AI Governance — What It Is and Why It Matters
1. A recruitment system rejects a candidate. She asks why. The system responds, 'The result is based on complex patterns.' What is missing here?
Transparency means: Decisions must be explainable.
2. What do the cases Amazon, COMPAS, and the credit card provider have in common?
No ill will — just a lack of oversight and accountability.
3. Which principle ensures that there is always a human responsible for an AI decision?
Accountability = Responsibility. AI cannot be responsible — humans must be.
4. An AI model for credit lending was never explicitly trained on discrimination — yet it systematically shows poorer results for a group. Which principle is violated?
Fairness means: actively check for bias, not just hope that none is present.
5. What does 'minimum data for the purpose' mean?
Data minimization is a GDPR principle: collect only the minimum.
6. You receive an automatic rejection for a job application. What right do you have according to GDPR Art. 22?
GDPR Art. 22: Automated decisions with significant impact must be contestable.
7. You are an HR manager evaluating an AI tool for pre-screening applicants. What questions must you ask before purchasing? (Multiple possible) 2 pts
Governance issues: Explainability, data provenance, liability. Speed is not a governance criterion.
8. What is the core difference between an organization with and without AI governance?
It's not about whether something goes wrong — but whether you can recognize it and respond.
9. Which statement about AI governance is correct?
Governance enables responsible action — not less AI, but better AI.