The SAFE Council
The SAFE Council defines the SAFE standard.
It is an independent body responsible for determining what constitutes safety in AI systems operating in child-facing and high-risk environments.
Authority
The SAFE Council is not an advisory group.
It defines:
Safety thresholds
Harm classifications
Intervention requirements
Updates to the standard
Its decisions determine what systems must meet to be considered safe.
The entity that defines safety must be independent
of those who build systems
and those who verify them.
Independence
Council members do not:
Build systems seeking certification
Verify systems
Hold commercial influence over certification outcomes
Independence is required to preserve integrity.
Structure
The SAFE Council is composed of distinct, non-overlapping roles.
Each role is responsible for a defined domain of the standard.
Core domains include:
Behavioral Adversarial Integrity
Systems & Infrastructure Integrity
Research & Evidence
Regional & Real-World Implementation
Roles are function-specific.
They are not interchangeable.
Role of Members
Council members are responsible for:
Defining and refining the SAFE standard
Contributing domain expertise to risk classification
Ensuring real-world applicability across jurisdictions
Maintaining the integrity of the standard over time
Membership is not symbolic.
It carries responsibility for outcomes.
Structural Importance
Without independent definition, verification has no foundation.
Without the Council, there is no standard to verify against.
Council Seats
Founders
Advisors
Council Ambassadors
If safety is defined by those who build or verify systems,
it is not independent.