CAIC CAIC AI Governance & Compliance 1 — Questions and Answers
Question 1: Which US federal framework provides guidelines for managing AI risks in organizations?
- NIST AI Risk Management Framework (AI RMF) (Correct answer)
- ISO 27001
- GDPR Article 22
- SOC 2 Type II
Correct answer: NIST AI Risk Management Framework (AI RMF)
The NIST AI RMF provides voluntary guidance to help organizations identify, assess, and manage AI risks across the AI system lifecycle.
Question 2: What is 'algorithmic accountability' in AI governance?
- Tracking compute costs of AI model training
- The obligation to explain, audit, and take responsibility for automated decision outcomes (Correct answer)
- Ensuring AI code is version-controlled
- Monitoring API rate limits on AI services
Correct answer: The obligation to explain, audit, and take responsibility for automated decision outcomes
Algorithmic accountability requires organizations to be able to explain how automated decisions are made and to take responsibility when those decisions cause harm.
Question 3: Under the EU AI Act, which category describes AI systems used in critical infrastructure that pose significant risk?
- Minimal risk
- Limited risk
- High risk (Correct answer)
- Unacceptable risk
Correct answer: High risk
The EU AI Act classifies AI used in critical infrastructure, employment, education, and essential services as 'high risk,' subject to strict conformity requirements.
Question 4: What is the purpose of an AI model card in responsible AI deployment?
- To store API authentication credentials securely
- To document a model's intended use, performance metrics, limitations, and ethical considerations (Correct answer)
- To define cloud infrastructure cost allocation
- To track user licensing agreements
Correct answer: To document a model's intended use, performance metrics, limitations, and ethical considerations
Model cards are structured documents that disclose key information about an AI model, including its training data, evaluation results, intended use cases, and known limitations.
Question 5: Which principle of responsible AI ensures that AI system decisions can be understood and traced by humans?
- Scalability
- Explainability (Correct answer)
- Redundancy
- Interoperability
Correct answer: Explainability
Explainability (also called interpretability) ensures that AI decisions and the reasoning behind them can be understood by developers, auditors, and affected individuals.
Question 6: A CAIC consultant discovers an AI hiring tool disproportionately rejects candidates from certain demographic groups. This is an example of:
- Model overfitting
- Algorithmic bias (Correct answer)
- Data pipeline latency
- Concept drift
Correct answer: Algorithmic bias
Algorithmic bias occurs when an AI system produces systematically unfair outcomes for specific groups, often due to biased training data or flawed model design.
CAIC CAIC AI Governance & Compliance 1
Which US federal framework provides guidelines for managing AI risks in organizations?