BRUSSELS, BELGIUM / RankWire.AI / – Effective as of August 2, 2026, the European Union has introduced new transparency standards for artificial intelligence, outlined in the EU AI Act, with Article 50 now specifying how organizations must identify certain AI-driven interactions and generated content. This legislation encompasses chatbots, deepfakes, media created artificially, and some texts of public interest, yet it does not mandate conspicuous labels on all AI-produced material; rather, the obligations vary depending on the specific system, its output, and the manner in which an organization presents it.
Entities operating interactive AI solutions are required to disclose when they are engaging with artificial intelligence, except in cases where a typical user can easily discern the artificial nature of the system. Furthermore, creators of generative AI must embed machine-readable information within synthetic text, images, audio, and video, designed to facilitate automated detection of manipulated or generated media. This requirement applies when current technology can reliably and effectively support such marking, ensuring the information is accessible for detection purposes.
Distributors of synthetic media are subject to additional public disclosure rules, which mandate the identification of deepfake images, audio, or videos that might be mistaken for authentic content. They must also label AI-generated texts concerning public interest topics under certain conditions. However, this labeling obligation is waived if individuals review the material and exercise editorial control, provided that a specific person or organization accepts responsibility for the publication before this exemption applies.
Focus on Deepfakes and Civic Information in Public Disclosures
These regulations extend to emotion recognition and biometric categorization systems that evaluate individuals, requiring operators to notify affected persons when such technologies are employed, except in limited situations authorized by law. The EU AI Act also grants greater flexibility in disclosing creative, fictional, artistic, and satirical works, allowing notices to maintain the normal presentation and experience of these works while still mandating clear disclosure if deepfake content is involved.
The European Commission has issued guidance clarifying the scope of Article 50 and outlining its exceptions, alongside a voluntary code aimed at companies developing or deploying generative AI systems. Participation in this code allows firms to demonstrate their adherence to legal standards through their labeling and marking strategies, although non-participation means relying on other effective compliance measures. While optional icons can aid public awareness, these symbols alone do not constitute compliance with the regulation.
Regulatory Bodies Are Authorized to Impose Heavy Penalties
Enforcement will primarily be led by national market surveillance authorities within EU member states, with the AI Office overseeing a select group of systems related to general-purpose AI models and major platforms, and the European Data Protection Supervisor monitoring relevant systems used by EU institutions and agencies. Violations may result in fines of up to 15 million euros, or penalties amounting to 3% of a company’s total worldwide annual turnover.
For generative AI systems already available before August 2, there is a transitional period extending until December 2, 2026, giving providers until that date to comply with the machine-readable marking obligation. This extension solely applies to the technical requirement for identifiable synthetic content and does not delay notices about chatbots, disclosures for deepfakes, or rules related to public-interest texts. Additionally, the law does not obligate organizations to label content created prior to the enactment of these transparency provisions.