Active

AI Security Review

The organization shall conduct periodic reviews of AI security activities to ensure their effectiveness, efficiency, and alignment with organizational needs. Reviews shall evaluate security controls, threat protection, monitoring, and incident response against established criteria and best practices, identifying opportunities for AI security improvement.

Requirement ID
AIG-005-R009
Parent Control
Domain Code

Requirement Details & Guidance

AIG-005-R009 · AIG-005
Why This Requirement Exists

This requirement exists to periodically review AI security controls, threat protection, monitoring, incident response, and resilience, ensuring AI systems are governed, secured, assessed, and operated in a controlled and accountable manner.

Implementation Guidance

Establish and document the process to periodically review AI security controls, threat protection, monitoring, incident response, and resilience. Define accountable roles, scope, required inputs and outputs, assessment or approval criteria, decision thresholds, review frequency, monitoring, reporting, exception handling, evidence requirements, and escalation paths. Apply the process across the AI lifecycle and integrate it with information security, privacy, risk, compliance, change management, incident management, and business governance as applicable.

Practical Application & Audit Readiness

AIG-005-R009 · Field Guidance

Business Examples

    Common Audit Findings

      Common Mistakes

        Best Practices & Ownership

        AIG-005 · Recommended Approach

        Recommended Practices

          Typical Owner

          Cross-Mapped Regulatory Frameworks

          Harmonized Control Mappings
          FRAMEWORK CODE FRAMEWORK NAME & DETAILS FAMILY / CATEGORY VERSION ACTION
          IRDAI IRDAI Information & Cyber Security Guidelines IRDAI information and cyber-security requirements... Clause: §2.16-3.6 – Information Security assessment
          India
          Guideline
          vLatest View →
          SEBI-CSCRF SEBI Cyber Security & Cyber Resilience Framework SEBI’s Cybersecurity and Cyber Resilience Framewor... Clause: 3.1 PR.AA: Identity Management, Authentication, and Access Control; 3.4 PR.IP: Information Protection Processes and Procedures
          India
          Regulation
          vLatest View →
          ISO27002 ISO/IEC 27002 ISO/IEC 27002 is a guidance standard containing a... Clause: 8.16 – Monitoring activities
          ISO
          Standard
          v2022 View →
          PCI-DSS PCI DSS PCI DSS is an industry security standard maintaine... Clause: 6.2.4 – Software engineering techniques or other; 6.3.1 – Security vulnerabilities are identified and
          PCI
          Standard
          v4 View →
          COBIT2019 COBIT COBIT is an enterprise governance and management f... Clause: 4.2 – Align, Plan and Organize (APO) > APO13 – Managed Security > APO13.01 – Establish and maintain an information security management system.
          COBIT
          Governance Framework
          v2019 View →
          NIST-CSF NIST Cybersecurity Framework (CSF) The NIST Cybersecurity Framework provides high-lev... Clause: PR.PS – Platform Security | PR.PS-06 – Secure software development practices are integrated, and their performance is monitored throughout the software development life cycle
          NIST
          Framework
          v2 View →
          NIST80053 NIST SP 800-53 NIST SP 800-53 Rev. 5 presents control families, i... Clause: 3.4 – ASSESSMENT, AUTHORIZATION, AND MONITORING; CA-7 – CONTINUOUS MONITORING, CA-2 – CONTROL ASSESSMENTS
          NIST
          Control Catalog
          vRev. 5 View →
          GDPR General Data Protection Regulation (GDPR) The General Data Protection Regulation is a bindin... Clause: Article 32(1) – Security of processing; Article 32(1)(d) – Testing, assessing and evaluating security measures
          EU
          Regulation
          vCurrent View →
          CERT-IN CERT-In Directions The CERT-In Directions issued under India’s Inform... Clause: Annexure I(xx) – Attacks or malicious/suspicious activities affecting systems/servers/software/applications related to Artificial Intelligence and Machine Learning
          India
          Direction
          v2022 View →
          SOC2 SOC 2 Trust Services Criteria SOC 2 is an independent examination and reporting... Clause: CC6.1 – Logical and Physical Access Controls – The entity implements logical access security software, infrastructure, and architectures over protected information assets to protect them from security events to meet the entity's objectives. CC7.1 – System Operations – To meet its objectives, the entity uses detection and monitoring procedures to identify (1) changes to configurations that result in the introduction of new vulnerabilities, and (2) susceptibilities to newly discovered vulnerabilities. CC7.3 – System Operations – The entity evaluates security events to determine whether they could or have resulted in a failure of the entity to meet its objectives (security incidents) and, if so, takes actions to prevent or address such failures. CC4.1 – Monitoring of Controls – COSO Principle 16: The entity selects, develops, and performs ongoing and/or separate evaluations to ascertain whether the components of internal control are present and functioning.
          AICPA
          Audit Framework
          vLatest View →

          Parent Domain Context

          Governance Domain Baseline
          Domain Name:
          AI Governance & Model Security (AIG)
          Purpose:
          To ensure the trustworthy, secure, and ethical development and deployment of AI.
          Scope:
          Enterprise

          Parent Control Context

          Control Baseline
          Control Name:
          AI Security & Operational Resilience (AIG-005)
          Summary:
          Establishes AI security including adversarial protection, prompt injection protection, model poisoning protection, access security, monitoring, incident response, resilience testing, review, and continuous improvement.
          Business Description:
          Establish and maintain controls to protect AI systems against adversarial attacks, model compromise, data poisoning, prompt injection, model theft, unauthorized use, operational failures, and AI-related security incidents.