implemented when current technology can reliably and effectively perform such marking.
Organizations distributing synthetic media are subject to separate disclosure regulations. They are obligated to identify deepfake images, audio, or video that could be mistaken for real. Moreover, they must label AI-generated content related to public interest topics in specific situations. This labeling requirement is not applicable when individuals review the material and exercise editorial oversight. Before this exemption applies, a designated person or organization must accept responsibility for the publication.
Public disclosure rules emphasize deepfakes and civic information
The regulations also extend to technologies like emotion recognition and biometric categorization systems that evaluate individuals. Operators are required to inform affected persons when employing such technologies, except in limited cases authorized by law. The EU AI Act allows greater flexibility in disclosing creative, fictional, artistic, and satirical works. Notices should preserve the usual presentation and enjoyment of these works, but operators must still provide an appropriate disclosure if a piece contains deepfake material.
European Commission guidance clarifies the scope of Article 50 and its exceptions. A voluntary code has also been issued for companies developing or deploying generative AI systems. This code helps demonstrate compliance with labeling and marking standards. Companies that opt not to participate are expected to implement other effective measures to ensure compliance. While optional icons can support public awareness, they do not automatically signify adherence to regulations.
Authorities have the power to impose substantial fines
Supervisory agencies across EU member states will oversee enforcement, led by national market authorities. The AI Office will focus on a limited set of systems associated with general-purpose AI models and large platforms. The European Data Protection Supervisor will monitor relevant systems used by EU institutions and agencies. Penalties for violations can reach up to 15 million euros, or 3% of a company’s global annual turnover.
Existing generative AI systems on the market before August 2 are granted a transitional period. Their providers have until December 2, 2026, to comply with the machine-readable marking requirement. This extension applies solely to the technical obligation of marking synthetic content and does not delay notices for chatbots, deepfake disclosures, or rules about public-interest texts. Additionally, organizations are not obliged to label content created prior to the enforcement of transparency rules.