ISACA AAISM Certification Exam Guide: Advanced in AI Security Management

Artificial intelligence is quickly becoming part of enterprise security architecture, business applications, cloud platforms, software development, and security operations. That creates opportunities, but it also introduces risks that traditional cybersecurity programs were not originally designed to handle. The AAISM Certification, formally known as ISACA Advanced in AI Security Management (AAISM), is designed for experienced security professionals who need to manage those risks at an enterprise level.

AAISM is not simply an artificial intelligence fundamentals exam. It combines AI governance, security management, risk assessment, data protection, AI architecture, incident response, vendor management, privacy, and technical controls. This makes the certification particularly interesting for security managers who already understand traditional information security but now need to apply those principles to AI systems.

This guide explains the AAISM exam structure, certification requirements, technical knowledge areas, career value, and a practical study strategy for candidates preparing for the exam.

What Is the ISACA AAISM Certification?

The Advanced in AI Security Management certification was developed by ISACA to validate the ability of experienced security professionals to secure enterprise AI solutions while managing the risks associated with their development, deployment, and operation.

ISACA positions AAISM as an AI-centric security management certification that builds on established information security management knowledge. Rather than asking candidates to become machine learning researchers or data scientists, the certification focuses on a more practical question:

How should an organization govern, secure, monitor, and manage the risks of AI systems?

This distinction is important. An AI engineer may understand model training and inference extremely well but have limited experience with enterprise risk governance. A security manager may understand access control, incident response, and security policy but have little exposure to model poisoning, prompt injection, AI supply chains, or model lifecycle risks.

AAISM sits between these worlds.

Is AAISM an Entry-Level AI Security Certification?

No. This is probably the most important fact for students and newcomers to understand.

The AAISM Certification is not an entry-level cybersecurity certification. ISACA requires candidates pursuing the credential to hold an active CISM (Certified Information Security Manager) or CISSP (Certified Information Systems Security Professional) certification.

ISACA also describes the target audience as professionals with proven security or advisory experience and some familiarity with assessing, implementing, or maintaining AI systems.

For students and junior network or security engineers, AAISM can still be useful as a learning roadmap. The exam objectives show where enterprise cybersecurity is moving and what skills security leaders will increasingly need. However, most beginners should first establish solid knowledge of networking, operating systems, cloud security, IAM, risk management, incident response, and security architecture before treating AAISM as an immediate certification target.

AAISM Exam Overview

The following table summarizes the current AAISM exam information. Exam policies and pricing may change, so candidates should always confirm the latest details with ISACA before registration.

One point worth remembering is that a 450 passing score is a scaled score. It should not be interpreted as simply answering 56.25% of the questions correctly. ISACA converts raw results to its standardized scoring scale.

Candidates who want to organize their preparation around the exam structure can also review these AAISM study resources while working through each domain.

AAISM Certification Exam Domains

The ISACA AAISM exam covers three major job-practice domains. Understanding their weighting should influence how you allocate your study time.

AI Technologies and Controls has the largest weighting at 38%, but the three domains are close enough that ignoring any one of them would be a poor exam strategy.

Domain 1: AI Governance and Program Management – 31%

Security technologies only work effectively when an organization has clear governance around how they should be used. The first AAISM domain therefore concentrates heavily on enterprise-level decision making.

Important areas include:

• AI governance structures
• Stakeholder roles and responsibilities
• Industry frameworks and regulatory requirements
• AI security policies and procedures
• AI asset inventories
• Data lifecycle management
• AI security program development
• Security metrics
• AI-related incident response
• Business continuity and disaster recovery

AI Governance Is More Than Writing a Policy

A common mistake is to think of AI governance as a document created by the compliance department. Real governance requires decisions about ownership, accountability, acceptable use, risk tolerance, data access, model approval, vendor selection, monitoring, and incident escalation.

Consider an enterprise that allows employees to use generative AI assistants. A mature security program needs to answer questions such as:

• What information may employees submit to external AI services?
• Who approves new AI platforms?
• How are AI applications added to the enterprise asset inventory?
• What logs must be retained?
• How should sensitive information be handled?
• Who investigates an AI-related security incident?
• When must legal, privacy, compliance, or executive teams become involved?

The AAISM perspective is therefore managerial as well as technical.

AI Asset and Data Lifecycle Management

Traditional asset inventories might contain servers, routers, applications, databases, and cloud resources. AI environments introduce additional assets that must be understood and classified.

These can include:

• Training datasets
• Fine-tuning datasets
• Foundation models
• Model weights
• Embeddings
• Vector databases
• System prompts
• AI agents
• Model APIs
• RAG knowledge sources
• Third-party AI services

A security manager needs visibility into these assets before meaningful controls can be designed.

Domain 2: AI Risk Management – 31%

The second major area of the AAISM certification exam focuses on identifying, assessing, treating, and continuously reviewing AI-related risk.

Major topics include:

• AI risk assessments
• Risk thresholds and risk appetite
• Risk treatment
• AI threat identification
• Vulnerability management
• AI impact assessments
• Vendor security
• Third-party AI services
• AI supply-chain risk
• Continuous risk reassessment

Why Traditional Risk Assessments Need an AI Perspective

Confidentiality, integrity, and availability remain important, but AI introduces additional characteristics that security teams must consider.

For example, organizations may need to evaluate:

• Model reliability
• Robustness
• Explainability
• Bias
• Data provenance
• Model integrity
• Human oversight
• Privacy impact
• Safety
• Unintended model behavior

The goal is not to replace traditional enterprise risk management. Instead, AI risks should be incorporated into the organization’s existing risk processes.

AI Threats Candidates Should Understand

The AAISM exam outline emphasizes AI threats and vulnerabilities. Candidates should therefore understand how conventional attacks interact with AI-specific weaknesses.

Useful examples to study include:

• Prompt injection: manipulating AI instructions so a model behaves in an unintended way.
• Indirect prompt injection: malicious instructions hidden in external content that an AI application later retrieves or processes.
• Data poisoning: manipulating training, fine-tuning, or retrieval data to influence model behavior.
• Model theft: unauthorized acquisition or reconstruction of valuable models or model parameters.
• Sensitive information disclosure: an AI system exposing confidential, personal, or proprietary information.
• Insecure output handling: blindly trusting model-generated output before passing it to another application or system.
• Excessive agency: allowing an AI agent to perform actions with more permissions or autonomy than necessary.
• AI supply-chain compromise: risk introduced through external models, datasets, libraries, APIs, plugins, and service providers.

Frameworks such as the OWASP GenAI Security Project and MITRE ATLAS are useful supplements because they provide concrete attack patterns that make abstract AI risk concepts easier to understand.

Domain 3: AI Technologies and Controls – 38%

At 38%, AI Technologies and Controls is the largest AAISM exam domain.

This section connects security management with the technical architecture of AI systems. Candidates do not need to become machine learning researchers, but they should understand enough about the AI lifecycle and AI architecture to make informed security decisions.

Key areas include:

• AI security architecture and design
• AI system lifecycle
• Model selection
• Model training
• Model validation
• Data security controls
• Privacy controls
• Ethical and trust controls
• AI safety controls
• Security monitoring

Secure AI Architecture

A production AI application usually contains much more than a model. Depending on the architecture, the complete solution might include an application frontend, APIs, model endpoint, authentication service, vector database, retrieval system, data pipeline, logging platform, cloud infrastructure, and external tools.

Each component creates a potential security boundary.

Security architects should consider controls such as:

• Strong authentication
• Least-privilege authorization
• Workload identities
• API gateways
• Network segmentation
• Encryption in transit and at rest
• Secrets management
• Input filtering
• Output validation
• Data loss prevention
• Model access controls
• Security logging
• Rate limiting
• Continuous monitoring

Securing Retrieval-Augmented Generation

Retrieval-Augmented Generation, or RAG, is widely used to connect large language models with enterprise information. It can improve relevance, but it also creates security challenges.

A poorly designed RAG environment might allow one employee to retrieve documents belonging to another department simply because the underlying vector database does not enforce the original access permissions.

Security teams should therefore think about:

• Document-level access controls
• Data classification
• Embedding security
• Vector database permissions
• Source validation
• Malicious document ingestion
• Indirect prompt injection
• Retrieval logging

This is a useful example of why AI security cannot be separated from traditional IAM, application security, and data security.

Human Oversight Still Matters

Automation is one of AI’s greatest advantages, but excessive automation can also increase operational risk.

If an AI security system identifies an account as malicious, should it immediately terminate every user session? If an AI agent believes that a server is compromised, should it automatically isolate the host from the network?

The answer depends on risk, confidence, business impact, and organizational policy.

Human oversight can be particularly important for high-impact decisions. Candidates should be comfortable thinking about where automation is appropriate and where approval, escalation, or verification should remain part of the process.

AI Security Management vs. Traditional Cybersecurity

Many traditional cybersecurity controls remain relevant in AI environments. Identity management, vulnerability management, encryption, logging, incident response, network segmentation, and secure software development do not suddenly disappear because an application contains an AI model.

The difference is that new assets and attack paths are added to the architecture.

This is one reason experienced CISM and CISSP professionals are a natural audience for AAISM. They already understand the security foundation and can concentrate on how AI changes the risk model.

Important AI Security Frameworks to Know

AAISM is not simply an exercise in memorizing one framework. However, several external resources are particularly valuable when studying AI governance and security.

NIST AI Risk Management Framework

The NIST AI Risk Management Framework (AI RMF) provides a structured approach for organizations designing, developing, deploying, or using AI systems.

Its core functions—Govern, Map, Measure, and Manage—provide a useful mental model for thinking about AI risk throughout the lifecycle.

NIST Generative AI Profile

NIST’s Generative Artificial Intelligence Profile extends the AI RMF with considerations specific to generative AI. It discusses areas such as information security, data privacy, content integrity, human-AI interaction, and other risks created or amplified by generative systems.

OWASP GenAI Security Project

OWASP provides practical guidance for securing LLM and generative AI applications. Its risk lists are especially useful for technical security professionals studying attacks such as prompt injection, sensitive information disclosure, supply-chain weaknesses, data and model poisoning, improper output handling, and excessive agency.

MITRE ATLAS

MITRE ATLAS provides a knowledge base of adversarial tactics and techniques targeting AI-enabled systems. Security professionals familiar with MITRE ATT&CK will find the structure useful for understanding AI-focused threat modeling and red-team activities.

ISO/IEC 42001

ISO/IEC 42001 defines requirements for establishing and continually improving an Artificial Intelligence Management System. It is particularly relevant when studying how AI governance can be integrated into enterprise management processes.

How AAISM Relates to CISM

CISM and AAISM are complementary rather than competing certifications.

CISM addresses the broader management of an enterprise information security program, including governance, risk management, security program development, and incident management.

AAISM takes that management foundation and applies it specifically to AI.

For a CISM holder moving into enterprise AI governance, AI security architecture, or AI risk leadership, AAISM provides a logical specialization.

How AAISM Relates to CISSP

CISSP provides broad cybersecurity knowledge across architecture, engineering, security operations, IAM, software development security, asset security, and risk management.

Those skills remain highly relevant when securing AI systems.

AAISM adds the AI-specific governance and management layer, including topics such as model lifecycle security, AI data risks, responsible AI, human oversight, model-related threats, and AI vendor risk.

A CISSP with architecture or security engineering experience may therefore find Domain 3 relatively familiar conceptually, while needing additional study in AI governance and AI-specific risk management.

Who Should Consider the AAISM Certification?

AAISM is most relevant to professionals who already work around security leadership, architecture, governance, risk, or enterprise AI.

Typical roles include:

• Information Security Manager
• Cybersecurity Manager
• Security Architect
• AI Security Architect
• Security Consultant
• AI Governance Manager
• Cyber Risk Manager
• GRC Manager
• Cloud Security Architect
• Security Program Manager
• Chief Information Security Officer
• AI Security or Responsible AI Advisor

AAISM may also be attractive to CISM and CISSP holders whose organizations are adopting generative AI platforms, copilots, machine learning applications, or autonomous AI agents.

Is the AAISM Certification Worth It?

The value of any certification depends on your role and career direction.

AAISM makes the most sense if you already have significant security knowledge and expect AI security to become part of your responsibilities. This may include reviewing AI projects, defining enterprise AI policies, assessing AI vendors, securing generative AI applications, overseeing AI-related incidents, or advising senior management on AI risk.

The credential is less compelling for someone who is only beginning to learn cybersecurity. In that situation, foundational networking, operating systems, cloud computing, security architecture, IAM, and incident response skills will normally provide more immediate value.

AAISM should also not be viewed as a substitute for hands-on experience. A certification can validate structured knowledge, but employers will still value professionals who can translate governance requirements into practical security architecture and operational controls.

How to Prepare for the AAISM Exam

A good AAISM exam preparation strategy should combine the official ISACA material with practical AI security resources and repeated review of scenario-based questions.

1. Start With the Official Exam Content Outline

Read every domain and subtopic before beginning detailed study. The exam content outline tells you what ISACA expects candidates to understand and prevents you from spending too much time on interesting AI subjects that are outside the exam’s management focus.

2. Use the AAISM Official Review Manual

The official review manual should form the backbone of your preparation because it follows ISACA’s exam domains.

As you study, organize notes into the three domains rather than simply reading from beginning to end.

3. Give Domain 3 Slightly More Study Time

AI Technologies and Controls represents 38% of the exam, compared with 31% for each of the other domains.

That does not justify ignoring governance or risk management, but it does mean architecture, lifecycle security, data protection, privacy, trust, safety, and monitoring deserve substantial attention.

4. Learn the Concepts Instead of Memorizing Definitions

ISACA exams commonly present management scenarios where several answers appear technically possible.

When answering a question, think about:

• Business objectives
• Risk
• Governance
• Accountability
• Policy
• Prioritization
• Appropriate stakeholder involvement
• Long-term control effectiveness

For example, immediately blocking an AI system might technically reduce a security risk, but it may not be the best management decision if the risk can be reduced to an acceptable level while supporting an important business objective.

5. Study Real AI Attack Patterns

Use resources such as OWASP and MITRE ATLAS to understand how attacks work in practice.

Being able to explain why prompt injection, data poisoning, excessive agency, or model supply-chain compromise matters is far more useful than memorizing a one-line definition.

6. Practice Scenario-Based Questions

Practice questions are useful for learning how ISACA structures its decisions. Concentrate on understanding why an answer is preferred rather than memorizing question wording.

After every practice session, identify weak domains and return to the related material. Candidates looking for a structured place to reinforce these areas can review the ISACA AAISM exam course alongside official ISACA documentation.

A Practical Eight-Week AAISM Study Plan

Experienced AI security professionals may require less time, while candidates whose background is primarily traditional security management may want additional time for AI architecture and machine learning concepts.

During the final weeks, it can be helpful to combine official documentation, personal notes, and focused AAISM certification exam preparation so that weak topics are identified before exam day.

Maintaining the AAISM Certification

Passing the exam is only part of the certification process.

Current ISACA requirements include holding an active qualifying CISM or CISSP credential, passing the AAISM exam, submitting the certification application, paying the application processing fee, agreeing to ISACA’s professional requirements, and meeting continuing professional education requirements.

After certification, AAISM holders are required to earn and report 10 CPE hours annually in the specialized domain of artificial intelligence, beginning in the calendar year after certification.

This continuing education requirement is especially relevant for AI security because the technology and threat landscape change much faster than many traditional areas of information security.

Final Thoughts

The ISACA AAISM Certification reflects an important shift in cybersecurity. AI security is no longer only a research problem or a responsibility for machine learning engineers. As AI becomes integrated into business processes and security operations, organizations need professionals who can connect AI technology with governance, risk management, security architecture, privacy, and incident response.

AAISM is designed for experienced CISM and CISSP professionals rather than beginners, but its exam domains provide a useful roadmap for anyone interested in the future of enterprise security.

If you are already responsible for information security management and your organization is adopting generative AI, machine learning platforms, AI copilots, or autonomous agents, studying for AAISM can help you develop a more structured approach to identifying and managing the risks those technologies introduce.

The strongest candidates will not simply memorize AI terminology. They will understand how business objectives, governance, risk, architecture, data, people, processes, and technical controls work together to create a secure and responsible AI environment.

If you are preparing for the exam, you can also explore the AAISM Advanced in AI Security Management course for additional study support and exam-focused preparation.

Frequently Asked Questions About AAISM

What does AAISM stand for?

AAISM stands for Advanced in AI Security Management, an ISACA certification focused on managing and securing enterprise artificial intelligence environments.

How many questions are on the AAISM exam?

The AAISM exam currently contains 90 multiple-choice questions.

How long is the AAISM exam?

Candidates currently have 150 minutes, or 2.5 hours, to complete the exam.

What score do you need to pass AAISM?

ISACA requires a scaled score of 450 or higher on its 200–800 scoring scale.

Do I need CISM before taking AAISM?

Candidates must hold an active CISM or CISSP credential to pursue AAISM. A candidate does not specifically need CISM if they already hold an active CISSP.

Is AAISM suitable for beginners?

AAISM is an advanced certification and is not designed as an entry-level cybersecurity credential. Beginners can use the AAISM syllabus as a learning roadmap, but they should normally build strong cybersecurity and risk-management foundations first.

Does AAISM cover generative AI and LLM security?

The exam covers AI technologies, security architecture, AI lifecycle controls, data management, security monitoring, risk, privacy, trust, and safety. Generative AI and LLM security concepts such as prompt injection, model supply-chain risk, sensitive information disclosure, and excessive agency are useful practical areas to understand when preparing for modern AI security scenarios.

Is AAISM better than CISM?

The certifications serve different purposes. CISM validates broad information security management expertise, while AAISM builds on established security-management knowledge with a specialization in AI security. For many professionals, the two credentials are complementary.

Official and Authoritative AAISM Resources

The following resources are useful for verifying current exam requirements and expanding your understanding of AI security management:

• ISACA – Advanced in AI Security Management (AAISM)
  Official certification information, eligibility requirements, registration, preparation materials, and certification details.
  https://www.isaca.org/credentialing/aaism
• ISACA – AAISM Exam Content Outline
  Official AAISM domains, weights, subtopics, and job-practice tasks.
  https://www.isaca.org/credentialing/aaism/aaism-exam-content-outline
• ISACA – Exam Candidate Guides
  Official information about exam scheduling, scoring, retakes, testing rules, and certification procedures.
  https://www.isaca.org/credentialing/exam-candidate-guides
• NIST – Artificial Intelligence Risk Management Framework
  A widely referenced framework for identifying and managing risks associated with AI systems.
  https://www.nist.gov/itl/ai-risk-management-framework
• NIST – Generative AI Profile
  NIST guidance extending the AI Risk Management Framework to risks associated with generative artificial intelligence.
  https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence
• OWASP GenAI Security Project – LLM Top 10
  Practical guidance covering major vulnerabilities and security risks affecting generative AI and LLM applications.
  https://genai.owasp.org/resource/owasp-genai-llm-top-10-2026/
• MITRE ATLAS
  A knowledge base describing adversarial tactics and techniques targeting artificial intelligence systems.
  https://atlas.mitre.org/
• ISO/IEC 42001 – Artificial Intelligence Management Systems
  International standard covering the establishment, implementation, maintenance, and continual improvement of AI management systems.
  https://www.iso.org/standard/42001