The European Commission began enforcing a new set of transparency obligations under the European Union’s Artificial Intelligence Act on August 2, 2026, marking one of the most consequential milestones yet in the bloc’s phased rollout of the world’s first comprehensive AI regulatory framework.
Under the new rules, providers of chatbots and other interactive AI systems must now clearly inform users that they are communicating with a machine rather than a human being, unless this is already obvious from the context. The obligation is designed to close a gap that critics say has allowed AI-driven customer service tools, virtual assistants, and companion applications to blur the line between human and automated interaction without users’ informed consent.
The rules also introduce mandatory labeling requirements for synthetic and manipulated media. Images, audio, and video that have been generated or significantly altered by artificial intelligence, including so-called deepfakes depicting real people, places, or events, must now carry a clear, visible, and machine-readable mark identifying them as AI-generated or edited. The European Commission said the goal is to help ordinary citizens recognize when content has been artificially produced and to make informed choices accordingly.
« The measures are intended to reduce deception and manipulation and help people make informed choices, » the Commission said in a statement announcing the start of enforcement. « They also give businesses clearer obligations and a practical way to show compliance. »
Enforcement will be carried out jointly by the Commission’s newly established AI Office and national market surveillance authorities across the 27 member states. Companies found in breach of the transparency obligations face fines of up to €15 million or 3 percent of their total worldwide annual turnover, whichever is higher. EU institutions, bodies, and agencies that fail to comply face penalties of up to €750,000, while small and medium-sized enterprises will benefit from a more proportional enforcement approach that accounts for their limited resources.
To ease the transition, more than 180 companies and organizations, including major technology firms, have already signed onto a voluntary Code of Practice on transparency, which offers a structured pathway toward compliance ahead of stricter enforcement. Brussels has also published detailed guidelines and standardized icons that developers can use to label AI-generated content across platforms.
Monday’s milestone follows earlier phases of the AI Act’s implementation, including the ban on a narrow set of « unacceptable risk » AI practices that took effect in February 2025 and obligations on providers of general-purpose AI models that began in August 2025. The transparency rules taking hold this week apply more broadly across the AI value chain, touching not only model developers but also any business or public body deploying AI tools that interact with the public or generate synthetic content.
With the transparency obligations now in force, attention in Brussels is turning to the next and arguably most demanding phase of the Act: the high-risk AI system requirements, which are set to apply from August 2027 and will impose strict conformity assessments on AI used in areas such as employment, law enforcement, and critical infrastructure.
