+AI Model Resilience
---+AI Model Pollution
---+Cascading Hallucination Defense
---+Resource Exhaustion & Denial of Service (DoS) Resilience
|
AI Model Resilience
Description
Mechanisms exist to ensure AI models are designed with resilience capabilities that are sufficient to withstand reasonable threats.
Possible Solutions & Considerations
Micro-Small Business (<10 staff) / BLS Firm Size Classes 1-2
∙ Choose AI tools from vendors with documented resilience capabilities
Small Business (10-49 staff) / BLS Firm Size Classes 3-4
∙ Resilience requirements in AI tool selection criteria
Medium Business (50-249 staff) / BLS Firm Size Classes 5-6
∙ Formal AI resilience requirements in development/procurement standards
Large Business (250-999 staff) / BLS Firm Size Classes 7-8
∙ AI resilience testing program
∙ Vendor resilience requirements
Enterprise (> 1,000 staff) / BLS Firm Size Class 9
∙ Enterprise AI resilience framework
∙ Adversarial robustness testing
∙ Redundancy and failover for AI systems
SCR-CMM
Level 0 Not Performed
Practices are non-existent, based on the inability to demonstrate an implemented and operational capability. A reasonable person would conclude the control is not being performed.
Level 1 Performed Informally
SCR-CMM Level 1 criteria definitions are not available for this control:
▪ A reasonable person would conclude this control requires a structured process.
▪ At this level of maturity, the "ad hoc" nature of performing a capability informally would indicate the intent of the control is not met due to a lack of consistency and formality.
Level 2 Planned Tracked
Artificial Intelligence and Autonomous Technology (AAT) capabilities are requirements-driven, but are not standardized across the entity (e.g., local/regional level consistency). Capability criteria associated with this control reasonably expect the following criteria to exist:
▪ Policies and standards associated with AAT domain capabilities are formally documented and centrally-managed by the entity.
▪ Standardized Operating Procedures (SOP) associated with AAT domain capabilities are documented and maintained by process owners.
▪ IT and/or cybersecurity personnel work with business stakeholders and process owners to appropriately scope and reasonably implement cybersecurity and data protection controls associated with AAT domain capabilities to address applicable statutory, regulatory and/or contractual requirements for Technology Assets, Applications, Services and/or Data (TAASD).
▪ Artificial Intelligence (AI)-related controls are primarily administrative and preventative in nature (e.g., policies, standards and procedures & guidelines).
▪ Asset management may be a defined function (e.g., team or department) or assigned as an additional duty to existing IT and/or cybersecurity personnel.
Level 3 Well Defined
Artificial Intelligence and Autonomous Technology (AAT) capabilities are standardized across the entity for applicability to People, Processes, Technologies, Data and/or Facilities (PPTDF) to ensure consistency for Technology Assets, Applications, Services and/or Data (TAASD). Capability criteria associated with this control reasonably expect the following criteria to exist:
▪ Policies and standards associated with AAT domain capabilities are formally documented and centrally-managed by the entity's Governance, Risk & Compliance (GRC) team, or similar function.
▪ Standardized Operating Procedures (SOP) associated with AAT domain capabilities are well-documented and kept current by process owners.
▪ An Artificial Intelligence Governance (AIG) team, or similar function, is appropriately staffed and supported to implement and maintain AAT domain capabilities.
▪ Technology is leveraged to enhance the efficiency and accuracy of AI governance, risk management and compliance operations (e.g., dedicated AI governance platform).
▪ The entity's Governance, Risk & Compliance (GRC) team, or similar function, works with business stakeholders and process owners to appropriately scope and reasonably implement cybersecurity and data protection controls associated with AAT domain capabilities to address Minimum Compliance Requirements (MCR) (e.g., applicable statutory, regulatory and/or contractual requirements) and Discretionary Security Requirements (DSR) (e.g., entity-required controls).
▪ An implemented and operational capability exists to ensure AI models are designed with resilience capabilities that are sufficient to withstand reasonable threats.
Level 4 Quantitatively Controlled
Utilize SCR-CMM Level 3 criteria definitions:
▪ There are no defined Level 4 criteria, since it is reasonable to assume a quantitatively-controlled process is not necessary to operationalize this control.
▪ While it may be possible to develop “metrics-driven” capabilities for this control, the criteria would be organization-specific to define.
Level 5 Continuously Improving
Utilize SCR-CMM Level 3 or Level 4 (if available) criteria definitions:
▪ There are no defined Level 5 criteria, since it is reasonable to assume a continuously-improving process is not necessary to operationalize this control.
▪ Level 5 capabilities should be considered “world-class” where the control builds on Level 4 capabilities, but are continuously improving through Artificial Intelligence (AI) and/or Machine Learning (ML) technologies.
▪ While it may be possible to develop responsive capabilities for this control through the use of AI and/or ML technologies, the criteria would be organization-specific to define.
1. Översikt
| Rubrik |
Standard |
|
AI Model Pollution
|
Description
Mechanisms exist to prevent "model pollution" due to accidental and/or malicious inputs by an AI agent that can negatively alter the AI model.
Possible Solutions & Considerations
Micro-Small Business (<10 staff) / BLS Firm Size Classes 1-2
∙ Restrict who can provide inputs to AI models
Small Business (10-49 staff) / BLS Firm Size Classes 3-4
∙ Access controls and input validation for AI models
Medium Business (50-249 staff) / BLS Firm Size Classes 5-6
∙ Formal model pollution prevention policy
∙ Input validation controls
Large Business (250-999 staff) / BLS Firm Size Classes 7-8
∙ AI model integrity program
∙ Input validation, sanitization, and access controls
Enterprise (> 1,000 staff) / BLS Firm Size Class 9
∙ Enterprise AI model integrity framework
∙ Differential privacy techniques
∙ Input validation and anomaly detection
SCR-CMM
Level 0 Not Performed
Practices are non-existent, based on the inability to demonstrate an implemented and operational capability. A reasonable person would conclude the control is not being performed.
Level 1 Performed Informally
SCR-CMM Level 1 criteria definitions are not available for this control:
▪ A reasonable person would conclude this control requires a structured process.
▪ At this level of maturity, the "ad hoc" nature of performing a capability informally would indicate the intent of the control is not met due to a lack of consistency and formality.
Level 2 Planned Tracked
Artificial Intelligence and Autonomous Technology (AAT) capabilities are requirements-driven, but are not standardized across the entity (e.g., local/regional level consistency). Capability criteria associated with this control reasonably expect the following criteria to exist:
▪ Policies and standards associated with AAT domain capabilities are formally documented and centrally-managed by the entity.
▪ Standardized Operating Procedures (SOP) associated with AAT domain capabilities are documented and maintained by process owners.
▪ IT and/or cybersecurity personnel work with business stakeholders and process owners to appropriately scope and reasonably implement cybersecurity and data protection controls associated with AAT domain capabilities to address applicable statutory, regulatory and/or contractual requirements for Technology Assets, Applications, Services and/or Data (TAASD).
▪ Artificial Intelligence (AI)-related controls are primarily administrative and preventative in nature (e.g., policies, standards and procedures & guidelines).
▪ Asset management may be a defined function (e.g., team or department) or assigned as an additional duty to existing IT and/or cybersecurity personnel.
Level 3 Well Defined
Artificial Intelligence and Autonomous Technology (AAT) capabilities are standardized across the entity for applicability to People, Processes, Technologies, Data and/or Facilities (PPTDF) to ensure consistency for Technology Assets, Applications, Services and/or Data (TAASD). Capability criteria associated with this control reasonably expect the following criteria to exist:
▪ Policies and standards associated with AAT domain capabilities are formally documented and centrally-managed by the entity's Governance, Risk & Compliance (GRC) team, or similar function.
▪ Standardized Operating Procedures (SOP) associated with AAT domain capabilities are well-documented and kept current by process owners.
▪ An Artificial Intelligence Governance (AIG) team, or similar function, is appropriately staffed and supported to implement and maintain AAT domain capabilities.
▪ Technology is leveraged to enhance the efficiency and accuracy of AI governance, risk management and compliance operations (e.g., dedicated AI governance platform).
▪ The entity's Governance, Risk & Compliance (GRC) team, or similar function, works with business stakeholders and process owners to appropriately scope and reasonably implement cybersecurity and data protection controls associated with AAT domain capabilities to address Minimum Compliance Requirements (MCR) (e.g., applicable statutory, regulatory and/or contractual requirements) and Discretionary Security Requirements (DSR) (e.g., entity-required controls).
▪ An implemented and operational capability exists to prevent "model pollution" due to accidental and/or malicious inputs by an AI agent that can negatively alter the AI model.
Level 4 Quantitatively Controlled
Utilize SCR-CMM Level 3 criteria definitions:
▪ There are no defined Level 4 criteria, since it is reasonable to assume a quantitatively-controlled process is not necessary to operationalize this control.
▪ While it may be possible to develop “metrics-driven” capabilities for this control, the criteria would be organization-specific to define.
Level 5 Continuously Improving
Utilize SCR-CMM Level 3 or Level 4 (if available) criteria definitions:
▪ There are no defined Level 5 criteria, since it is reasonable to assume a continuously-improving process is not necessary to operationalize this control.
▪ Level 5 capabilities should be considered “world-class” where the control builds on Level 4 capabilities, but are continuously improving through Artificial Intelligence (AI) and/or Machine Learning (ML) technologies.
▪ While it may be possible to develop responsive capabilities for this control through the use of AI and/or ML technologies, the criteria would be organization-specific to define.
|
|
Cascading Hallucination Defense
|
Description
Mechanisms exist to detect and prevent the propagation of false data (e.g., hallucinations) within the AI model or between AI agents.
Possible Solutions & Considerations
Micro-Small Business (<10 staff) / BLS Firm Size Classes 1-2
∙ Review AI outputs for signs of hallucination before use
Small Business (10-49 staff) / BLS Firm Size Classes 3-4
∙ Procedures to detect and prevent AI hallucination propagation
Medium Business (50-249 staff) / BLS Firm Size Classes 5-6
∙ Formal hallucination detection and prevention policy
Large Business (250-999 staff) / BLS Firm Size Classes 7-8
∙ AI hallucination detection program
∙ Technical controls to isolate false data propagation
Enterprise (> 1,000 staff) / BLS Firm Size Class 9
∙ Enterprise AI hallucination prevention framework
∙ RAG architecture
∙ Automated hallucination detection tools
SCR-CMM
Level 0 Not Performed
Practices are non-existent, based on the inability to demonstrate an implemented and operational capability. A reasonable person would conclude the control is not being performed.
Level 1 Performed Informally
SCR-CMM Level 1 criteria definitions are not available for this control:
▪ A reasonable person would conclude this control requires a structured process.
▪ At this level of maturity, the "ad hoc" nature of performing a capability informally would indicate the intent of the control is not met due to a lack of consistency and formality.
Level 2 Planned Tracked
Artificial Intelligence and Autonomous Technology (AAT) capabilities are requirements-driven, but are not standardized across the entity (e.g., local/regional level consistency). Capability criteria associated with this control reasonably expect the following criteria to exist:
▪ Policies and standards associated with AAT domain capabilities are formally documented and centrally-managed by the entity.
▪ Standardized Operating Procedures (SOP) associated with AAT domain capabilities are documented and maintained by process owners.
▪ IT and/or cybersecurity personnel work with business stakeholders and process owners to appropriately scope and reasonably implement cybersecurity and data protection controls associated with AAT domain capabilities to address applicable statutory, regulatory and/or contractual requirements for Technology Assets, Applications, Services and/or Data (TAASD).
▪ Artificial Intelligence (AI)-related controls are primarily administrative and preventative in nature (e.g., policies, standards and procedures & guidelines).
▪ Asset management may be a defined function (e.g., team or department) or assigned as an additional duty to existing IT and/or cybersecurity personnel.
Level 3 Well Defined
Artificial Intelligence and Autonomous Technology (AAT) capabilities are standardized across the entity for applicability to People, Processes, Technologies, Data and/or Facilities (PPTDF) to ensure consistency for Technology Assets, Applications, Services and/or Data (TAASD). Capability criteria associated with this control reasonably expect the following criteria to exist:
▪ Policies and standards associated with AAT domain capabilities are formally documented and centrally-managed by the entity's Governance, Risk & Compliance (GRC) team, or similar function.
▪ Standardized Operating Procedures (SOP) associated with AAT domain capabilities are well-documented and kept current by process owners.
▪ An Artificial Intelligence Governance (AIG) team, or similar function, is appropriately staffed and supported to implement and maintain AAT domain capabilities.
▪ Technology is leveraged to enhance the efficiency and accuracy of AI governance, risk management and compliance operations (e.g., dedicated AI governance platform).
▪ The entity's Governance, Risk & Compliance (GRC) team, or similar function, works with business stakeholders and process owners to appropriately scope and reasonably implement cybersecurity and data protection controls associated with AAT domain capabilities to address Minimum Compliance Requirements (MCR) (e.g., applicable statutory, regulatory and/or contractual requirements) and Discretionary Security Requirements (DSR) (e.g., entity-required controls).
▪ An implemented and operational capability exists to detect and prevent the propagation of false data (e.g., hallucinations) within the AI model or between AI agents.
Level 4 Quantitatively Controlled
Utilize SCR-CMM Level 3 criteria definitions:
▪ There are no defined Level 4 criteria, since it is reasonable to assume a quantitatively-controlled process is not necessary to operationalize this control.
▪ While it may be possible to develop “metrics-driven” capabilities for this control, the criteria would be organization-specific to define.
Level 5 Continuously Improving
Utilize SCR-CMM Level 3 or Level 4 (if available) criteria definitions:
▪ There are no defined Level 5 criteria, since it is reasonable to assume a continuously-improving process is not necessary to operationalize this control.
▪ Level 5 capabilities should be considered “world-class” where the control builds on Level 4 capabilities, but are continuously improving through Artificial Intelligence (AI) and/or Machine Learning (ML) technologies.
▪ While it may be possible to develop responsive capabilities for this control through the use of AI and/or ML technologies, the criteria would be organization-specific to define.
|
|
Resource Exhaustion & Denial of Service (DoS) Resilience
|
Description
Mechanisms exist to monitor and prevent resource overload or Denial of Service (DoS) conditions through:
(1) Enforcement of quotas;
(2) Workload controls; and
(3) Auto-suspension of runaway AI agent processes.
Possible Solutions & Considerations
Micro-Small Business (<10 staff) / BLS Firm Size Classes 1-2
∙ Monitor AI tool resource usage and set limits
Small Business (10-49 staff) / BLS Firm Size Classes 3-4
∙ Resource quotas for AI tool usage
∙ Monitor for unusual resource consumption
Medium Business (50-249 staff) / BLS Firm Size Classes 5-6
∙ Formal AI resource management policy
∙ Quotas, workload controls, and monitoring
Large Business (250-999 staff) / BLS Firm Size Classes 7-8
∙ AI resource governance program
∙ Automated monitoring and auto-suspension capabilities
Enterprise (> 1,000 staff) / BLS Firm Size Class 9
∙ Enterprise AI resource management platform
∙ Automated quota enforcement
∙ Auto-scaling and auto-suspension
SCR-CMM
Level 0 Not Performed
Practices are non-existent, based on the inability to demonstrate an implemented and operational capability. A reasonable person would conclude the control is not being performed.
Level 1 Performed Informally
SCR-CMM Level 1 criteria definitions are not available for this control:
▪ A reasonable person would conclude this control requires a structured process.
▪ At this level of maturity, the "ad hoc" nature of performing a capability informally would indicate the intent of the control is not met due to a lack of consistency and formality.
Level 2 Planned Tracked
Artificial Intelligence and Autonomous Technology (AAT) capabilities are requirements-driven, but are not standardized across the entity (e.g., local/regional level consistency). Capability criteria associated with this control reasonably expect the following criteria to exist:
▪ Policies and standards associated with AAT domain capabilities are formally documented and centrally-managed by the entity.
▪ Standardized Operating Procedures (SOP) associated with AAT domain capabilities are documented and maintained by process owners.
▪ IT and/or cybersecurity personnel work with business stakeholders and process owners to appropriately scope and reasonably implement cybersecurity and data protection controls associated with AAT domain capabilities to address applicable statutory, regulatory and/or contractual requirements for Technology Assets, Applications, Services and/or Data (TAASD).
▪ Artificial Intelligence (AI)-related controls are primarily administrative and preventative in nature (e.g., policies, standards and procedures & guidelines).
▪ Asset management may be a defined function (e.g., team or department) or assigned as an additional duty to existing IT and/or cybersecurity personnel.
Level 3 Well Defined
Artificial Intelligence and Autonomous Technology (AAT) capabilities are standardized across the entity for applicability to People, Processes, Technologies, Data and/or Facilities (PPTDF) to ensure consistency for Technology Assets, Applications, Services and/or Data (TAASD). Capability criteria associated with this control reasonably expect the following criteria to exist:
▪ Policies and standards associated with AAT domain capabilities are formally documented and centrally-managed by the entity's Governance, Risk & Compliance (GRC) team, or similar function.
▪ Standardized Operating Procedures (SOP) associated with AAT domain capabilities are well-documented and kept current by process owners.
▪ An Artificial Intelligence Governance (AIG) team, or similar function, is appropriately staffed and supported to implement and maintain AAT domain capabilities.
▪ Technology is leveraged to enhance the efficiency and accuracy of AI governance, risk management and compliance operations (e.g., dedicated AI governance platform).
▪ The entity's Governance, Risk & Compliance (GRC) team, or similar function, works with business stakeholders and process owners to appropriately scope and reasonably implement cybersecurity and data protection controls associated with AAT domain capabilities to address Minimum Compliance Requirements (MCR) (e.g., applicable statutory, regulatory and/or contractual requirements) and Discretionary Security Requirements (DSR) (e.g., entity-required controls).
▪ An implemented and operational capability exists to monitor and prevent resource overload or Denial of Service (DoS) conditions through:
(1) Enforcement of quotas;
(2) Workload controls; and
(3) Auto-suspension of runaway AI agent processes.
Level 4 Quantitatively Controlled
Utilize SCR-CMM Level 3 criteria definitions:
▪ There are no defined Level 4 criteria, since it is reasonable to assume a quantitatively-controlled process is not necessary to operationalize this control.
▪ While it may be possible to develop “metrics-driven” capabilities for this control, the criteria would be organization-specific to define.
Level 5 Continuously Improving
Utilize SCR-CMM Level 3 or Level 4 (if available) criteria definitions:
▪ There are no defined Level 5 criteria, since it is reasonable to assume a continuously-improving process is not necessary to operationalize this control.
▪ Level 5 capabilities should be considered “world-class” where the control builds on Level 4 capabilities, but are continuously improving through Artificial Intelligence (AI) and/or Machine Learning (ML) technologies.
▪ While it may be possible to develop responsive capabilities for this control through the use of AI and/or ML technologies, the criteria would be organization-specific to define.
|
1.1 Referenser
1.2 Identifierade krav
1.3 Relaterade regleringar
2. Identifierade krav
3. Relaterade regleringar
Regleringar
| Källa |
Reglering |
Länkade ärenden
- Secure Controls Framework -
EN: GRCA includes the Secure Controls Framework® (SCF) content, © Secure Controls Framework Council, LLC, used under the Creative Commons Attribution-NoDerivatives 4.0 International Public License. SCF control numbers, names, descriptions and mappings appear exactly as published by the SCF Council – https://securecontrolsframework.com. GRCA may group or reorder SCF content for display, which does not change SCF content; the mappings are shown as links to the frameworks they name. Learn more at https://securecontrolsframework.com.
DE: GRCA enthält Inhalte des Secure Controls Framework® (SCF), © Secure Controls Framework Council, LLC, genutzt unter der Creative Commons Attribution-NoDerivatives 4.0 International Public License. Nummern, Namen, Beschreibungen und Zuordnungen der SCF-Controls erscheinen genau so, wie der SCF Council sie veröffentlicht – https://securecontrolsframework.com. GRCA gruppiert oder sortiert SCF-Inhalte für die Anzeige, ohne sie zu verändern; die Zuordnungen erscheinen als Verknüpfungen zu den genannten Regelwerken. Mehr unter https://securecontrolsframework.com.
|