Isaca AAISM Online Practice Questions and Exam Preparation

Isaca AAISM Online Questions & Answers

• Question 51:

  Which of the following is the MOST effective strategy for penetration testers assessing the security of an AI model against membership inference attacks?

  A. Disabling AI model logging to reduce noise during testing
  B. Measuring AI model accuracy on the test set
  C. Analyzing AI model confidence scores to indicate training data
  D. Generating synthetic data to replace the training data
  C. Analyzing AI model confidence scores to indicate training data

  ---

  AAISM identifies confidence-score analysis as a principal technique for evaluating exposure to membership inference: models often yield measurably higher confidence for points seen during training. Testers compare output probabilities/entropies for known in-training vs. out-of-training samples to assess leakage. Disabling logs (A) reduces evidence; test-set accuracy (B) does not measure privacy leakage; synthetic data generation (D) is a mitigation strategy, not a penetration-testing method.

  References: AI Security Management (AAISM) Body of Knowledge - Model Privacy Threats:

  Membership Inference; Red/Blue Team Evaluation Techniques; Confidence/Entropy-based Privacy Testing.

• Question 52:

  Which of the following controls BEST mitigates the risk of bias in AI models?

  A. Robust access control techniques
  B. Regular data reconciliation
  C. Cryptographic hash functions
  D. Diverse data sourcing strategies
  D. Diverse data sourcing strategies

  ---

  Bias in AI models primarily stems from limitations or imbalances in training data. The AAISM study materials emphasize that the most effective way to mitigate this risk is through diverse data sourcing strategies that ensure coverage across demographics, scenarios, and contexts. Access controls protect data security, not fairness. Data reconciliation ensures accuracy but does not address representational imbalance. Cryptographic hashing preserves integrity but has no impact on bias mitigation. To reduce

  systemic unfairness, the critical control is sourcing diverse and representative data.

  References:

  AAISM Exam Content Outline - AI Technologies and Controls (Bias and Fairness Management)

  AI Security Management Study Guide - Data Governance and Bias Reduction Strategies

• Question 53:

  A school district contracts a third-party provider for AI-based curriculum recommendations. Which of the following is the BEST way to ensure the vendor uses AI responsibly?

  A. Confirming the AI solution supports single sign-on (SSO)
  B. Verifying the vendor has updated terms of service
  C. Requiring the vendor to provide the model card
  D. Ensuring the vendor offers 24/7 technical support
  C. Requiring the vendor to provide the model card

  ---

  AAISM emphasizes transparency artifacts from vendors to enable due diligence and assurance. A model card documents intended use, data sources, limitations, performance across subgroups, known risks, and evaluation procedures-information necessary to assess safety, fairness, and compliance for sensitive contexts like education. SSO and support are useful operational features; generic ToS updates are insufficient without model-specific disclosures.

  References: AI Security Management (AAISM) Body of Knowledge - Third-Party and Supply Chain Governance; Transparency Artifacts (Model Cards, Datasheets). AAISM Study Guide - Vendor Due Diligence Requirements; Documentation for Risk, Fairness, and Intended Use.

• Question 54:

  To ensure the ethical and responsible use of AI, which of the following AI usage policy metrics is MOST important for an organization to monitor?

  A. Frequency of policy consultations by employees
  B. Number of reported policy violations
  C. Number of AI projects that have undergone policy compliance review
  D. Frequency of policy reviews and updates
  C. Number of AI projects that have undergone policy compliance review

  ---

  AAISM emphasizes governance effectiveness metrics tied to real lifecycle checkpoints. The count (and percentage) of AI projects that completed policy compliance review before deployment is a leading indicator of policy enforcement and assurance. It directly reflects whether responsible-AI requirements (risk assessment, impact assessment, data/privacy checks, security controls) are embedded in practice. Consult frequency (A) and review cadence (D) are activity metrics, not outcomes. Reported violations (B) are lagging indicators and can be deceptive (low numbers may indicate under-reporting). References:. AI Security Management (AAISM) Body of Knowledge: Program KPIs-policy adoption, stage-gate compliance, audit readiness. AAISM Study Guide: Governance metrics for Responsible AI-coverage of reviews, pass/fail rates, exceptions handling

• Question 55:

  A financial institution wants to adopt an LLM-based assistant for loan-eligibility inquiries. What is the FIRST action to ensure regulatory compliance?

  A. Establish human-in-the-loop review for all decisions
  B. Validate fairness metrics against demographic subgroups
  C. Conduct an AI impact assessment D. Test the model for bias and drift
  C. Conduct an AI impact assessment D. Test the model for bias and drift

  ---

  According to AAISM, high-impact use cases involving financial decisions require an impact assessment FIRST to understand legal, societal, fairness, and rights-based implications. Fairness testing (B) and bias/drift evaluations (D) occur later. HITL review (A) is a control but not the initial step.

  References:

  AAISM Study Guide ?AI Governanceand; Impact Assessment Requirements.

• Question 56:

  A financial organization relies on AI-based identity verification and fraud detection services. Which of the following BEST integrates AI security risk into the business continuity plan (BCP)?

  A. Using explainable AI to document decision paths
  B. Periodic retraining using pre-labeled data
  C. Including AI model supporting infrastructure in disaster recovery scenarios
  D. Duplicating AI microservices across multiple availability zones
  C. Including AI model supporting infrastructure in disaster recovery scenarios

  ---

  AAISM stresses that AI systems and their supporting infrastructure must be explicitly included in disaster recovery and continuity planning, since disruptions to models, feature stores, or pipelines can halt critical business functions. Explainability (A) and retraining (B) are operational improvements, not continuity mechanisms. Multi-zone redundancy (D) improves availability but does not represent complete BCP integration.

  References: AAISM Study Guide - AI Business Continuity and Resilience Requirements.

• Question 57:

  Which of the following is BEST for analyzing true positives, true negatives, false positives, and false negatives produced by an AI model?

  A. Hyperparameter tuning
  B. Precision
  C. Confusion matrix
  D. Recall
  C. Confusion matrix

  ---

  A confusion matrix is explicitly defined in AAISM as the framework used to interpret classification performance by listing: true positives true negatives false positives false negatives

  Precision (B) and recall (D) are derived metrics that use parts of the matrix but do not show the full picture. Hyperparameter tuning (A) is unrelated.

  References: AAISM Study Guide - AI Model Evaluation Metrics; Confusion Matrix.

• Question 58:

  Which of the following is the MOST effective way to prevent a model inversion attack?

  A. Monitor model output for anomalies
  B. Utilize data pseudonymization
  C. Implement differential privacy during model training
  D. Ensure data minimization
  C. Implement differential privacy during model training

  ---

  AAISM identifies differential privacy as the primary mitigation technique against model inversion attacks, which attempt to reconstruct sensitive training data by probing model outputs. Pseudonymization (B) and minimization (D) reduce exposure but do not prevent inversion. Output monitoring

  (A) detects anomalies but doesn't block reconstruction.

  References: AAISM Study Guide - Privacy Attacks and Defenses; Differential Privacy.

• Question 59:

  A large pharmaceutical company using a new AI solution to develop treatment regimens is concerned about potential hallucinations with the introduction of real-world data. Which of the following is MOST likely to reduce this risk?

  A. Penetration testing
  B. Human-in-the-loop
  C. AI impact analysis
  D. Data asset validation
  B. Human-in-the-loop

  ---

  AAISM materials identify human-in-the-loop governance as the most effective safeguard against risks such as hallucinations in AI systems used in high-stakes domains like healthcare. By ensuring that human experts validate outputs before they influence patient treatment decisions, organizations preserve accountability, safety, and accuracy. Penetration testing is a cybersecurity measure, not relevant to hallucination risk. AI impact analysis helps evaluate systemic effects but does not directly prevent faulty

  outputs. Data validation improves input quality but cannot fully prevent generative hallucinations. The key safeguard is human-in-the-loop oversight.

  References:

  AAISM Study Guide - AI Governance and Program Management (Human Oversight in High-Risk AI)

  ISACA AI Security Management - Mitigating Hallucinations in Generative AI

• Question 60:

  A military contractor discovered that its large language model (LLM) is at high risk of being targeted by advanced persistent threat (APT) actors seeking to exploit the model to access confidential information. Which of the following attacks is the HIGHEST priority to protect against?

  A. Model inversion
  B. Data poisoning
  C. Unauthorized tuning
  D. Model distillation
  A. Model inversion

  ---

  AAISM classifies model inversion as a privacy/information-leakage threat where adversaries infer or reconstruct sensitive training data or attributes from model outputs-directly jeopardizing confidential information targeted by APTs. While data poisoning, unauthorized tuning, and model distillation present material risks (integrity, governance/IP theft), the scenario's stated objective-accessing confidential information-most directly maps to inversion. Accordingly, AAISM prioritizes defenses such as output regularization, confidence suppression/calibration, overfitting controls, privacy-preserving techniques, and strict access/telemetry on inference interfaces. References:. AI Security Management. (AAISM) Body of Knowledge: Model Security-Inference-Time Threats (Inversion, Membership Inference) and Confidentiality Risks. AAISM Study Guide: Leakage Mitigations-Regularization, Output Minimization/Calibration, Access Controls and Monitoring on Model Interfaces