Back to the AI Control Library
36 controls | 18 core | 18 supplemental
Why this family matters
AI risk begins with the information a system learns from, retrieves, processes and retains. This family addresses how that information is identified, classified, protected and governed throughout its lifecycle. It connects data ownership and privacy expectations to the technical controls that govern access, movement, retention and disposal.
What this family covers
The controls cover inventories, data flows, classification, privacy assessments, sensitive information, residency, poisoning prevention, encryption and sub-processing. Supplemental controls provide additional depth in encryption and key management, production-data use and disclosure processes.
How the four control types apply
Operational controls govern data ownership, permitted use, retention and disposal. Technical controls enforce encryption, protection and deletion. Governance controls establish applicable external privacy and data obligations. Contextual constraints arise from the purpose, setting and expectations surrounding the information being used.
Where to begin
Trace the data used by each system, identify its owner and sensitivity, and define permitted uses and retention before expanding access.
Controls in this family
This family contains 36 controls: 18 core and 18 supplemental. Each page provides requirements, implementation steps, verification criteria, evidence, escalation and a worked example. Applicability depends on the system, information and risks involved.
Core controls
- CF05-001: Data Security & Privacy Policy for AI
- CF05-002: AI Data Inventory
- CF05-003: AI Data Classification
- CF05-004: AI Data Flow Documentation
- CF05-005: AI Data Ownership & Stewardship
- CF05-006: AI Data Protection by Design & Default
- CF05-007: AI Data Protection Impact Assessment (DPIA)
- CF05-008: AI Sensitive Data Transfer Controls
- CF05-009: AI Data Subject Rights Implementation
- CF05-010: AI Data Purpose Limitation
- CF05-011: AI Data Retention & Deletion
- CF05-012: AI Sensitive Data Protection
- CF05-013: AI Data Location & Residency Controls
- CF05-014: AI Training Data Poisoning Prevention & Detection
- CF05-015: Privacy-Enhancing Technologies for AI
- CF05-016: AI Data Secure Disposal
- CF05-017: AI Data Encryption at Rest & in Transit
- CF05-018: AI Data Sub-processing Controls
Supplemental controls
- CF05-S01: AI Encryption & Key Management Policy
- CF05-S02: AI Encryption Roles & Responsibilities
- CF05-S03: AI Encryption Change Management
- CF05-S04: Encryption Change Cost-Benefit Analysis
- CF05-S05: AI Encryption Risk Management
- CF05-S06: Customer-Managed Encryption Keys for AI
- CF05-S07: AI Encryption Key Generation
- CF05-S08: AI Encryption Key Purpose Restriction
- CF05-S09: AI Encryption Key Rotation
- CF05-S10: AI Encryption Key Revocation
- CF05-S11: AI Encryption Key Destruction
- CF05-S12: AI Encryption Key Lifecycle Management (Activation/Suspension/Deactivation)
- CF05-S13: AI Encryption Key Archival
- CF05-S14: AI Encryption Key Compromise Response
- CF05-S15: AI Encryption Key Recovery
- CF05-S16: AI Encryption Key Inventory Management
- CF05-S17: Production AI Data Use Limitation
- CF05-S18: AI Personal Data Law Enforcement Disclosure
Applying this family
Begin with data ownership, classification and flow visibility. Connect privacy and use decisions to practical controls on collection, retrieval, transfer, retention and disposal. Include copies created through logs, indexes, embeddings, backups and suppliers where relevant. Encryption and key management support these decisions, but do not by themselves establish that a use is permitted or appropriate.
Policy and assessment controls define permitted handling. Technical controls enforce access, location, encryption and deletion where feasible. Customer-managed keys, cryptographic lifecycle records and provider arrangements must be evaluated together with recovery needs and actual service capabilities.
Version 1.0
