IT Cert Ready

ISACA

AAISM

Advanced in AI Security Management

Start practicing

About this certification

AAISM (Advanced in AI Security Management) is a ISACA certification. IT Cert Ready offers 1,497 practice questions with domain quizzes, full mock exams, and detailed explanations. Practice by official domain with 3 blueprint-aligned domains.

Official exam domain weights

Verified blueprint weights from the official exam outline.

  • Domain 1 · AI Governance and Program Management · 31%
  • Domain 2 · AI Risk Management · 31%
  • Domain 3 · AI Technologies and Controls · 38%

Sample AAISM practice questions

Preview realistic multiple-choice items from the IT Cert Ready question bank. Expand each answer for the correct choice and explanation—the same format used in domain quizzes and mock exams.

  1. Sample question 1 · AI Governance and Program Management (31%)

    Veltronix Precision Components, a mid-sized German automotive parts manufacturer, operates computer-vision (CV) models on its stamping lines to detect surface defects in real time. During a routine shift handover, a line technician named Markus notices that the CV system has been flagging an unusually high number of false positives for the past 90 minutes, causing repeated line stoppages. Markus assumes it is a camera calibration issue and manually overrides the AI rejection flags to keep production running. He does not log the override or notify the AI security team. Two hours later, the quality assurance supervisor discovers that 4,200 potentially defective parts have shipped to a Tier-1 customer. The AI security program lead is now reviewing the incident and finds that Veltronix has no formal AI-specific acceptable use guideline addressing operator override authority or escalation triggers for anomalous model behavior. As the AI security program lead at Veltronix, which action should you prioritize FIRST to prevent a recurrence of this type of incident?

    • A. Immediately retrain the CV model using the flagged false-positive images to improve its accuracy before issuing any policy update.
    • B. Establish an AI acceptable use guideline that defines operator override authority limits, mandatory logging requirements, and explicit escalation triggers when model behavior deviates from baseline thresholds.
    • C. Require all production line technicians to complete a general cybersecurity awareness course covering phishing and password hygiene within the next 30 days.
    • D. Commission a third-party audit of the CV model's training data to determine whether data poisoning caused the false-positive surge before taking any policy action.
    View answer & explanation

    Correct answer: B. Establish an AI acceptable use guideline that defines operator override authority limits, mandatory logging requirements, and explicit escalation triggers when model behavior deviates from baseline thresholds.

    [Domain 1 - AI Governance and Program Management, Subtopic 1B - AI-Related Strategies, Policies, and Procedures - Area 1 - AI-Related Strategies, Policies, and Procedures] Core Concept: AI-specific acceptable use guidelines must define the boundaries of human interaction with AI systems, including who may override model outputs, under what conditions, and what escalation steps are mandatory. Without these controls, operators act on personal judgment, creating unlogged, ungoverned interventions that bypass the AI system's intended safety function and obscure security or quality incidents. Correct Answer (B): The root cause of this incident is the absence of a policy governing operator override authority and escalation obligations. Establishing an acceptable use guideline that codifies override limits, mandatory logging, and escalation triggers directly addresses the governance gap and prevents future unilateral, unrecorded interventions. Correct Answer (B): Correct Answer (B): Correct Answer (B): This is the foundational control that all subsequent technical or training measures depend upon. Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Option A: : : : Retraining the model addresses a potential technical symptom but does not resolve the governance failure — operators would still have no policy guidance on when and how overrides are permitted, leaving the same behavioral gap in place. Option C: : : : A generic cybersecurity awareness course covering phishing and password hygiene is irrelevant to the specific AI-human interaction risk exposed here; AI-specific acceptable use training tied to model override procedures is what is required. Option D: : : : Commissioning a data-poisoning audit may be warranted as a parallel investigation, but delaying policy action until the audit concludes leaves the override gap open and does not address the immediate governance deficiency that caused the business impact.

  2. Sample question 2 · AI Technologies and Controls (38%)

    Caelum Digital, a fast-growing startup in Singapore providing AI-powered contract analysis services to legal firms across Southeast Asia, recently onboarded a third-party LLM vendor, Veridian NLP, to power its document summarization feature. Veridian NLP operates under a shared-responsibility model in which Caelum Digital fine-tunes the base model using its own client document corpus. Three weeks after the fine-tuning job completes and the updated model is deployed to production, a legal firm client reports that the model is generating summaries that include verbatim excerpts from documents belonging to a different client — a clear data leakage event. Caelum Digital's AI security program lead is notified at 9:00 PM on a Friday. The affected clients include a multinational corporation and a government-linked entity, both with strict confidentiality obligations. As the AI security program lead at Caelum Digital, what is the MOST appropriate sequence of actions to take immediately upon receiving the data leakage report?

    • A. Notify Veridian NLP's technical team and request that they investigate the fine-tuning pipeline before taking any action on the production model, to avoid disrupting client services over the weekend.
    • B. Disable the fine-tuned model in production, activate the incident response plan, notify affected clients and relevant data protection authorities within required timeframes, and preserve all fine-tuning artifacts and logs for forensic review.
    • C. Retrain the model immediately using a sanitized version of the document corpus to eliminate the leakage, then redeploy to production before business hours on Monday to minimize client impact.
    • D. Conduct an internal root cause analysis over the weekend, and if the leakage is confirmed to be systemic, escalate to executive leadership on Monday morning before notifying clients or regulators.
    View answer & explanation

    Correct answer: B. Disable the fine-tuned model in production, activate the incident response plan, notify affected clients and relevant data protection authorities within required timeframes, and preserve all fine-tuning artifacts and logs for forensic review.

    [Domain 3 - AI Technologies and Controls, Subtopic 3B - AI Life Cycle (Model Selection, Training, and Validation) - Area 1 - AI Life Cycle (e.g., Model Selection, Training, and Validation)] Core Concept: When a fine-tuned LLM deployed in production begins leaking training data across client boundaries, this represents both a model life cycle control failure — specifically a failure in training data isolation and output validation — and a data breach requiring immediate incident response. The AI security program lead must treat this as a security incident with regulatory notification obligations, not merely a model performance issue. Preserving forensic artifacts from the fine-tuning pipeline is essential for root cause analysis and regulatory accountability. Correct Answer (B): Disabling the fine-tuned model immediately stops the ongoing leakage and protects all clients from further exposure. Activating the incident response plan ensures the organization follows its documented escalation and notification procedures. Notifying affected clients and data protection authorities within required timeframes (e.g., Singapore's PDPA mandates notification within three business days for significant breaches) fulfills legal obligations. Correct Answer (B): Correct Answer (B): Correct Answer (B): Preserving fine-tuning artifacts and logs prevents evidence destruction and supports the forensic investigation needed to determine whether the leakage stems from inadequate data partitioning, memorization during fine-tuning, or a misconfigured inference pipeline. Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Option A: : : : Delegating the initial response to the vendor before taking containment action allows the leakage to continue and shifts accountability inappropriately. The AI security program lead owns the incident response for Caelum Digital's production environment regardless of the shared-responsibility model with Veridian NLP. Option C: : : : Retraining and redeploying the model over the weekend without completing a root cause analysis risks deploying a model that still contains the underlying vulnerability. It also destroys or overwrites forensic artifacts needed to understand the scope of the breach and fulfill regulatory reporting obligations. Option D: : : : Delaying client and regulatory notification until Monday morning to complete an internal analysis violates data breach notification requirements and prolongs client exposure. Regulatory frameworks do not permit organizations to delay notification pending internal convenience; the clock for mandatory reporting begins at the point of discovery.

  3. Sample question 3 · AI Governance and Program Management (31%)

    Northgate Meridian Broadcasting (NMB), a mid-sized European media company operating under EU AI Act obligations, deployed a content moderation AI called FilterCore to screen user-submitted comments across its 14 regional news portals. Three weeks after launch, the security operations team detects a pattern of coordinated prompt injection attempts: external actors are submitting specially crafted comment strings designed to override FilterCore's moderation logic and allow hate speech to pass undetected. The AI security program lead reviews incident logs and discovers that fewer than 20% of the 340 content moderators and portal administrators who interact with FilterCore daily have completed the mandatory AI acceptable use training. The remaining staff were granted system access provisionally pending training completion. The CISO demands an immediate briefing on how the training gap contributed to the incident and what corrective action should be taken first. As the AI security program lead at NMB, which action should you prioritize FIRST to address the training gap's contribution to the ongoing prompt injection incident?

    • A. Suspend FilterCore operations entirely until all 340 staff complete the full AI security awareness curriculum, preventing further exposure while training is remediated.
    • B. Issue an emergency just-in-time training module specifically covering prompt injection recognition and acceptable use boundaries for FilterCore, targeting the untrained 80% before restoring full portal access, and update the acceptable use guidelines to require training completion as a hard prerequisite for system access.
    • C. Escalate the incident to the EU AI Act supervisory authority immediately, citing the training deficiency as a compliance breach, and defer internal corrective action until regulatory guidance is received.
    • D. Require the vendor that built FilterCore to patch the model against prompt injection vulnerabilities, treating the incident as a product defect rather than a training or policy failure.
    View answer & explanation

    Correct answer: B. Issue an emergency just-in-time training module specifically covering prompt injection recognition and acceptable use boundaries for FilterCore, targeting the untrained 80% before restoring full portal access, and update the acceptable use guidelines to require training completion as a hard prerequisite for system access.

    [Domain 1 - AI Governance and Program Management, Subtopic 1B - AI-Related Strategies, Policies, and Procedures - Area 1 - AI-Related Strategies, Policies, and Procedures] Core Concept: AI-specific security awareness training and acceptable use guidelines must be maintained as living controls that adapt to emerging threats; when a training gap is identified during an active incident, the program lead's first obligation is to close the knowledge gap in a targeted, risk-proportionate manner without unnecessarily halting business operations. Acceptable use guidelines should enforce training completion as a hard access prerequisite, and emergency training interventions must be scoped to the specific threat vector in play. Correct Answer (B): Issuing a targeted just-in-time module on prompt injection recognition directly addresses the knowledge gap that enabled the incident to escalate, while updating the acceptable use guidelines to enforce training-before-access closes the systemic policy failure. Correct Answer (B): Correct Answer (B): Correct Answer (B): This approach balances business continuity — keeping portals operational under heightened monitoring — with immediate risk reduction, which is the correct management posture when an active threat is linked to a training deficiency. Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Option A: : : : A full operational suspension is disproportionate to the risk and would cause significant business impact across 14 regional portals; acceptable use and training frameworks are designed to enable controlled operations, not mandate blanket shutdowns whenever a gap is found. Option C: : : : Regulatory escalation may eventually be required, but deferring all internal corrective action pending regulatory guidance abdicates the program lead's primary responsibility to contain the incident and remediate the control failure immediately. Option D: : : : While vendor patching may be a valid long-term control, framing the incident solely as a product defect ignores the documented training gap; acceptable use guidelines and staff awareness are independent, complementary controls that the AI security program lead owns and must address regardless of vendor remediation timelines.

  4. Sample question 4 · AI Governance and Program Management (31%)

    Fortbridge National Bank, a large multinational institution headquartered in the United Kingdom and subject to FCA and EU AI Act obligations, has embedded an AI-powered transaction monitoring system into its enterprise security metrics program. During a quarterly security review, the AI security program lead discovers that the bank's general cybersecurity awareness training module—last updated 18 months ago—contains no content specific to AI risks, adversarial manipulation of AI models, or employee responsibilities when interacting with AI-generated alerts. Front-line compliance analysts have been dismissing AI-generated suspicious transaction alerts at a rate 40% higher than the industry benchmark, citing alert fatigue. A recent internal audit finding noted that no AI-specific acceptable use guidelines exist separate from the general IT acceptable use policy. The AI security program lead must address the training and policy gap identified at Fortbridge National Bank. Which approach BEST aligns with leading practice for AI security awareness and acceptable use under the applicable regulatory context?

    • A. Reduce the volume of AI-generated alerts by adjusting the model's sensitivity threshold so that compliance analysts receive fewer notifications, thereby lowering the dismissal rate.
    • B. Incorporate AI-specific modules into the security awareness training program covering adversarial AI risks, responsible use of AI-generated outputs, and escalation obligations, and publish a standalone AI acceptable use guideline referencing EU AI Act human oversight requirements.
    • C. Require compliance analysts to complete the existing general IT acceptable use policy acknowledgment annually and add a brief AI disclaimer paragraph to the current cybersecurity training without creating separate AI-specific content.
    • D. Engage an external AI vendor to deliver a one-time AI literacy seminar for compliance analysts and defer policy updates until the next scheduled enterprise policy review cycle.
    View answer & explanation

    Correct answer: B. Incorporate AI-specific modules into the security awareness training program covering adversarial AI risks, responsible use of AI-generated outputs, and escalation obligations, and publish a standalone AI acceptable use guideline referencing EU AI Act human oversight requirements.

    [Domain 1 - AI Governance and Program Management, Subtopic 1B - AI-Related Strategies, Policies, and Procedures - Area 1 - AI-Related Strategies, Policies, and Procedures] Core Concept: Task 21 requires the AI security program lead to develop and maintain AI-specific security awareness training and acceptable use guidelines as distinct artifacts from general IT policies. Under the EU AI Act, high-risk AI systems such as transaction monitoring tools require documented human oversight mechanisms and trained personnel who understand their responsibilities when interacting with AI outputs. A standalone AI acceptable use guideline ensures clarity of roles, escalation paths, and responsible use obligations. Correct Answer (B): Developing AI-specific training modules that address adversarial risks, responsible interpretation of AI-generated alerts, and escalation obligations directly resolves the awareness gap driving the high dismissal rate. Correct Answer (B): Correct Answer (B): Correct Answer (B): Publishing a standalone AI acceptable use guideline that references EU AI Act human oversight requirements satisfies both the internal audit finding and the regulatory obligation to document human oversight for high-risk AI systems, making this the most comprehensive and compliant response. Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Option A: : : : Adjusting the model's sensitivity threshold is a technical tuning action that may reduce alert volume but does not address the root cause—employees lack the training and policy guidance to correctly evaluate and escalate AI-generated alerts. Option C: : : : Adding a brief AI disclaimer to existing general training and re-acknowledging the current IT policy is insufficient; it does not create the AI-specific content or standalone guidelines required by Task 21 or the EU AI Act's human oversight provisions. Option D: : : : A one-time external seminar is not a sustainable, maintained training program, and deferring policy updates to the next enterprise review cycle leaves the bank exposed to regulatory non-compliance and continued high alert dismissal rates in the interim.

  5. Sample question 5 · AI Governance and Program Management (31%)

    Caldera Community Credit Union, a small financial cooperative in Canada with approximately 85 employees, recently integrated an AI-based member authentication and fraud detection tool into its online banking platform. The tool is managed by a third-party fintech vendor. The AI security program lead, who also serves as the sole information security officer, is preparing to integrate AI security metrics into the credit union's enterprise information security program for the first time. During onboarding interviews with branch staff, the lead discovers that employees are unaware that the authentication tool uses AI, do not know how to recognize or report an AI-related authentication anomaly, and have received no guidance on what constitutes acceptable versus prohibited use of AI outputs in member-facing decisions. A member complaint about an erroneous account lock triggered by the AI system went unescalated for six business days. Given the six-day escalation delay and the staff awareness gaps identified at Caldera Community Credit Union, what should the AI security program lead do FIRST to establish a foundational control?

    • A. Terminate the contract with the fintech vendor immediately due to the AI system's erroneous account lock and procure a replacement solution with a more transparent AI model.
    • B. Conduct a full NIST AI RMF GOVERN function assessment of the vendor's AI system before taking any internal training or policy actions.
    • C. Develop and deliver a concise AI security awareness briefing for all staff that defines AI-related incident indicators specific to the authentication tool, establishes a clear escalation timeline, and outlines acceptable use boundaries for acting on AI-generated authentication decisions.
    • D. Escalate the six-day delay directly to the credit union's board of directors and request emergency budget approval for a dedicated AI security team before implementing any awareness or policy measures.
    View answer & explanation

    Correct answer: C. Develop and deliver a concise AI security awareness briefing for all staff that defines AI-related incident indicators specific to the authentication tool, establishes a clear escalation timeline, and outlines acceptable use boundaries for acting on AI-generated authentication decisions.

    [Domain 1 - AI Governance and Program Management, Subtopic 1B - AI-Related Strategies, Policies, and Procedures - Area 1 - AI-Related Strategies, Policies, and Procedures] Core Concept: In resource-constrained organizations, the AI security program lead must prioritize foundational awareness and acceptable use controls that directly address identified gaps causing business impact. Task 21 calls for developing and maintaining AI-specific security awareness training and acceptable use guidelines; in a small institution with no prior AI governance, the first actionable step is establishing baseline staff knowledge of AI incident indicators, escalation obligations, and use boundaries before more complex governance activities are undertaken. Correct Answer (C): Delivering a targeted AI security awareness briefing addresses all three identified gaps simultaneously: staff ignorance of the AI tool's existence, inability to recognize or report authentication anomalies, and absence of acceptable use guidance. Correct Answer (C): Correct Answer (C): Correct Answer (C): This directly prevents recurrence of the six-day escalation delay by giving employees the knowledge and procedural clarity needed to act promptly on AI-related incidents, and it is executable immediately without requiring additional budget or vendor action. Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Why the other options are incorrect: Option A: : : : Terminating the vendor contract is a disproportionate response to a single erroneous lock event and does not address the internal awareness and policy gaps that caused the escalation failure; a new vendor would face the same unaddressed staff knowledge gaps. Option B: : : : Conducting a full NIST AI RMF GOVERN assessment is a valuable governance activity but is not the most urgent first step when staff cannot recognize or escalate AI incidents; the immediate business impact risk from uninformed employees must be addressed first. Option D: : : : Escalating to the board for emergency budget approval before implementing any awareness measures delays the most immediately actionable control and conflates a governance escalation with an operational response; the program lead has authority to deliver awareness training without board approval.

AAISM exam prep FAQ

What is the passing score for the AAISM exam?
Passing scores for AAISM are set by ISACA and can change between exam versions. Check the latest official exam outline on the ISACA website for the current passing standard.
How many questions are on the AAISM certification exam?
Official AAISM exam length and item types are defined by ISACA. Review the current exam blueprint for the latest question counts and timing.
How do domain quizzes help with AAISM exam prep?
Domain quizzes on IT Cert Ready let you practice AAISM one blueprint domain at a time across 3 official exam domains. You get immediate feedback, detailed explanations, and can revisit weak areas before taking a full timed mock exam.
How many AAISM practice questions are on IT Cert Ready?
IT Cert Ready includes 1,497 published AAISM practice questions with domain quizzes, full-length mock exams, and answer explanations mapped to the official outline.