AI Ethics and Machine Learning Compliance Issues

5 minutes 5 Questions

Test mode:
Start practice test

CIPP/E - AI Ethics and Machine Learning Compliance Issues Example Questions

Test your knowledge of AI Ethics and Machine Learning Compliance Issues

Question 1

An organization deploys a machine-learning model to screen job applicants. The model produces automated rejections with no human involvement, and applicants receive only a score. A rejected candidate asks why they were turned down. The privacy team must decide how to bring this processing into line with data protection law. Which measure best addresses the core compliance gap here?

Correct Answer: Provide meaningful information about the logic involved and arrange for human review of the decision

Article 22 GDPR prohibits solely automated decisions producing legal or similarly significant effects unless safeguards apply, and the data subject must be able to obtain human intervention, express their view, and contest the decision, plus receive meaningful information about the logic involved under Articles 13 to 15.

Publishing the training dataset is neither required nor practical and could breach others' rights. Legitimate interests plus audit logging does not satisfy the Article 22 safeguards or the applicant's rights. A transfer impact assessment addresses international transfers, which is not the issue here.

Key concept: safeguards for solely automated individual decision-making under Article 22.

Question 2

Which quality of a machine-learning model measures how consistently its predictions can be interpreted and traced back to specific input features by a human reviewer?

Correct Answer: Interpretability of the model's decision logic

Interpretability describes how well a human reviewer can understand, trace, and explain a model's predictions back to specific input features, which is central to transparency and accountability requirements under the GDPR.

Robustness concerns resilience to malicious or noisy inputs, not human understanding. Accuracy measures correctness of predictions, not explainability. Reproducibility refers to obtaining consistent results on repeated runs, which is about stability rather than traceability to inputs.

Key concept: model interpretability and explainability in the context of automated decision-making transparency.

Question 3

An organization procures a third-party machine-learning tool that ranks loan applicants. The vendor treats the model as confidential, so the deploying organization has little insight into how features drive outcomes. Data subjects begin questioning the fairness of decisions affecting them. What is the strongest reason this arrangement is problematic under European data protection law?

Correct Answer: It undermines the accountability principle, as the controller cannot demonstrate lawful and fair processing

The correct answer is right because Article 5(2) requires the controller to demonstrate compliance with lawfulness, fairness and transparency; if the vendor's confidential model prevents insight into how features drive outcomes, the deploying controller cannot evidence fair processing or meet transparency and Article 22 obligations regarding automated decisions.

Controllership is not automatically shifted to a vendor acting as processor, so obligations remain. Storage limitation concerns data retention, not model opacity. A lawful basis is always required and cannot be waived, and legitimate interests are not automatically valid for automated ranking.

Key concept tested is the accountability principle.

More AI Ethics and Machine Learning Compliance Issues questions
27 questions (total)
Start 27 question test

Other Compliance with European Data Protection Law and Regulation concepts

Adequacy Decisions (Article 45)AI Compliance Under GDPRBinding Corporate Rules (BCRs)Cookies and Tracking TechnologiesCooperation and Consistency MechanismsCorrective Measures and Enforcement PowersData Protection Authorities (DPAs)Direct Marketing and ePrivacyEU Data Act and AI Act ImplicationsEuropean Data Protection Board (EDPB) and EDPSEU-U.S. Data Privacy FrameworkFines and Penalties (Article 83)International Data Transfers Framework (Chapter V)One-Stop-Shop Mechanism (Article 56)Processing Employee Data Under GDPRSchrems I and Schrems II RulingsStandard Contractual Clauses (SCCs)Transfer Derogations (Article 49)
All Compliance with European Data Protection Law and Regulation questions · CIPP/E practice test