Apply / your decision record

Make the call.
Explain your decision.

Identify the facts that matter, any missing evidence and the next action. Name the responsible role and explain what would change your decision. Do not assume that an unmentioned approval, exception or assessment exists. Allow around 8–12 minutes.

One useful record

YOUR CASE / RISK MANAGEMENT AND DATA GOVERNANCE

A provider is developing a high-risk recruitment-ranking system. Its dataset reflects past hiring decisions, has few examples from one applicant group and contains duplicate CVs across training and testing. Aggregate accuracy looks strong, but a pilot shows unequal false-rejection rates.

Your output / a practical record

AI risk register and dataset review

Make the next decision explicit. Connect it to evidence, unresolved questions and an accountable role.

Draft: AI risk register and dataset review

1. Intended purpose, affected people and foreseeable misuse2. Hazard, harm pathway and initial risk rationale3. Dataset source, rights, splits and representativeness4. Mitigation, owner and measurable acceptance criterion5. Test evidence, subgroup results and residual risk6. Release authority, monitoring trigger and review date

Check your reasoning

I linked duplicate CVs, under-represented groups and unequal false rejections to specific harm pathways.I separated aggregate accuracy from evidence about subgroup performance and dataset suitability.I assigned mitigations, acceptance criteria and a reasoned release or hold decision.I recorded dataset provenance, test evidence, residual risk and monitoring triggers.

This is a self-review. Compare with the worked example if you need help. Every field needs a response of at least 10 characters; text length does not assess its quality.