AICM AtlasCSA AI Controls Matrix
CEK · Cryptography, Encryption & Key Management
CEK-11Cloud & AI Related

Key Purpose

Specification

Manage cryptographic secret and private keys that are provisioned for a unique purpose.

Threat coverage

Model manipulation
Data poisoning
Sensitive data disclosure
Model theft
Model/Service Failure
Insecure supply chain
Insecure apps/plugins
Denial of Service
Loss of governance

Architectural relevance

Physical infrastructure
Network
Compute
Storage
Application
Data

Lifecycle

Preparation

Data storage

Development

Guardrails

Evaluation

Not applicable

Deployment

Orchestration, AI Services supply chain, AI applications

Delivery

Operations, Maintenance, Continuous monitoring, Continuous improvement

Retirement

Archiving, Data deletion, Model disposal

Ownership / SSRM

PI

Shared Cloud Service Provider-Model Provider (Shared CSP-MP)

The CSP and MP are jointly responsible and accountable for the design, development, implementation, and enforcement of the control to mitigate security, privacy, or compliance risks associated with Large Language Model (LLM)/GenAI technologies in the context of the services or products they develop and offer.

Model

Owned by the Model Provider (MP)

The model provider (MP) designs, develops, and implements the control as part of their services or products to mitigate security, privacy, or compliance risks associated with the Large Language Model (LLM). Model Providers are entities that develop, train, and distribute foundational and fine-tuned AI models for various applications. They create the underlying AI capabilities that other actors build upon. Model Providers are responsible for model architecture, training methodologies, performance characteristics, and documentation of capabilities and limitations. They operate at the foundation layer of the AI stack and may provide direct API access to their models. Examples: OpenAI (GPT, DALL-E, Whisper), Anthropic(Claude), Google(Gemini), Meta(Llama), as well as any customized model.

Orchestrated

Shared Model Provider-Orchestrated Service Provider (Shared MP-OSP)

The MP and OSP are jointly responsible and accountable for the design, development, implementation, and enforcement of the control to mitigate security, privacy, or compliance risks associated with Large Language Model (LLM)/GenAI technologies in the context of the services or products they develop and offer.

Application

Shared Orchestrated Service Provider-Application Provider (Shared OSP-AP)

The OSP and AP are jointly responsible and accountable for the design, development, implementation, and enforcement of the control to mitigate security, privacy, or compliance risks associated with Large Language Model (LLM)/GenAI technologies in the context of the services or products they develop and offer.

Implementation guidelines

[All Actors]
Applies to all Roles (Baseline) before application of role context.
1. Establish and document policies and procedures for Cryptography, Encryption, and Key Management.

2. Approve the policies and procedures through formal governance processes (e.g., security committee, CISO).

3. Communicate the policies and procedures to all relevant stakeholders.

4. Apply the approved policies and procedures to all systems, services, and processes under the role’s control.

5. Evaluate the effectiveness of policy and procedure implementation using internal audits, technical reviews, 
and encryption control validations.

6. Review and update the policies and procedures at least annually, or when significant system, model, or 
regulatory changes occur.

Auditing guidelines

1. Verify that the CSP assigns cryptographic keys and secrets (e.g., tenant keys, HSM-stored credentials, API tokens) to a unique purpose (e.g., encryption, signing, authentication) and that purpose separation is enforced across cloud services, secret management systems, and cryptographic APIs.

2. Confirm that each key and secret is mapped to its intended function (e.g., volume encryption, TLS, identity verification) and that this mapping is documented, reviewed periodically, and integrated into key management systems.

3. Verify that technical and procedural controls prevent the reuse of a single key or secret for multiple cryptographic purposes across different cloud services or layers.

4. Review cryptographic service configurations (e.g., KMS, HSM policies) to ensure that key and secret usage is restricted to their assigned purpose within compute, storage, or networking scopes.

5. Confirm that access to purpose-bound keys and secrets is limited to authorized personnel, services, or systems based on their designated function and aligned with the principle of least privilege.

6. Validate that secrets and keys used in AI-related services (e.g., encrypted inference logs, signed outputs, model container authentication) are provisioned for distinct functions and not shared across services or tenants.

7. Review whether key and secret metadata includes attributes (e.g., tags, labels, purpose descriptors) that specify intended use and that enforcement of these attributes is implemented across cloud key management and monitoring tools.

8. Confirm that infrastructure-as-code (IaC), orchestration templates, and cloud deployment scripts enforce key and secret purpose separation, rejecting configurations that apply multi-use materials.

9. Verify that logging and audit mechanisms capture purpose-related usage metadata and detect any misuse of keys or secrets outside their defined function, with alerting or remediation triggers in place.

10. Validate that keys or secrets exposed to upstream or downstream parties (e.g., APs, AICs) are purpose-scoped and contractually or technically restricted from being reused for unrelated tasks or services.

From CCM:

1. Obtain copies of the policy and procedures detailing the management of secret and private cryptographic keys.
2. Identify cryptographic secret and private keys that have been provisioned for a unique purpose.
3. Ascertain that these keys are being managed in accordance with policy and procedures.

Standards mappings

ISO 42001Partial Gap
No Mapping for ISO 42001
ISO 27001: A.8.24
ISO 27002: 8.24
Addendum

ISO 42001 does not mandate managing keys with a requirement for unique provisioning per purpose, a critical detail for secure AI operations. Add a control for this requirement, ensuring each key is assigned to a specific use case with documented policies and procedures, addressing the gap in ISO 42001:2023’s lack of specific key management requirements. Include guidance on key segregation and purpose-specific provisioning, enhancing ISO 27001 (A.8.24) and ISO 27002 (8.24) for AI-specific cryptographic integrity.

EU AI ActFull Gap
No Mapping
Addendum

Cover the AICM control.

NIST AI 600-1Full Gap
No Mapping
Addendum

No (implicit/explicit) reference to cryptography, encryption, or key management is made in the NIST AI 600-1 standard, let alone to the requirement of managing cryptographic keys in line with their set purpose.

BSI AIC4No Gap
CRY-04
Addendum

N/A

AI-CAIQ questions (1)

CEK-11.1

Are cryptographic secrets and private keys that are provisioned for a unique purpose properly managed?