The European Union has moved its Artificial Intelligence Act into a new enforcement phase, with transparency obligations applying from 2 August 2026. The European Commission’s AI Office and national authorities can now supervise compliance by providers and deployers of covered systems rather than treating the rules as a future implementation target.
Chatbots, AI agents, avatars and other interactive systems must clearly inform users when they are interacting with a machine rather than a person. The obligation focuses on the identity of the interaction and is meant to prevent users from mistaking an automated service for human communication.
Defined forms of generated or manipulated content also require disclosure. Deepfake images, audio and video that resemble real people, objects, places, entities or events must be visibly labelled and carry machine-readable marks. The technical mark is intended to remain detectable even when a person encounters the content outside its original platform.
The transparency regime also covers emotion-recognition and biometric-categorisation tools, as well as AI-generated text published to inform the public on matters of public interest when it has not undergone human review or editorial control. The requirements therefore do not impose an identical label on every use of AI; scope depends on the system, content and deployment context.
Enforcement is divided among national market-surveillance authorities, the European AI Office for systems under its supervision and the European Data Protection Supervisor when EU institutions are providers or deployers. The Commission has also opened complaint, whistleblower and downstream-provider channels to bring suspected non-compliance into the supervisory process.
Companies can face fines of up to EUR 15 million or 3% of global annual turnover for breaches of the transparency obligations. EU institutions, bodies and agencies can face up to EUR 750,000. The framework requires proportionality for small and medium-sized enterprises and small mid-cap companies rather than applying the maximum mechanically.
The Commission has issued guidelines, icons and a transparency code of practice to operationalise compliance. More than 180 organisations appeared on its first list of code signatories. Participation can help demonstrate a compliance method, but companies remain responsible for whether their systems and published content satisfy the binding legal requirements.
The AI Act entered into force in August 2024 and applies in stages, making the August 2026 date one milestone rather than the beginning of the entire law. The practical tests now are consistent labels across platforms, durable machine-readable marking and aligned enforcement among EU and national bodies when synthetic content crosses borders.
