A structured assessment identifying affected stakeholders: Which assessment most fully meets the objective of
An AI impact assessment maps affected stakeholders, tests necessity and proportionality, and analyzes rights risks and safeguards, unlike narrower security, accuracy, or cost reviews.
The question
A public agency plans an AI system that scores welfare applicants for fraud risk, a use that touches vulnerable groups and legally significant decisions. Leadership wants a single assessment that best captures potential harms to affected individuals and communities before build sign-off. Which assessment most fully meets the objective of performing an impact assessment on the AI system?
Preparing for AIGP? Take the free 5-min readiness quiz →
- A penetration test and threat-model exercise that probes the system's attack surface, enumerates the exploit paths available, and documents the security controls needed to protect the scoring pipeline end to end.Plausible because security testing is a valid governance activity, but it evaluates confidentiality and integrity, not the societal and rights harms an impact assessment is meant to surface.
- A model accuracy and performance benchmark that measures precision, recall and calibration against a holdout set to confirm the fraud-scoring model meets its technical acceptance thresholds.Plausible because performance evaluation is essential, but metric benchmarking measures technical quality, not the broader impacts on rights and vulnerable populations an impact assessment covers.
- A financial cost-benefit analysis quantifying projected fraud savings against build and operating costs to establish that the scoring program delivers a positive return.Plausible because business justification matters for the use case, but an ROI study addresses economic value, not the harms to individuals and communities central to an impact assessment.
- A structured assessment identifying affected stakeholders, evaluating necessity and proportionality, and analyzing risks to individuals' rights alongside mitigating safeguards before deployment. ✓Correct because an AI impact assessment systematically maps affected stakeholders, tests necessity and proportionality, and analyzes rights impacts and mitigations, which is exactly what this rights-sensitive use requires.
The trap
Assuming a security test, accuracy benchmark, or ROI study can stand in for an impact assessment focused on rights and societal harms. How to remember it
An AI impact assessment maps affected stakeholders, tests necessity and proportionality, and analyzes rights risks and safeguards, unlike narrower security, accuracy, or cost reviews.
How many of these would you get right?
One of 1581 AIGP questions on Certsqill. Take a free five-minute check and see your score per domain — not one number, but which section to open tonight.
Test your AIGP readiness — freeMore Understanding How to Govern AI Development questions
- Specify the intended purpose: Which action most directly fulfills that objective? →
- Enabling a qualified person to review: When applying best practices to designing and building an AI system, →
- To prioritize identified risks by combining: What is the primary purpose of this probability/severity harms →
- All 426 Understanding How to Govern AI Development questions →
Part of the Certsqill AIGP question bank · Understanding How to Govern AI Development ·
Every answer, right and wrong, comes with its own explanation.
