Healthcare
Governed AI in Healthcare
Illustrative reference architecture, not a client implementation or measured case study.
Behavioral containment for clinical decision support and workload isolation for PHI-handling agents. The pattern supports safeguards relevant to HIPAA and FDA SaMD reviews; it does not by itself make a system compliant.
Enforcement Layers
L3Drift Guard
L4Gated Substrate
Enforcement Points
Behavioral containment for clinical decision support
Workload isolation for PHI-handling agents
Commercial Mapping
Evidence for HIPAA safeguard reviews
FDA SaMD regulatory alignment
Patient safety risk containment
Audit defensibility for clinical AI