The rim · governance
CORE, an AI governance framework: the unlock, not the brake.
Four pillars that run across every stage. The weakest pillar sets your trust ceiling.
Updated
Four pillars
No owner, no system.
From The AI Flywheel, Chapters 14 to 17.
- C
Compliance and Fairness
Lawful and fair by design: risk tiers, fairness tests, data residency.
Tool: Compliance and Sovereignty Readiness Checklist · Ch 15
- O
Openness and Transparency
Disclose the AI, document it with model and agent cards, keep a trace.
Tool: Model card / Agent card · Ch 16
- R
Responsibility and Safety
A named owner, guardrails and a kill switch you have tested.
Tool: Agent Accountability Register · Ch 16–17
- E
Empowerment and Human Agency
Human judgment stays in command. Autonomy is a dial, not a switch.
Tool: Autonomy-Level Decision Guide · Ch 17
Quick answers
CORE, answered.
What does CORE stand for?
Compliance and Fairness, Openness and Transparency, Responsibility and Safety, and Empowerment and Human Agency. CORE is the four-pillar governance discipline at the rim of the AI Flywheel, and it runs across all four stages of AI (The AI Flywheel, Ch 14).
What is the trust ceiling?
The trust ceiling is the height of value and autonomy at which an organisation will, and may, deploy AI. The weakest CORE pillar sets it: weak governance keeps it low, mature governance raises it. Governance is the unlock, not the brake.
Next
An AI competency framework: five levels in each stage.
From Aware (C1) to Architect (C5). Each level is earned only with evidence from real work.