BRUSSELS, BELGIUM / RankWire.AI / – Effective as of August 2, 2026, new European Union rules regarding transparency in artificial intelligence are now in force. Article 50 of the EU AI Act outlines how organizations must identify certain AI-driven interactions and artificially generated material. The legislation encompasses chatbots, deepfakes, media created by AI, and specific public-interest texts. It does not mandate prominent labels on all AI-produced content but instead assigns duties based on the nature of the system, its output, and how it is presented by the organization.
Entities operating interactive AI platforms are required to disclose when users are communicating with artificial intelligence. This obligation does not apply if the system’s artificial origin is immediately obvious to a typical user. Providers of generative AI must also embed information that can be read by machines within synthetic media such as text, images, audio, and video. This data aims to facilitate automated detection of altered or generated content. The implementation of this requirement depends on the availability of reliable and effective marking technology.
Organizations distributing synthetic media are subject to distinct public disclosure obligations. They need to identify deepfake visuals, audio, or videos that could be mistaken for genuine. They are also required to label AI-generated text concerning public-interest issues in certain scenarios. However, this requirement does not apply if the material is reviewed and edited by a person or organization. Responsibility for publication must be accepted by a named individual or entity before this exemption can be claimed.
Disclosures in public focus on deepfakes and civic information
The regulations extend to emotion recognition and biometric categorization tools that analyze individuals. Operators must notify affected parties when employing these technologies unless legally authorized to do otherwise. The EU AI Act offers more leniency for creative, fictional, artistic, and satirical works, allowing disclosures that do not disrupt their usual presentation or enjoyment. Nonetheless, operators are still obliged to clearly disclose when a work contains deepfake elements.
European Commission has issued guidance clarifying the scope of Article 50 and its exemptions. Additionally, it has published a voluntary code for organizations involved in developing or deploying generative AI. This code helps demonstrate how their labeling and marking procedures comply with legal standards. Companies choosing not to participate must rely on other effective compliance strategies. While optional icons can help raise public awareness, they do not by themselves establish legal adherence.
Authorities have the power to levy substantial fines
Enforcement will be primarily carried out by national market oversight authorities across EU member states. The European Data Protection Supervisor will oversee systems connected to EU institutions and agencies. The AI Office will supervise a narrower range of systems, including those related to general-purpose AI models and major platforms. Penalties for violations can reach up to 15 million euros, or 3% of a company’s global annual turnover, whichever is higher.
Systems in the generative AI market prior to August 2 are granted a limited transition period. Providers have until December 2, 2026, to comply with the machine-readable marking requirement. This extension only applies to the technical aspect of detectable synthetic content and does not delay notices about chatbots, disclosures related to deepfakes, or rules concerning public-interest content. The legislation also clarifies that organizations are not required to label material created before the transparency rules came into effect.