Programme 01
AI Governance & Intelligent Systems
How institutions can adopt AI systems with documented oversight, accountability and limits that are stated rather than assumed.
Research objectives
- 1.Establish assurance requirements for AI-assisted institutional workflows.
- 2.Define what evidence an AI-assisted decision must leave behind.
- 3.Identify where human sign-off must remain mandatory.
Research questions
- —Which decisions should never be delegated to automated systems?
- —What documentation makes an AI-assisted outcome reviewable after the fact?
- —How should model limitations be recorded and communicated to operators?
Areas of investigation
- —AI risk management and assurance frameworks
- —Human-in-the-loop control design
- —Decision logging and explainability
Methodology
Comparative review of published standards and institutional guidance, translated into control requirements and reference architecture.
Publications
- —Reference notes in development — no publication date announced.
Reference frameworks
- —NIST AI Risk Management Framework
- —ISO/IEC 42001 (AI management systems)
- —OECD AI Principles
Future research opportunities
Joint evaluation of governance frameworks · Academic review of proposed control requirements