NIST AI RMF：AI-FW 控制与组织责任
了解框架义务、AI-FW 技术控制和组织责任。
NIST AI RMF：AI-FW 控制与组织责任 · 自愿框架
本页说明框架与 AI-FW 的关系。产品支持部分运行时控制，但不替代组织的合规或审计责任。
AI-FW 的定位
- 说明：AI-FW provides visibility and enforcement for observed traffic: identity, content, destinations, decisions and risk signals. It does not replace governance or system-level evaluation.
产品不负责的事项
- 说明：Setting governance or risk appetite.
- 说明：Mapping every context, stakeholder and impact.
- 说明：Measuring every model property.
- 说明：Selecting treatment or accepting residual risk.
- 说明：Covering unobserved systems.
控制映射
- 说明：GOVERN: 说明：Define policy and oversight. AI-FW: 说明：Technical policies, roles and audit records.
- 说明：MAP: 说明：Understand systems, uses and impacts. AI-FW: 说明：Model inventory and identity for observed traffic.
- 说明：MEASURE: 说明：Analyse risks and incidents. AI-FW: 说明：Inspection, scoring and risk signals.
- 说明：MANAGE: 说明：Prioritise and respond to risks. AI-FW: 说明：Blocking, masking, audits and exports.
证据预期
- 说明：Coverage: 说明：Observed channels and models.
- 说明：Controls: 说明：Rules and destinations.
- 说明：Measures: 说明：Decisions, flags and metrics.
组织责任
- 说明：Define governance and accountability.
- 说明：Map systems, uses and impacts.
- 说明：Measure models and review bias.
- 说明：Validate risk mitigations.
- 说明：Improve the complete system.
常见问题
说明：Does AI-FW implement NIST AI RMF?
说明：It supports technical evidence within an organisational programme.
说明：Does a dashboard prove trustworthiness?
说明：No. It shows observed gateway metrics.
说明：Is NIST AI RMF a certification?
说明：No. It is voluntary.
一般信息，不构成法律或审计建议。