Leadership for AI qualifications (Level 3, 5 & 7)
What the AI leadership exams test
CMI Leadership for AI qualifications at Levels 3, 5 and 7 test AI strategy, governance, data ethics, risk, workforce impact and human oversight. Level 3 focuses on responsible use; Level 5 on managing AI projects; Level 7 on board-level governance and value.
Core rules
- Governance principles: accountability, transparency, fairness, safety, privacy and human oversight.
- Data protection: lawful basis, minimisation, retention and a DPIA for high-risk processing.
- Model risk: data quality, bias testing, validation, monitoring and drift.
- Keep a human in the loop for high-impact decisions.
- Know risk tiers such as the EU AI Act categories and relevant UK sector regulators.
Worked example
A CV-screening AI needs a DPIA, bias audit across protected groups, explainability, human review of rejections, an appeal route, supplier due diligence and monitoring. Accuracy alone is insufficient. A vendor claim of bias-free operation is not evidence.
Traps that cost marks
- Treating AI as an IT-only project.
- Ignoring data quality and provenance.
- Assuming explainability is automatic.
- Deploying without monitoring or a kill switch.