BRUSSELS, BELGIUM / RankWire.AI / – Effective from August 2, 2026, the European Union has implemented new transparency regulations for artificial intelligence. Under Article 50 of the EU AI Act, companies are now required to identify certain AI interactions and synthetic media. These regulations encompass chatbots, deepfakes, generated visual and audio content, as well as some publicly relevant textual materials. However, not all AI-produced content must carry overt labels. Instead, obligations vary based on the system, the nature of the output, and how the organization presents it.
Any business operating interactive AI technologies must inform users when they are engaging with artificial intelligence. This requirement is waived when a typical user can easily recognize the system’s artificial origins. Providers of generative AI must also include machine-readable data within synthetic images, audio, text, and video to facilitate automated detection of manipulated or generated content. Such measures should be implemented when current technology can reliably support effective labeling.
Distinct disclosure rules apply to organizations distributing synthetic media. They are mandated to identify deepfake images, audio, or video that could be mistaken for real. Additionally, AI-generated content concerning public-interest issues must be labeled under certain circumstances. These labeling obligations do not apply if the material is reviewed and editorially controlled by a person or organization, which must also accept responsibility for the publication before the exemption applies.
Public disclosure requirements emphasize deepfakes and civic-related information
The regulations also cover systems that utilize emotion recognition and biometric categorization to assess individuals. Operators are required to notify those affected when these technologies are deployed, with exceptions only in limited legal circumstances. The EU AI Act provides more lenient disclosure rules for creative, fictional, artistic, and satirical works, allowing disclosures that do not interfere with their typical presentation or enjoyment. Nonetheless, when a work includes deepfake content, operators must still provide an appropriate warning.
The European Commission issued guidance clarifying the scope of Article 50 and its exceptions. A voluntary code has also been published for companies developing or deploying generative AI. This code allows participants to demonstrate how their labeling and marking processes fulfill the legal requirements. Companies opting out must rely on other effective compliance methods. While optional icons can help raise public awareness, these symbols alone do not constitute compliance.
Authorities may impose hefty fines for violations
Enforcement will primarily be carried out by national market supervision agencies within each EU member state. The AI Office will oversee a narrower set of systems, specifically those associated with general-purpose AI models and major platforms. The European Data Protection Supervisor will monitor systems used by EU institutions and agencies. Penalties for non-compliance can reach up to 15 million euros or 3% of a company’s total global annual turnover.
Existing generative AI systems in the market prior to August 2 benefit from a limited transition period. Providers have until December 2, 2026, to comply with the machine-readable marking obligation. This extension applies solely to the technical requirement for detectable synthetic content and does not delay notices related to chatbots, deepfake disclosures, or public-interest textual content. Additionally, organizations are not required to label content created before the law’s transparency rules came into force.