SAFE Governance
SAFE governance is independent by design.
The entity that defines safety
is not the entity that builds systems
and is not the entity that verifies them.
The Failure
AI systems are governed by the organisations that build them.
Developers define safety thresholds.
Companies assess their own systems.
Compliance is internally interpreted.
The builder is also the assessor.
This creates structural conflict.
No entity can define, implement, and verify its own safety.
Structural Separation
SAFE introduces separation of roles.
SAFE
Defines what safety is
SAFE Labs
Verifies systems independently
SAIL
Implements safety within systems
These roles cannot be combined.
Role of the Council
The SAFE Council defines the SAFE standard.
It is responsible for:
Defining safety thresholds
Determining harm classifications
Setting intervention requirements
Governing updates to the standard
The SAFE Council is not advisory. It holds authority over the standard.
Independence
Council members do not build systems seeking certification.
They do not verify systems.
They do not hold commercial influence over outcomes.
Independence is required to preserve integrity.
Enforcement
Governance without verification cannot enforce safety.
SAFE governance is inseparable from SAFE Labs certification.
Without independent verification, governance has no authority.
Governance Whitepaper
The complete governance standard in detail.
Structural separation — definition, verification, implementation
SAFE-001 risk taxonomy — ten behavioural harm stages (P1–P10)
Council authority, validation gates, and capture prevention
Certification tiers — SAFE-Ready, SAFE-Certified, SAFE-Endorsed
Cross-jurisdictional regulatory alignment
If safety is not independently defined and enforced,
it is not safety.