EU AI Act

The EU AI Act regulates artificial intelligence by risk. It bans a small set of unacceptable uses, puts strict obligations on high-risk systems, adds transparency duties for limited-risk uses, and leaves minimal-risk systems largely free. To comply you first need to know which AI systems you use and where each one falls. turingsecure helps you inventory those systems and manage the controls around them.

Regulation by Risk

Four Risk Classes, Four Levels of Duty

The AI Act sorts every system into one of four risk classes. Unacceptable-risk practices, such as social scoring or manipulative systems, are prohibited outright. High-risk systems, for example those used in recruitment, credit scoring or critical infrastructure, carry the bulk of the obligations. Limited-risk systems, such as chatbots, mainly owe transparency: people must know they are dealing with AI. Minimal-risk systems face no new duties.

The practical first step is the same whatever you build or buy: you cannot classify or govern AI you have not catalogued. An inventory of the AI systems in use, with their purpose and risk class, is the foundation everything else rests on.

Your Duties

What the AI Act Requires of You

Risk Classification: Determine the risk class of every AI system you develop or deploy. The class, from unacceptable through high, limited and minimal, decides which obligations apply.
AI System Inventory: Keep an inventory of the AI systems in use, with their purpose, provider, risk class and the role you play as provider or deployer.
Conformity Assessment: For high-risk systems, run a conformity assessment and maintain the technical documentation, risk management and logging the Act requires before and after they go live.
Transparency and Human Oversight: Tell people when they interact with AI or see AI-generated content, and keep high-risk systems under meaningful human oversight.