Orientation
The EU Artificial Intelligence Legal Ecosystem — Map
A layered map of the instruments in this area and how they stack.
The EU artificial intelligence legal ecosystem
Work in progress. This area of the knowledge base is a structured stub: the instrument list and key dates below are drawn from primary sources, but the deep dives are still being written. Anything marked ⚠️ VERIFY has not been independently confirmed.
EU AI law is anchored by a single horizontal instrument — the AI Act — with a ring of product-safety and liability rules around it:
| Layer | Question it answers | Main instruments |
|---|---|---|
| 1. Horizontal AI rules | What AI may be placed on the EU market, and under what obligations? | AI Act (risk-based: prohibited / high-risk / GPAI / transparency) |
| 2. Product safety | Is AI embedded in machinery and products safe? | Machinery Regulation, sectoral product law |
| 3. Liability | Who pays when AI causes damage? | New Product Liability Directive (software & AI in scope) |
| 4. Adjacent | Data, platforms, cybersecurity of AI | GDPR (see Data area), DSA (see Digital area), AI Act Art. 15 cybersecurity (see Cyber area) |
Where to start
- AI Act — the risk pyramid and the staged application timeline (prohibitions → GPAI → high-risk).
- New Product Liability Directive — strict liability now clearly covers software and AI systems.
- Machinery Regulation — safety requirements for machinery with self-evolving behaviour.