Transparency and AI Literacy Duties Were Not Deferred

The postponement of the AI Act's high-risk obligations is being read more broadly than it should be, and analysts tracking the Digital Omnibus have started saying so directly. The deferral moved two specific deadlines. It left the rest of the Act where it was.

Article 50, the transparency chapter, is unaffected. Systems that interact directly with people must make clear that a person is dealing with an AI system unless that is obvious from context. Synthetic audio, image, video and text content must be marked in a machine-readable format. Emotion recognition and biometric categorisation systems must inform the people exposed to them. Deep fake content must be disclosed as artificially generated. None of these duties moved.

Article 4, the AI literacy duty, is likewise untouched and has applied since 2 February 2025. It requires providers and deployers to take measures ensuring a sufficient level of AI literacy among staff and others operating AI systems on their behalf, taking into account their technical knowledge, experience and the context of use. It is one of the few obligations in the Act that falls on organisations simply using AI rather than building it.

Also unchanged are the prohibitions in Article 5, in force since February 2025, and the general-purpose AI model obligations that began applying in August 2025.

The phrase being used is deferred, not cancelled, and it is doing real work. An organisation that reads the headline as a general reprieve and stops its AI Act programme will find that several duties applying to it today were never part of the deferral, and that the high-risk work it paused has a hard date at the end of it.