AIGP Tests Proportionate AI Governance Decisions — Not Definitions of Machine Learning
For privacy, legal, risk and product people governing AI. Body of Knowledge v2.1 frames items as role-and-stage decisions: who is provider or deployer, which gate applies, and which control is proportionate.
Check Your Readiness →Most candidates understand IAPP Artificial Intelligence Governance Professional (AIGP) concepts — and still fail. This exam tests how you apply knowledge under pressure.
IAPP publishes item ranges per domain (16-20, 19-23, 21-25 and 21-25 of 85 scored items), so these percentages are midpoints. Expect short scenarios: identify the organisation's role and the life cycle stage, then choose the proportionate control rather than the most restrictive one.
Ban all generative AI use across the company
Bans push use into unmanaged personal accounts. Govern instead: approve enterprise tools with suitable data terms, issue an acceptable use policy, train each audience and record the tools in the AI inventory.
Allow it because the tool was already approved
This is function creep. A new purpose affecting different people needs a fresh impact assessment, and employment use may move the system into the EU AI Act's Annex III high-risk category.
Accept the vendor audit and deploy the tool
That audit reflects the vendor's test data, not this employer's candidates or workflow. Test locally, assign trained human overseers and keep accountability in house; the deployer remains exposed under employment and AI laws.
Release now and monitor the disparity after launch
A known high-severity issue in a high-impact use must be mitigated and retested, or escalated for an accountable decision, before release. Post-launch monitoring catches unknown problems; it does not excuse known ones.
Under the EU AI Act the vendor placing the system on the market is the provider. The bank is a deployer with oversight, logging and monitoring duties, and becomes a provider only if it rebrands or substantially modifies the system.
High risk follows Annex III purposes such as employment, credit, education or law enforcement. An ordinary customer chatbot is limited risk, with Article 50 transparency duties: people must be told they are interacting with AI.
The NIST AI RMF, with its Govern, Map, Measure and Manage functions, is voluntary guidance without any certification scheme. ISO/IEC 42001 is the certifiable AI management system, and even that certificate does not equal EU AI Act compliance.
Proxies such as postal code still produce disparate impact, and intent is not required. Fairness needs outcome testing by subgroup, because aggregate accuracy that meets its target can hide poor performance for a protected group.
Retrieval-augmented generation keeps content current, citable and removable without touching weights. Fine-tuning bakes in information that goes stale and is hard to unlearn, which also complicates deletion requests involving personal data.
Strengthen weak areas with exam-style practice questions and detailed explanations.
Simulate the real exam experience and assess your readiness under timed conditions.
Review key concepts, objectives, and exam topics in one place.
Get personalized explanations, learning recommendations, and instant answers.
Follow a structured learning path designed to help you prepare efficiently.
AIGP rewards proportionate, role-aware AI governance. Test whether you place each duty on the right party.