Managing Health AI Governance

Structures AI/ML governance for healthcare applications with inventory management, risk classification, validation requirements, bias assessment, regulatory compliance, and ongoing monitoring frameworks. This skill covers the organizational governance structure needed to safely adopt, deploy, and monitor AI/ML tools across clinical, operational, and administrative healthcare functions.

Why This Skill Exists

Healthcare organizations are rapidly deploying AI/ML tools: EHR-embedded predictions (sepsis, deterioration, readmission), imaging AI (radiology CAD, pathology), ambient clinical documentation, coding/billing AI, operational AI (scheduling, staffing, capacity), and generative AI (clinical summarization, patient communication). Without governance, organizations face: patient safety risks from unvalidated models, health equity harms from biased algorithms, regulatory exposure (FDA SaMD, state AI transparency laws, CMS AI requirements), legal liability for AI-influenced clinical decisions, and reputational damage from publicized AI failures. Governance provides the organizational structure to manage these risks while enabling beneficial AI adoption.

High-profile AI failures in healthcare --- the Optum algorithm that systematically under-referred Black patients, Epic sepsis model performance questions, and concerns about ambient documentation hallucinations --- demonstrate that AI governance is not theoretical. President Biden's Executive Order 14110 on AI Safety (2023) and emerging state AI laws create regulatory obligations. The Coalition for Health AI (CHAI) and the AAMC are developing healthcare-specific AI governance standards that will shape expectations.

Checkpoint A --- Intake & Scoping

Answer every question before proceeding. Mark unknowns with [VERIFY].