BRUSSELS, BELGIUM / RankWire.AI / – Effective from August 2, 2026, the European Union has introduced updated transparency regulations for artificial intelligence. Under Article 50 of the EU AI Act, companies now must specify how they identify certain AI interactions and synthetic content. This legislation encompasses chatbots, deepfakes, generated media, and some public-interest texts, but it does not mandate visible labels for all AI-produced material. Instead, obligations vary depending on the system, output, and how an organisation presents it.
Operators of interactive AI systems are required to notify users when they are engaging with artificial intelligence, except when an average user can easily recognize the AI’s artificial nature. Generative AI providers must also embed machine-readable data in synthetic texts, images, audio, and video to facilitate automated detection of manipulated or generated content. These markings are mandatory where available technology can reliably and effectively implement them.
Entities distributing synthetic media must adhere to separate public disclosure requirements, which involve identifying deepfake images, audio, or video that may seem authentic. They must also label AI-created content related to public interest issues under specific circumstances. If the material is reviewed and editorial control is exercised by individuals, this labeling obligation is waived, provided a named person or organisation accepts responsibility for publication beforehand.
Focus on Deepfakes and Civic Discourse
The regulations also extend to emotion recognition and biometric categorisation technologies that evaluate individuals. Operators are obliged to notify affected persons when such systems are used unless legally authorized exceptions apply. The EU AI Act provides more leniency for creative, fictional, artistic, and satirical works, allowing disclosures to be made in a manner that does not disrupt the normal presentation and enjoyment. Nonetheless, a proper notice must still be issued if a work includes deepfake elements.
The European Commission has issued guidance clarifying the scope of Article 50 and its exemptions, alongside a voluntary code for companies involved in developing or deploying generative AI. This code enables firms to demonstrate how their labeling and marking processes fulfill legal standards. Those choosing not to participate must implement other effective compliance measures, with optional icons available to raise public awareness—though such symbols alone do not constitute compliance.
Enforcement and Penalties for Non-Compliance
Supervision across EU member states will be conducted by national market surveillance authorities, while the AI Office will oversee a select group of systems related to general-purpose AI models and large platforms. The European Data Protection Supervisor will monitor systems used by EU institutions and agencies. Violations can result in fines reaching up to 15 million euros or penalties equivalent to 3% of a company’s global annual turnover.
Existing generative AI systems on the market before August 2 benefit from a transitional period until December 2, 2026, to comply with the machine-readable marking requirement. This extension pertains solely to the technical aspect of marking synthetic content but does not postpone chatbot notices, deepfake disclosures, or rules concerning public-interest texts. Furthermore, organizations are not obligated to label content created prior to the implementation of transparency provisions.