EU AI Act: Article 5
The regulations detailed in the EU AI Act distinguish AI systems into three (3) categories, viz., Prohibited AI, High-Risk AI, and Moderate/Low Risk AI.
Title II of the EU AI Act establishes the list of prohibited AI. The list includes AI systems that, for example, could violate fundamental rights, manipulate persons through subliminal techniques, or assist in social scoring of natural humans for general purposes.
The control questions, generated from the texts in Articles 4, 4a, 4b, and 5, are meant to assist an organization in properly classifying the AI systems under development or under consideration for deployment or being put into market.