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
Q391 An enterprise fine-tunes a general model and sells the result as a product to other firms. Which statement is MOST accurate about its position in the AI supply chain?
- A. It has no role because the base model came from another firm
- B. It becomes the regulator for the customers who buy the product
- C. It may act as a provider for its customers and carry provider responsibilities
- D. It is only a deployer and carries no compliance duty for the product, since it did not train the base model
Answer: C
How an enterprise uses AI determines its supply chain role; supplying a solution to others makes it a provider with compliance and documentation duties. Being a deployer of the base model does not remove that. It is never a regulator.
Domain 2: AI Risk Management
Q392 During vendor vetting, which factor is the MOST common reason for NOT selecting an AI vendor?
- A. Poor security practices or financial instability
- B. A high number of social media followers and press mentions
- C. A newly renovated headquarters location in a central business district
- D. A large number of satisfied attendees at the vendor's own conferences
Answer: A
Financial instability, regulatory noncompliance, poor security practices, reputational risk and weak operational resilience are standard reasons to reject a vendor, amplified for AI. Office location and followers do not measure risk.
Domain 2: AI Risk Management
Q393 Which concern is MOST specific to evaluating a third-party generative AI solution?
- A. The vendor's holiday calendar
- B. Lack of content provenance mechanisms
- C. Elevator capacity at the vendor's building
- D. The vendor's choice of office furniture
Answer: B
Lack of content provenance mechanisms for generative AI appears in the manual's list of primary concerns, along with transparency, ethics, privacy and lock-in. The other items have no bearing on AI risk.
Domain 2: AI Risk Management
Q394 A vendor refuses to allow due diligence, audits or security assessments of its AI service. How should the organization MOST appropriately view this?
- A. As an issue only for the vendor's lawyers to resolve in contract talks
- B. As a positive sign of vendor confidence in its own security posture
- C. As a significant concern that may place the vendor outside acceptable risk
- D. As irrelevant to the decision if the vendor's price is low enough
Answer: C
Resistance to due diligence, auditing and security assessments is listed as a primary concern and may make the vendor exceed risk appetite. Price does not offset inability to assess risk. It is an organizational risk issue, not only a legal one.
Domain 2: AI Risk Management
Q395 An organization chooses to use a vendor it considers high risk. Under what condition is this MOST justifiable?
- A. The vendor's staff were cooperative during the sales presentations
- B. The vendor offers the lowest price among the shortlisted providers
- C. The associated risk can be adequately mitigated, transferred or accepted
- D. The vendor is well known and frequently covered in the technology press
Answer: C
A high-risk vendor may still be selected if the risk can be mitigated, transferred or accepted within appetite. Price, fame or friendliness do not address the risk.
Domain 2: AI Risk Management
Q396 A vendor's AI technology is so immature that its potential harm cannot be controlled by contract, technical safeguards or procedures. What is the BEST decision?
- A. Allow the vendor to run in production but without any monitoring
- B. Do not engage the vendor because the risk exceeds appetite and benefits do not outweigh it
- C. Transfer all accountability to the vendor through a contract clause
- D. Sign the contract and rely on the vendor's written assurances of safety, since the vendor knows the technology best
Answer: B
When a vendor's characteristics create impacts the organization cannot understand, control or mitigate, the benefits do not outweigh the risk and the vendor should not be used. Assurances alone do not change this, and accountability cannot be fully transferred.
Domain 2: AI Risk Management
Q397 Which business consideration for vendor selection BEST addresses the vendor's long-term fit with digital transformation plans?
- A. The number of regional offices the vendor maintains worldwide
- B. The vendor's pricing for printed documentation and training books
- C. Alignment with business strategy rather than a one-off experiment
- D. The vendor's published brand color palette and marketing campaigns
Answer: C
Vendor selection includes confirming the provider understands the strategy and that its AI aligns with broader digital transformation initiatives. Brand colors, printing prices and office count do not show strategic fit.
Domain 2: AI Risk Management
Q398 Which vendor selection area BEST addresses who owns the data and outputs and how they are handled?
- A. Data handling and legal and ethical implications
- B. Seating arrangements and facilities provided at the vendor's offices
- C. The marketing slogans used on the vendor's product web pages
- D. Payroll processing timelines used by the vendor's finance team
Answer: A
Data handling and legal and ethical considerations cover ownership, privacy, security and use of data and outputs. Seating, slogans and payroll are not part of AI vendor selection considerations.
Domain 2: AI Risk Management
Q399 A company deploys a vendor model and assumes the vendor is solely accountable for outcomes. What is the MAIN weakness in this view?
- A. Deployers keep responsibilities such as oversight, data control and monitoring
- B. Providers handle human oversight on behalf of every deployer
- C. The vendor owns all data supplied to it, so deployers carry no duties once the contract has been signed
- D. Deployers have no duties once a contract is signed with a provider
Answer: A
Deployers are responsible for using the system as directed, controlling data inputs, human oversight and monitoring regardless of provider duties. Responsibility is shared, not wholly transferred.
Domain 2: AI Risk Management
Q400 Under an AI shared responsibility model, how should responsibilities be MOST appropriately defined?
- A. By the size of the model that the vendor chooses to supply
- B. By assuming the vendor covers every responsibility by default, since the vendor operates the platform and holds the model
- C. Informally through email exchanges between account managers
- D. Contractually, with clear expectations on SLAs, incident response and notification
Answer: D
A shared approach must be addressed contractually, with defined expectations for SLAs, incident response activities and notification requirements. Informal or assumed arrangements leave gaps.