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 2: AI Risk Management
Q331 A company plans to deploy an AI system classed as high risk under the EU AI Act. Which assessment is MOST specifically required before a public body first uses it?
- A. A hardware refresh plan
- B. A marketing impact review
- C. A cost allocation study
- D. A fundamental rights impact assessment
Answer: D
The Act requires a FRIA before deploying high-risk systems, especially for public bodies or private entities providing public services, to ensure individual rights are not harmed. The other items are not required assessments.
Domain 2: AI Risk Management
Q332 Which element is MOST characteristic of a fundamental rights impact assessment?
- A. A hardware inventory of the servers used for model training
- B. A list of marketing slogans approved for the AI product launch
- C. Identification of affected groups and oversight measures such as HITL
- D. Projected vendor profit forecasts for the AI system over five years
Answer: C
A FRIA covers usage description, operational details, affected groups, risk identification, oversight measures and mitigation plans. Marketing or profit information is irrelevant. Hardware is not the focus.
Domain 2: AI Risk Management
Q333 A privacy team assumes a GDPR data privacy impact assessment satisfies the FRIA requirement. What is the MAIN problem with this assumption?
- A. FRIAs are optional for high-risk systems when a DPIA exists
- B. A DPIA covers a wider range of rights than a FRIA does
- C. A DPIA is always forbidden for AI systems that process personal data
- D. A DPIA does not fully cover the scope of rights addressed by a FRIA
Answer: D
A data privacy impact assessment under GDPR does not fully cover the scope of the rights mandated by a FRIA, so both may be needed. FRIAs are required for high-risk systems. A DPIA is narrower, not broader.
Domain 2: AI Risk Management
Q334 When is it MOST necessary to update a fundamental rights impact assessment?
- A. Every month regardless of change
- B. Never after the first use
- C. Only after a regulator audit
- D. When any of the conditions assessed change
Answer: D
The assessment must be completed before first use and updated if conditions change, with findings submitted to the authority. A fixed monthly cycle or audit-only update is not what is required.
Domain 2: AI Risk Management
Q335 Which evaluation BEST demonstrates that a high-risk AI provider's quality management system and technical documentation meet regulatory standards?
- A. A financial audit
- B. A conformity assessment
- C. A penetration test
- D. A customer satisfaction survey
Answer: B
Conformity assessments evaluate the provider's quality management system and technical documentation. They may be performed internally or by an independent expert. Penetration tests, financial audits and surveys do not demonstrate regulatory conformity.
Domain 2: AI Risk Management
Q336 When is a new conformity assessment MOST likely required?
- A. After substantial modifications to a high-risk system
- B. Only if the vendor changes its logo, brand name or corporate ownership during the contract period
- C. At the end of each business day
- D. Whenever a user logs in
Answer: A
New assessments are required after any substantial modification to high-risk AI systems. Logins, daily cycles and branding have no bearing on conformity.
Domain 2: AI Risk Management
Q337 Who may PRIMARILY perform a conformity assessment, depending on system type and standards used?
- A. Only a regulator's technology or marketing team that has been appointed
- B. Only an unrelated vendor
- C. The provider's organization or an independent external expert
- D. Only the end user
Answer: C
The assessment may be performed by the provider's organization or through external review by an independent expert. End users and unrelated vendors are not designated. Regulators evaluate, not market.
Domain 2: AI Risk Management
Q338 A company introduces a new AI feature that uses personal data in a way not previously done. What should it do FIRST regarding privacy?
- A. Conduct a new privacy impact assessment
- B. Wait for a complaint
- C. Remove all logging
- D. Rely on the old assessment unchanged
Answer: A
New PIAs should be performed when a change likely involves a new use of personal data or significantly changes handling, such as adopting AI. Waiting for complaints or reusing old assessments leaves new risk unexamined.
Domain 2: AI Risk Management
Q339 Which risk is MOST specific to AI in privacy impact assessments, compared with traditional privacy risk?
- A. Unauthorized physical access to servers hosting the AI platform, for example by contractors visiting the data center
- B. Weak door locks on the office where staff process personal data
- C. Loss of backup tapes containing archived customer records
- D. Inference of sensitive personal details beyond those explicitly provided
Answer: D
AI-driven privacy risk includes algorithmic bias, inference of sensitive details and lack of explainability. Physical access and backup loss are traditional risks. They remain relevant but are not AI-specific.
Domain 2: AI Risk Management
Q340 Why should a PIA for an AI system PRIMARILY use an established government or industry methodology?
- A. It reduces the number of stakeholders who must be consulted to zero, since the methodology supplies all of the assumptions
- B. It makes the PIA repeatable and consistent from one assessment to the next
- C. It guarantees zero risk
- D. It removes the need for any assessment
Answer: B
A recognized methodology ensures the PIA process is repeatable and can be performed consistently. It does not eliminate risk or remove the need for assessment.