AI Regulation ReferenceDoes OECD AI Principles require Risk Assessment?
OECD • voluntary
Yes — 1 provision
Requirements at a glance
This regulation imposes 4 specific requirements for Risk Assessment across 1 provision:
- Risk management — AI systems should not pose unreasonable risks; manage risks throughout lifecycle
- Robustness — Ensure AI systems function as intended under normal and adversarial conditions
- Security — Address cybersecurity risks in AI systems
- Traceability — Enable traceability of AI system outcomes to data and processes
Robustness, Security and Safety (Principle 1.4) #
Requirements
| Requirement | Details |
|---|---|
| Risk management | AI systems should not pose unreasonable risks; manage risks throughout lifecycle |
| Robustness | Ensure AI systems function as intended under normal and adversarial conditions |
| Security | Address cybersecurity risks in AI systems |
| Traceability | Enable traceability of AI system outcomes to data and processes |
Penalties
| Violation | Fine |
|---|---|
| Non-compliance | Voluntary — no binding enforcement mechanism |