Set Up an AI Governance Committee: Charters & RACI Roles October 2026

A proposed 2025 California class-action lawsuit alleging that a health system used an AI scribe to record and transmit patient conversations without consent illustrates the governance risks that can arise when AI deployments lack clear oversight, although the claims have not been adjudicated. The recommended model distinguishes an AI governance committee, which has delegated authority to approve, suspend, or require remediation of AI systems, from a steering committee focused on strategy and investment priorities. It argues that effective committees should include legal, compliance, security, privacy, engineering, business leadership, and an executive sponsor with escalation or suspension authority, supported by a signed charter, defined quorum and recusal rules, a current AI-system inventory, and a RACI matrix assigning responsibility and accountability for decisions. The discussion links this structure to EU AI Act requirements, particularly deployer obligations to assign competent human oversight for high-risk systems, while noting that the law does not itself require a committee. It recommends documented workflows for risk classification, deployment approval, exceptions, policy review, production monitoring, incidents, and compliance evidence, along with metrics such as inventory coverage, approval speed, exception aging, remediation time, audit findings, and framework coverage. It also presents Openlayer as a platform for maintaining AI inventories, mapping controls to regulatory frameworks, testing systems, monitoring production behavior, enforcing guardrails on routed traffic, and creating request-level audit records.

Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.