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
Q451 A team needs to know where all data elements feeding an AI process originate and flow. Which artifact is MOST useful?
- A. A procurement price list for storage and compute purchases
- B. A dataflow map aligned with the data dictionary or inventory
- C. A staff phone directory for the data governance office
- D. A holiday calendar maintained by the data engineering team
Answer: B
A dataflow map of all source data elements, aligned with the data inventory, supports security reviews and incident response evidence collection. The other documents do not describe data flows.
Domain 3: AI Technologies and Controls
Q452 Which access principle is MOST appropriate for datasets used by AI processes?
- A. Open access for all staff
- B. Least privilege
- C. Access granted by seniority only
- D. Shared administrator accounts
Answer: B
The principle of least privilege should be applied to AI datasets so people only have the access needed for their role. Open access, shared accounts and seniority-based access violate need-to-know.
Domain 3: AI Technologies and Controls
Q453 Temporary copies of data are created during an AI pipeline run. What should the security team MOST importantly ensure?
- A. The location of temporary data is known and the data is securely deleted after processing
- B. Temporary data is copied to personal drives for convenience
- C. Temporary data is retained indefinitely in case the pipeline is rerun, so that results can always be reproduced exactly
- D. Temporary data is emailed to the team for peer review
Answer: A
The architecture determines where temporary data sits, which must be known and securely deleted promptly after processing. Indefinite retention or personal copies expand exposure.
Domain 3: AI Technologies and Controls
Q454 A model will import data from an external source for the first time. Which control is MOST appropriate?
- A. Follow risk management processes each time data is imported from external sources
- B. Import the data first and assess the associated risk later, once the model owner has confirmed that the source performs well
- C. Disable monitoring for the import to speed up transfer
- D. Skip the review when the source is a well-known brand
Answer: A
Risk management processes should be followed every time data is imported from external sources. Importing first or skipping review for reputation ignores supply chain risk. Monitoring should stay on.
Domain 3: AI Technologies and Controls
Q455 An AI system creates new synthetic data on its own and stores it. How should the organization BEST treat this data?
- A. Classify it as public by default since no person supplied it
- B. Apply the same acquisition and governance controls as other data
- C. Exempt it from governance controls because it is synthetic
- D. Delete it without review once the model has finished training
Answer: B
Data generated by an automated AI system requires the acquisition processes and governance controls applied to other data. Synthetic origin does not remove obligations such as classification and retention.
Domain 3: AI Technologies and Controls
Q456 Which data governance step is MOST aligned with limiting what personal data is gathered for training?
- A. Data replication
- B. Data sharding
- C. Data hoarding
- D. Data minimization
Answer: D
Data minimization collects and retains only the data necessary for the purpose. Replication and sharding concern storage design, and hoarding is the opposite.
Domain 3: AI Technologies and Controls
Q457 When should data used for AI be destroyed according to sound data governance, at the LEAST?
- A. Never, because retained data always improves later model accuracy
- B. At the end of its defined retention period or when no longer needed
- C. Only when the vendor requests deletion in writing
- D. Only when the employees who collected it leave the company
Answer: B
Retention and destruction policies should set when data is deleted, based on purpose, legal requirements and necessity. Keeping data forever increases exposure.
Domain 3: AI Technologies and Controls
Q458 Which role is MOST accountable for ensuring data is cleansed, standardized and fit for AI use?
- A. The help desk
- B. The marketing team
- C. The facilities manager
- D. The data steward
Answer: D
Data stewards oversee quality, cleansing, standardization and compliance of datasets. Help desk, marketing and facilities are not responsible for data fitness.
Domain 3: AI Technologies and Controls
Q459 Which technique is MOST appropriate to protect sensitive training data at rest?
- A. Shared passwords
- B. Plain text storage
- C. Public cloud buckets without access control
- D. Encryption with managed keys
Answer: D
Encryption at rest with proper key management protects confidentiality. Plain text, shared passwords and open buckets weaken security.
Domain 3: AI Technologies and Controls
Q460 A team needs to use production data for model testing without exposing identities. Which technique is MOST suitable?
- A. Masking or anonymization
- B. Copying data to laptops
- C. Publishing the dataset
- D. Emailing samples
Answer: A
Masking or anonymization reduces exposure of identifiers while preserving utility. Copying, emailing or publishing real data increases risk.