This is not a replacement for the official study guide. Please read the official study guide first — use this exam prep only to reinforce or practice your understanding.
Free AAISM exam practice questions with answers and explanations, organized by domain and part.
Domain 3: AI Technologies and Controls
Q471 A hiring tool consistently ranks candidates from one demographic lower. Which ethical principle is MOST directly at risk?
- A. Cost control over the recruitment advertising budget
- B. Availability of the hiring portal during peak hours
- C. Efficiency of the interview scheduling workflow
- D. Fairness and non-discrimination in automated decisions
Answer: D
Biased ranking harms specific groups, so fairness is the principle at risk. Availability, scheduling efficiency and budget are operational concerns that do not describe discriminatory outcomes.
Domain 3: AI Technologies and Controls
Q472 An organization wants employees to challenge decisions made by an AI system that affects them. Which ethical principle does this BEST support?
- A. Cost control over the vendor licence for the AI platform
- B. Availability of the AI service during peak usage periods
- C. Efficiency of the process used for scheduling reviews
- D. Accountability and the ability to contest outcomes
Answer: D
Letting people challenge AI outcomes supports accountability and human rights. Throughput, storage cost and release cadence are operational measures unrelated to contestability.
Domain 3: AI Technologies and Controls
Q473 Which practice is MOST effective for embedding ethics in AI development?
- A. Performing ethical impact assessments early and at each significant change
- B. Reviewing ethics only after public complaints have been received and the communications team has prepared a response
- C. Delegating ethical review entirely to the model vendor
- D. Treating ethics statements as a marketing exercise for the brand
Answer: A
Ethical impact assessments early in development and on changes surface harm before release. Reacting only to complaints, delegating entirely, or using ethics for marketing weakens governance.
Domain 3: AI Technologies and Controls
Q474 Which attribute is MOST associated with trustworthy AI according to common risk frameworks?
- A. Popular among competitors and widely promoted in the press, with a large community of enthusiastic early adopters
- B. Fast to deploy, cheap to run and easy for teams to adopt
- C. Largest in model size and broadest in available features
- D. Valid and reliable, safe, secure, accountable, explainable, privacy-enhanced and fair
Answer: D
Trustworthiness attributes in frameworks such as NIST include validity and reliability, safety, security and resilience, accountability and transparency, explainability, privacy and fairness. Speed, cost, popularity and size do not establish trust.
Domain 3: AI Technologies and Controls
Q475 A model's decisions are accurate but nobody can describe how they were reached. Which trust attribute is MOST lacking?
- A. Availability of the service during business hours
- B. Portability of the model between hosting environments
- C. Explainability and interpretability
- D. Scalability to handle peak workloads without errors
Answer: C
Explainability lets people understand how outputs are produced, which is needed for trust, challenge and audit. The system may be available, scalable and portable yet still opaque.
Domain 3: AI Technologies and Controls
Q476 What is the MOST likely effect of poor transparency about AI use on customers?
- A. Higher customer satisfaction because users prefer simpler messaging
- B. Erosion of trust and increased complaints or regulatory scrutiny
- C. Immediate regulatory approval because no claims are made
- D. Lower litigation risk because fewer details are disclosed
Answer: B
Lack of transparency erodes trust and can trigger complaints and scrutiny. It doesn't raise satisfaction or lower legal risk, and regulators do not approve opaque use.
Domain 3: AI Technologies and Controls
Q477 An AI-enabled device might injure users if its model behaves unexpectedly. Which governance objective is MOST relevant?
- A. Safety, including safeguards and human oversight
- B. Brand color guidelines applied to the device packaging
- C. Marketing reach across regions where the device is sold
- D. Subscription pricing structure for the connected device
Answer: A
Safety requires safeguards, testing and oversight to prevent physical or psychological harm. Marketing, pricing and branding do not address harm prevention.
Domain 3: AI Technologies and Controls
Q478 A human reviews and approves each AI recommendation before action is taken. Which oversight model does this BEST describe?
- A. Fully autonomous operation
- B. Human-out-of-the-loop
- C. Human-on-the-loop
- D. Human-in-the-loop
Answer: D
In human-in-the-loop, a person approves or intervenes in each decision. Human-on-the-loop supervises and can intervene, while autonomous or out-of-the-loop operation proceeds without per-decision human involvement.
Domain 3: AI Technologies and Controls
Q479 Which factor MOST favors requiring human-in-the-loop for an AI process?
- A. High-impact decisions that affect rights, safety or finances
- B. Fully tested internal calculations with no external effect
- C. Low-impact, reversible tasks
- D. Cosmetic formatting changes
Answer: A
HITL is most appropriate where decisions can significantly affect individuals' rights, safety or finances. Low-impact or reversible tasks can use lighter oversight.
Domain 3: AI Technologies and Controls
Q480 A reviewer routinely accepts AI suggestions without scrutiny. Which risk is MOST evident?
- A. Hardware aging in the workstation used by the reviewer
- B. Automation bias weakening human oversight
- C. Excessive transparency about how the recommendation was produced
- D. Network congestion that delays the delivery of recommendations
Answer: B
Automation bias is overreliance on AI outputs, which defeats the purpose of oversight. Congestion, hardware aging and transparency are not at issue.