AI Governance Three Lines Model (Hong Kong)
Tags: Frameworks
TL;DR
- Governance assigns responsibilities across project team, oversight committees, and executive review.
- High-risk AI requires senior review, monitoring, and optional Ethical AI Committee advice.
Why it matters for HK marketers: Clear decision rights speed approvals and reduce risk for AI-driven campaigns and tools.
Governance model
- First line — Project Team
- Develop AI applications, evaluate risks via AI Assessment, implement mitigations, document throughout the lifecycle.
- Second line — PSC/PAT (Project Steering Committee / Project Assurance Team)
- Define acceptance criteria, provide independent review, ensure Ethical AI Principles are addressed via completed AI Assessment, approve AI applications.
- Third line — IT Board/CIO (with optional Ethical AI Committee)
- Review, advise, and monitor high-risk AI; Ethical AI Committee may include external advisors to strengthen competency.
Approval flow highlights
- PSC/PAT sign-off required before delivery.
- Independent review monitors controls and assessments; ongoing monitoring for high-risk cases.
- Senior management directives and review invoked for high-risk AI applications.
3 lines of defence define roles from build to executive oversight.
So what for marketers
Stand up a lightweight PSC/PAT-equivalent for martech AI and define “high-risk” escalation to CIO/IT Board to avoid late-stage blockers.
← Back to Knowledge Base