architecture
Runtime AI Governance
ArchitectureTA-14 AI Governance Libraryfoundational
Summary
Governance applied while AI systems are executing rather than only before deployment.
Why It Matters
Execution-time governance enables admissibility checks, intervention, and outcome preservation.
Key Topics
runtime controlsexecutionmonitoringinterventionadmissibilityevidence
Related Records
TA-14 Admissible Execution
An evidence-bound governance architecture that determines whether a proposed action is admissible before execution and preserves the resulting outcome evidence.
AI Incident Response
Practices for identifying, triaging, containing, investigating, documenting, and resolving AI incidents.
Model Governance
Governance practices covering model approval, versioning, validation, monitoring, retirement, and accountability.