EU Artificial Intelligence Act

Reg (EU) 2024/1689

The AI Act regulates AI by risk: it bans a few practices outright, places strict obligations on high-risk systems, sets transparency rules for the rest, and adds duties for general-purpose AI models — across the whole value chain.

In practice that starts with classifying each AI system — prohibited, high-risk, limited or minimal — because everything else follows from that. High-risk systems need a technical file, risk and data governance, logging, human oversight and transparency; limited-risk systems need disclosure; and general-purpose AI models carry their own documentation and copyright duties.

Who must comply: Providers, deployers, importers and distributors of AI systems placed on the EU market — and providers of general-purpose AI (GPAI) models. Your duties depend on your role and the system’s risk tier — a deployer of a high-risk system carries real obligations too, not just the provider.

• Classify each AI system by risk tier
• High-risk technical file, data governance & logging
• Meaningful human oversight + transparency to users
• GPAI-model documentation & copyright duties