Transparency Obligation

What is the Transparency Obligation?

The Transparency Obligation under the AI Act requires that certain AI systems or their outputs be identifiable as such to humans, regardless of whether the system itself is classified as a high-risk AI system.

When transparency is required

The obligation applies, among other things, to chatbots, which must be identifiable to users as AI systems unless this is already obvious from the context of use. It also applies to AI-generated or AI-manipulated image, audio, or video content such as deepfakes, as well as to emotion recognition and biometric categorization systems.

Practical significance

The Transparency Obligation primarily applies to providers and deployers of the affected systems, regardless of the risk classification of the AI system itself. It complements the more specific obligations for high-risk AI systems by introducing a general requirement that certain AI systems and their outputs be recognizable.

Example

A chatbot on a company’s website informs users at the beginning of the conversation that its responses are generated by an AI system. Likewise, AI-generated images may be subject to transparency obligations depending on the nature of the content.

Further reading

Related terms
→ AI Act
→ Provider
→ Deployer