Pass IAPP AIGP Exam in First Attempt Guaranteed!
Get 100% Latest Exam Questions, Accurate & Verified Answers!
30 Days Free Updates, Instant Download!
IAPP AIGP Practice Test 70
Question 690
A hospital buys a medical triage model and assumes the vendor will also train clinicians and manage local patient complaints. Which additional observation would most strongly confirm the particular risk or control gap?
Explanation
Correct Answer
The supplier contract specifies model updates but no party owns onsite training or clinical overrides. It supports the targeted response: Define vendor/provider duties separately from hospital deployment, monitoring, staff training and local escalation duties.
Incorrect Answers
Procurement finds paid model subscriptions absent from security and risk registers. It supports the targeted response: Define vendor/provider duties separately from hospital deployment, monitoring, staff training and local escalation duties. The supplier contract specifies model updates but no party owns onsite training or clinical overrides.
Incident reports show staff misunderstand which shipment fields may enter third-party AI services. It supports the targeted response: Define vendor/provider duties separately from hospital deployment, monitoring, staff training and local escalation duties. The supplier contract specifies model updates but no party owns onsite training or clinical overrides.
Local teams cannot document country-specific impact requirements in the headquarters-only template. It supports the targeted response: Define vendor/provider duties separately from hospital deployment, monitoring, staff training and local escalation duties. The supplier contract specifies model updates but no party owns onsite training or clinical overrides.
Question 691
A consumer app vendor claims its new AI moderation is safe based on a hand-picked demonstration of 20 prompts. Which governance intervention most directly addresses this specific situation?
Explanation
Correct Answer
Design representative benchmarks and controlled pilot tests before approving rollout. The action addresses benchmark contamination and pilots rather than treating every AI control as interchangeable. The decisive signal to verify is that the shown prompts avoid ambiguous, multilingual and adversarial inputs that real customers send. This matters because evidence from curated demos does not establish robustness across deployment conditions. Success should be measured by confirming that the pilot records performance under realistic difficult cases and meets pre-agreed guardrail criteria.
Incorrect Answers
Conduct an impact assessment including affected individuals, error severity and available recourse. Evidence from curated demos does not establish robustness across deployment conditions. Design representative benchmarks and controlled pilot tests before approving rollout.
Threat-model malicious training input and design validation and isolation controls before ingestion. Evidence from curated demos does not establish robustness across deployment conditions. Design representative benchmarks and controlled pilot tests before approving rollout.
Define uncertainty thresholds and human escalation for unsupported or high-stakes requests. Evidence from curated demos does not establish robustness across deployment conditions. Design representative benchmarks and controlled pilot tests before approving rollout.
Question 692
A fitness insurer wants to feed every historical health record into a reminder model that only needs a renewal date. Which governance reasoning best supports the appropriate response?
Explanation
Correct Answer
Data minimization links each personal-data element to a legitimate, necessary purpose. This reasoning fits minimization of health data in ai because the observed gap is that the feature inventory includes diagnoses with no demonstrated predictive value for the reminder task. It points to the practical intervention: Restrict training features to relevant fields, justify any sensitive data use and reduce identifiability. The outcome should then show that the input dataset and model pipeline exclude unnecessary medical fields and document the rationale.
Incorrect Answers
A supplier's privacy notice does not establish the customer controller's own legal grounds. This reasoning fits minimization of health data in ai because the observed gap is that the feature inventory includes diagnoses with no demonstrated predictive value for the reminder task. Data minimization links each personal-data element to a legitimate, necessary purpose.
High-impact automation can trigger special legal duties beyond generic website disclosure. This reasoning fits minimization of health data in ai because the observed gap is that the feature inventory includes diagnoses with no demonstrated predictive value for the reminder task. Data minimization links each personal-data element to a legitimate, necessary purpose.
An AI output can be a security and privacy incident that triggers data-breach duties. This reasoning fits minimization of health data in ai because the observed gap is that the feature inventory includes diagnoses with no demonstrated predictive value for the reminder task. Data minimization links each personal-data element to a legitimate, necessary purpose.
Question 693
A tax adviser needs a chatbot to cite frequently updated public guidance without embedding every update in model weights. Which governance intervention most directly addresses this specific situation?
Explanation
Correct Answer
Evaluate approved retrieval-augmented generation with source freshness and permission controls. The action addresses retrieval augmented generation fit rather than treating every AI control as interchangeable. The decisive signal to verify is that the model's static knowledge misses recent guidance and invented citations appear in tests. This matters because retrieval can ground changing factual content without treating the model as the authority. Success should be measured by confirming that responses link to current approved documents, with unsupported queries escalated.
Incorrect Answers
Assess staff readiness, operating procedures and override skills before launch. Retrieval can ground changing factual content without treating the model as the authority. Evaluate approved retrieval-augmented generation with source freshness and permission controls.
Assess licenses, support, security, update controls, data use terms and governance before choosing. Retrieval can ground changing factual content without treating the model as the authority. Evaluate approved retrieval-augmented generation with source freshness and permission controls.
Compare an existing governed service with in-house development on requirements, risk, support and total cost. Retrieval can ground changing factual content without treating the model as the authority. Evaluate approved retrieval-augmented generation with source freshness and permission controls.
Question 694
A chatbot team closes a harmful-output ticket once it edits the prompt, without testing whether the problem is resolved. Which governance reasoning best supports the appropriate response?
Explanation
Correct Answer
Issue management needs goal evidence of mitigation, not simply completion of a work item. This reasoning fits issue tracking and closure evidence because the observed gap is that the ticket lists a configuration change but contains no reproduction or retest result.
Incorrect Answers
A post-release AI incident may demand broader governance response than ordinary bug triage. This reasoning fits issue tracking and closure evidence because the observed gap is that the ticket lists a configuration change but contains no reproduction or retest result. Issue management needs goal evidence of mitigation, not simply completion of a work item.
Safe AI release management includes the ability to reverse harmful changes. This reasoning fits issue tracking and closure evidence because the observed gap is that the ticket lists a configuration change but contains no reproduction or retest result. Issue management needs goal evidence of mitigation, not simply completion of a work item.
Model and tool changes can create security risks even if original penetration tests passed. This reasoning fits issue tracking and closure evidence because the observed gap is that the ticket lists a configuration change but contains no reproduction or retest result. Issue management needs goal evidence of mitigation, not simply completion of a work item.
Question 695
A robotics maker deploys obstacle-avoidance AI and assumes standard software terms waive every product safety responsibility. Which governance intervention most directly addresses this specific situation?
Explanation
Correct Answer
Evaluate applicable product-safety and liability requirements alongside rigorous fault testing. The action addresses defective ai-enabled physical product rather than treating every AI control as interchangeable. The decisive signal to verify is that a robot makes unsafe movements after a foreseeable sensor failure, with no documented hazard analysis. This matters because aI embedded in a product can create liability for defects despite software disclaimers. Success should be measured by confirming that safety testing and corrective controls address foreseeable failure modes before distribution.
Incorrect Answers
Review insurance fairness rules and model factors with legal and actuarial oversight. AI embedded in a product can create liability for defects despite software disclaimers. Evaluate applicable product-safety and liability requirements alongside rigorous fault testing.
Review source-license restrictions and output rights before commercial distribution. AI embedded in a product can create liability for defects despite software disclaimers. Evaluate applicable product-safety and liability requirements alongside rigorous fault testing.
Substantiate numerical performance claims on representative data and disclose limitations. AI embedded in a product can create liability for defects despite software disclaimers. Evaluate applicable product-safety and liability requirements alongside rigorous fault testing.
Question 696
An energy provider's data science team approves a customer forecasting model without consulting safety, accessibility or customer-service teams. Which governance reasoning best supports the appropriate response?
Explanation
Correct Answer
Interdisciplinary governance surfaces risks invisible to a single technical specialty. This reasoning fits cross-functional risk council because the observed gap is that review minutes include model metrics but omit operational harms and stakeholder perspectives. It points to the practical intervention: Create a cross-functional review with technical, legal, domain and affected-user expertise at design gates. The outcome should then show that approval packets include the required disciplines and recorded resolution of raised issues.
Incorrect Answers
Building and using an AI system create different accountability obligations even when shared among organizations. This reasoning fits cross-functional risk council because the observed gap is that review minutes include model metrics but omit operational harms and stakeholder perspectives. Interdisciplinary governance surfaces risks invisible to a single technical specialty.
Governance communication should make accountable choices possible, not simply circulate raw metrics. This reasoning fits cross-functional risk council because the observed gap is that review minutes include model metrics but omit operational harms and stakeholder perspectives. Interdisciplinary governance surfaces risks invisible to a single technical specialty.
AI incidents can be safety, fairness or trust failures without fitting conventional security categories. This reasoning fits cross-functional risk council because the observed gap is that review minutes include model metrics but omit operational harms and stakeholder perspectives. Interdisciplinary governance surfaces risks invisible to a single technical specialty.
Question 697
A customer support agent was secure at launch, but new tools allow it to access financial accounts. Which governance reasoning best supports the appropriate response?
Explanation
Correct Answer
Model and tool changes can create security risks even if original penetration tests passed. This reasoning fits periodic red-team exercise because the observed gap is that the threat model predates a tool that can initiate account changes. It points to the practical intervention: Repeat adversarial testing when capabilities or access expand and at planned intervals. The outcome should then show that updated red-team tests verify boundaries and documented remediation of emerging attack paths.
Incorrect Answers
A post-release AI incident may demand broader governance response than ordinary bug triage. This reasoning fits periodic red-team exercise because the observed gap is that the threat model predates a tool that can initiate account changes. Model and tool changes can create security risks even if original penetration tests passed.
Safe AI release management includes the ability to reverse harmful changes. This reasoning fits periodic red-team exercise because the observed gap is that the threat model predates a tool that can initiate account changes. Model and tool changes can create security risks even if original penetration tests passed.
Responsible AI validation is continuous because populations and workflows change. This reasoning fits periodic red-team exercise because the observed gap is that the threat model predates a tool that can initiate account changes. Model and tool changes can create security risks even if original penetration tests passed.
Question 698
A loan recommendation model becomes less reliable after the bank changes product eligibility policies. Which governance reasoning best supports the appropriate response?
Explanation
Correct Answer
AI model performance depends on a changing operational context, not just code stability. This reasoning fits concept drift after policy change because the observed gap is that the source eligibility rule changes while the model's prior score distribution remains unchanged. It points to the practical intervention: Monitor outcome drift relative to changed policy and initiate governed revalidation. The outcome should then show that a validated update or suspension restores alignment with the current documented policy.
Incorrect Answers
AI failures can arise from data and human processes even when compute is operational. This reasoning fits concept drift after policy change because the observed gap is that the source eligibility rule changes while the model's prior score distribution remains unchanged. AI model performance depends on a changing operational context, not just code stability.
Risk documentation should evolve with the actual AI system, not remain frozen at initial approval. This reasoning fits concept drift after policy change because the observed gap is that the source eligibility rule changes while the model's prior score distribution remains unchanged. AI model performance depends on a changing operational context, not just code stability.
Transparency requires actionable deployment information, not just generic AI branding. This reasoning fits concept drift after policy change because the observed gap is that the source eligibility rule changes while the model's prior score distribution remains unchanged. AI model performance depends on a changing operational context, not just code stability.
Question 699
A consumer app vendor claims its new AI moderation is safe based on a hand-picked demonstration of 20 prompts. Which additional observation would most strongly confirm the particular risk or control gap?
Explanation
Correct Answer
The shown prompts avoid ambiguous, multilingual and adversarial inputs that real customers send.
Incorrect Answers
The proposed pipeline treats anonymous labels as authoritative without provenance checks. The shown prompts avoid ambiguous, multilingual and adversarial inputs that real customers send.
Low-confidence advice is sent automatically while specialists never receive the flagged conversation. The shown prompts avoid ambiguous, multilingual and adversarial inputs that real customers send.
Evaluation measures cost savings but not delayed treatment reimbursements or appeal burden. The shown prompts avoid ambiguous, multilingual and adversarial inputs that real customers send.
Question 700
A firm keeps indefinitely every employee prompt to a drafting assistant although most contain short-lived customer queries. Which governance reasoning best supports the appropriate response?
Explanation
Correct Answer
AI-specific data flows should be covered by lifecycle retention governance. This reasoning fits data-retention policy for prompts because the observed gap is that configuration exports show indefinite logging with no retention owner or deletion route. It points to the practical intervention: Set purpose-based retention periods and verified deletion for conversational records. The outcome should then show that retention tests demonstrate logs age out or are deleted according to the approved schedule.
Incorrect Answers
Third-party AI assurance extends beyond uptime and price to lifecycle risk and provenance. This reasoning fits data-retention policy for prompts because the observed gap is that configuration exports show indefinite logging with no retention owner or deletion route. AI-specific data flows should be covered by lifecycle retention governance.
AI procurement is an organizational risk decision, not merely a software purchase. This reasoning fits data-retention policy for prompts because the observed gap is that configuration exports show indefinite logging with no retention owner or deletion route. AI-specific data flows should be covered by lifecycle retention governance.
Responsible AI requirements belong in the product design, not only in a final release memo. This reasoning fits data-retention policy for prompts because the observed gap is that configuration exports show indefinite logging with no retention owner or deletion route. AI-specific data flows should be covered by lifecycle retention governance.
Artificial Intelligence Governance Professional (AIGP) Practice Tests
- IAPP AIGP Practice Test 68 - 10 Questions (Last Reviewed 11-Nov-2026)
- IAPP AIGP Practice Test 69 - 10 Questions (Last Reviewed 10-Nov-2026)
- IAPP AIGP Practice Test 70 - 10 Questions (Last Reviewed 09-Nov-2026)
- IAPP AIGP Practice Test 71 - 10 Questions (Last Reviewed 08-Nov-2026)
- IAPP AIGP Practice Test 72 - 10 Questions (Last Reviewed 07-Nov-2026)
