ISO 27001 to ISO 42001

ISO/IEC 27001:2022 and ISO/IEC 42001:2023 are designed to operate together. This map shows where each ISO 42001 Annex A control draws on an information-security control you may already run, and where it adds something specific to AI.

Carried over

Specific to AI

Security-only

What ISO 42001 adds for AI

The controls an AI management system introduces, where an information-security standard was not written to reach.

A.5.2AI system impact assessment process
Specific to AI
A.5.3Documentation of AI system impact assessments
Specific to AI
A.5.4Assessing AI system impact on individuals or groups
Specific to AI
A.5.5Assessing societal impacts of AI systems
Specific to AI
A.6.1.2Objectives for responsible development of AI
Specific to AI
A.7.4Quality of data for AI systems
Specific to AI
A.7.5Data provenance
Specific to AI
A.9.3Objectives for responsible use of AI systems
Specific to AI
A.9.4Intended use of the AI system
Specific to AI

Each ISO 42001 Annex A control, grouped by objective, and the ISO 27001 controls it draws on.

- A.2.2AI Policy
  Carried over
  Relates to
  A.5.1 Policies for information security
  Add an AI policy as part of the overall InfoSec policy framework.
- A.2.3Alignment with other organizational policies
  Carried over
  Relates to
  A.5.1 Policies for information security
  Ensure the AI policy aligns with InfoSec policies and the broader MS.
  A.5.4 Management responsibilities
  Management must integrate AI responsibilities into overall governance.
- A.2.4Review of the AI policy
  Carried over
  Relates to
  A.5.1 Policies for information security
  Periodically review the AI policy alongside InfoSec.
  A.5.36 Compliance with policies, rules & standards
  Integrate AI policy reviews in compliance checks.
- A.3.2AI roles & responsibilities
  Carried over
  Relates to
  A.5.2 Information security roles & responsibilities
  Document AI roles (e.g., AI Ethics Lead, AI Dev) in parallel with InfoSec roles.
- A.3.3Reporting of concerns
  Carried over
  Relates to
  A.5.24 InfoSec incident mgmt planning & prep
  AI "ethical" or security concerns can follow incident management workflow.
  A.6.4 The disciplinary process (partial)
  Partial synergy: potential disciplinary steps if AI misuse or unethical practices are reported.
- A.4.2Resource documentation
  Carried over
  Relates to
  A.5.9 Inventory of information & other associated assets
  Keep AI data/models, computing resources in your asset inventory.
- A.4.3Data resources
  Carried over
  Relates to
  A.5.9 Inventory of information & other associated assets
  Include AI datasets, pipelines, etc.
  A.5.14 Information transfer
  Manage and secure data flows for AI.
- A.4.4Tooling resources
  Carried over
  Relates to
  A.5.9 Inventory of information & other associated assets
  Also track specialized AI tools or frameworks as assets.
  A.8.18 Use of privileged utility programs
  Restrict or monitor advanced AI scripts/utilities.
- A.4.5System & computing resources
  Carried over
  Relates to
  A.8.6 Capacity management
  Overlaps in GPU/ML resource management.
  A.5.23 InfoSec for use of cloud services
  If AI is cloud-based, apply relevant cloud security controls.
- A.4.6Human resources
  Carried over
  Relates to
  A.6.1 Screening
  Screen or verify staff for AI competencies and trustworthiness.
  A.6.3 InfoSec awareness, education & training
  Extend training to include AI ethics, responsible development, etc.
- A.5.2AI system impact assessment process
  Specific to AI
  Relates to
  Clause 6.1 Risk assessment (ISO 27001 Clause 6.1)
  Integrate AI-specific risk and impact into your standard risk approach.
  A.5.7 Threat intelligence (partial)
  Factor new AI threats (e.g., data poisoning) into risk approach.
  A.5.25 Assessment & decision on InfoSec events (partial)
  Evaluate AI events in your existing process.
- A.5.3Documentation of AI system impact assessments
  Specific to AI
  Relates to
  Clause 6.1 Risk management documentation (ISO 27001 Clause 6.1)
  Store AI impact results in your risk register or risk docs.
- A.5.4Assessing AI system impact on individuals or groups
  Specific to AI
  Relates to
  A.5.34 Privacy & protection of PII
  Evaluate potential personal data usage, bias, fairness.
- A.5.5Assessing societal impacts of AI systems
  Specific to AI
  Relates to
  A.5.34 Privacy & protection of PII (partial)
  Goes beyond typical privacy to societal-level impacts, but some overlap remains.
- A.6.1.2Objectives for responsible development of AI
  Specific to AI
  Relates to
  A.5.4 Management responsibilities
  Leadership sets AI dev goals as part of broader management system.
- A.6.1.3Processes for responsible AI system design & development
  Carried over
  Relates to
  A.8.25 Secure development life cycle
  Integrate AI-lifecycle steps (data quality, bias checks) into secure DevOps.
  A.5.8 InfoSec in project management (partial)
  Incorporate AI-lifecycle steps in general project planning.
- A.6.2.2AI system requirements & specification
  Carried over
  Relates to
  A.8.26 Application security requirements
  Overlaps with specifying security needs for AI solutions.
  A.5.8 InfoSec in project management (partial)
  Document AI specs as part of the project scope.
  A.8.9 Configuration management (partial)
  Initial system specs feed into configuration management.
- A.6.2.3Documentation of AI system design & development
  Carried over
  Relates to
  A.5.37 Documented operating procedures
  Keep AI design/dev docs with your standard documented procedures.
  A.8.9 Configuration management (partial)
  Track changes to design docs as part of configuration management.
- A.6.2.4AI system verification & validation
  Carried over
  Relates to
  A.8.29 Security testing in development & acceptance
  Incorporate fairness, bias, and model vulnerability testing.
  A.8.28 Secure coding (partial)
  Overlaps with verifying code security and correctness.
- A.6.2.5AI system deployment
  Carried over
  Relates to
  A.8.19 Installation of software on operational systems
  Align AI deployment with standard release procedures.
  A.8.31 Separation of dev, test & production
  Keep AI dev/test separate from production.
- A.6.2.6AI system operation & monitoring
  Carried over
  Relates to
  A.8.16 Monitoring activities
  Monitor AI performance, detect drift, anomalies.
  A.8.32 Change management (partial)
  Oversee updates or retraining in production.
- A.6.2.7AI system technical documentation
  Carried over
  Relates to
  A.5.37 Documented operating procedures
  Maintain detailed AI system docs.
- A.6.2.8AI system recording of event logs
  Carried over
  Relates to
  A.8.15 Logging
  Log model inputs/outputs, performance, etc.
- A.7.2Data for development & enhancement of AI system
  Carried over
  Relates to
  A.5.9 Inventory of info & other assets
  Include training/validation datasets in the asset register.
  A.5.12 Classification of information (partial)
  Consider confidentiality or compliance constraints.
  A.8.25 Secure development life cycle (partial)
  Integrate data usage in dev/test phases.
- A.7.3Acquisition of data
  Carried over
  Relates to
  A.5.14 Information transfer
  Securely transfer external data for AI.
  A.5.19-A.5.22 Supplier mgmt (partial)
  If data is purchased or obtained from external suppliers.
- A.7.4Quality of data for AI systems
  Specific to AI
  Relates to
  A.5.12 Classification of information (partial)
  Data classification can incorporate data quality/bias checks.
  A.8.11 Data masking (partial)
  Use masking/anonymization to maintain data quality & privacy.
  A.8.12 Data leakage prevention (partial)
  Ensuring data integrity also helps prevent leakage or unauthorized use.
- A.7.5Data provenance
  Specific to AI
  Relates to
  A.5.9 Inventory of info & other assets
  Keep data lineage as part of your asset inventory approach.
  A.5.14 Information transfer (partial)
  Track origin/destination of datasets.
- A.7.6Data preparation
  Carried over
  Relates to
  A.8.10 Information deletion (partial)
  Removing/sanitizing data used for AI.
  A.8.33 Test information
  Overlaps in how test/training data is prepared or anonymized.
- A.8.2System documentation & information for users
  Carried over
  Relates to
  A.5.37 Documented operating procedures
  Provide AI usage documentation within standard operating procedures.
- A.8.3External reporting
  Carried over
  Relates to
  A.5.31 Legal, statutory, regulatory & contractual requirements
  If AI is regulated, you may need external disclosures.
- A.8.4Communication of incidents
  Carried over
  Relates to
  A.5.24 InfoSec incident mgmt planning & prep
  Integrate AI incidents with overall incident response.
  A.6.8 InfoSec event reporting
  Extend event reporting channels to AI issues.
- A.8.5Information for interested parties
  Carried over
  Relates to
  A.5.31 Legal, statutory, regulatory & contractual (partial)
  If laws require stakeholder transparency about AI.
- A.9.2Processes for responsible use of AI systems
  Carried over
  Relates to
  A.5.15 Access control
  Limit who can use or train AI models.
  A.5.18 Access rights
  Tie specific AI usage rights to user roles.
  A.8.2 Privileged access rights
  Overlaps in controlling privileged AI usage.
  A.8.3 Information access restriction
  Determine which data/models a user can access.
  A.8.18 Use of privileged utility programs
  Restrict advanced AI debugging or development tools.
- A.9.3Objectives for responsible use of AI systems
  Specific to AI
  Relates to
  A.5.4 Management responsibilities
  Management sets usage objectives (e.g., ethical usage).
- A.9.4Intended use of the AI system
  Specific to AI
  Relates to
  No direct ISO 27001 equivalent
  27001 does not require "intended use" documentation.
- A.10.2Allocating responsibilities
  Carried over
  Relates to
  A.5.2 Information security roles & responsibilities (partial)
  Clarify cross-organizational responsibilities for AI.
  A.5.31 Legal, statutory, regulatory & contractual requirements (partial)
  Include AI roles in contracts if needed.
- A.10.3Suppliers
  Carried over
  Relates to
  A.5.19 Information security in supplier relationships
  Oversee AI suppliers with your standard supplier management approach.
  A.5.20 Addressing InfoSec within supplier agreements
  Add AI-specific clauses in supplier contracts.
  A.5.21 Managing InfoSec in the ICT supply chain
  Extend supply chain security to AI model/data providers.
  A.5.22 Monitoring, review & change mgmt of supplier services
  Continual oversight of AI suppliers.
  A.8.30 Outsourced development (partial)
  If outsourcing AI development, apply the same supplier dev controls.
- A.10.4Customers
  Carried over
  Relates to
  A.5.19 Information security in supplier relationships (partial)
  Reverse relationship if you deliver AI solutions to clients.
  A.5.31 Legal, statutory, regulatory & contractual requirements (partial)
  Contracts for AI solutions to customers.

Each ISO 27001:2022 Annex A control and its place in an AI management system.

Relationship

ISO 27001 theme

Click to sort rows by Control header in ascending order	Click to sort rows by Relates to header in ascending order	Click to sort rows by Relationship header in ascending order
A.5.1 Policies for information security	A.2.2 AI Policy +2 more	Carried over
A.5.2 Information security roles & responsibilities	A.3.2 AI roles and responsibilities	Carried over
A.5.3 Segregation of duties	No distinct ISO 42001 equivalent	Security-only
A.5.4 Management responsibilities	A.2.3 Alignment with other organizational policies +2 more	Carried over
A.5.5 Contact with authorities	No distinct ISO 42001 equivalent	Security-only
A.5.6 Contact with special interest groups	No distinct ISO 42001 equivalent	Security-only
A.5.7 Threat intelligence	A.5.2 AI system impact assessment process	Carried over
A.5.8 Information security in project management	A.6.1.3 Processes for responsible AI system design & development +1 more	Carried over
A.5.9 Inventory of information & other associated assets	A.4.2 Resource documentation +1 more	Carried over
A.5.10 Acceptable use of information & other associated assets	No distinct ISO 42001 equivalent	Security-only
A.5.11 Return of assets	No distinct ISO 42001 equivalent	Security-only
A.5.12 Classification of information	A.7.4 Quality of data for AI systems	Carried over
A.5.13 Labelling of information	No distinct ISO 42001 equivalent	Security-only
A.5.14 Information transfer	A.4.3 Data resources +1 more	Carried over
A.5.15 Access control	A.9.2 Processes for responsible use of AI systems	Carried over
A.5.16 Identity management	No distinct ISO 42001 equivalent	Security-only
A.5.17 Authentication information	No distinct ISO 42001 equivalent	Security-only
A.5.18 Access rights	A.9.2 Processes for responsible use of AI systems	Carried over
A.5.19 Information security in supplier relationships	A.10.3 Suppliers	Carried over
A.5.20 Addressing InfoSec within supplier agreements	A.10.2 Allocating responsibilities +1 more	Carried over
A.5.21 Managing InfoSec in the ICT supply chain	A.10.3 Suppliers	Carried over
A.5.22 Monitoring, review & change mgmt of supplier services	A.10.3 Suppliers	Carried over
A.5.23 InfoSec for use of cloud services	A.4.5 System & computing resources	Carried over
A.5.24 InfoSec incident mgmt planning & preparation	A.8.4 Communication of incidents	Carried over
A.5.25 Assessment & decision on InfoSec events	A.5.2 AI system impact assessment process	Carried over

Items per page:

1–25 of 93 items

of 4 pages