Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Ai Governance topic
No spam. Unsubscribe anytime.
Researchers, ethicists and unions urge state governance, monitoring and clearer liability for health‑care AI
Summary
Legal scholars, patient-rights advocates, clinicians and union representatives told lawmakers that market incentives alone will not produce safe AI deployments and urged the Assembly to require governance structures, standardized transparency and post‑deployment monitoring as licensure or procurement conditions.
Experts at the Assembly hearing pressed the committee to address a governance gap: many developers disclaim liability in contracts, and most health‑care purchasers have not stood up rigorous pre‑deployment validation or ongoing monitoring for AI tools.
Stanford researcher and lawyer Michelle Mello said market incentives encourage experiments but not the personnel and processes required for robust safety oversight. “The incentives are there for health care organizations to want to try things with AI. But when it comes to the difficult task of actually governing them to ensure safe and responsible use, there are not incentives sufficient for a lot of health care organizations,” she told the committee.
Mello recommended a state framework that requires health facilities to maintain governing bodies and review procedures for AI — modeled loosely on institutional review boards used for human‑subjects research — as part of licensure or certification. She said the goal is to push hospitals to measure safety‑oriented outcomes, not just whether software is running.
Patient‑rights and civil‑society panelists asked for stronger, enforceable patient protections: notice when AI affects a care decision, patient participation in governance, clear pathways for legal recourse when errors cause harm, and privacy safeguards against re‑identification. Christine Von Riesfeld of the Light Collective proposed a seven‑point Patient AI Rights framework to codify these protections.
Labor representatives also urged stronger worker protections. Chris Nielsen of the California Nurses Association warned that employers could rely on AI to justify understaffing and called for workers to have authority to override tools and veto deployments that undermine safety.
Industry and coalition representatives offered practical complements: standardized model cards or “fact labels” (Championed by the Coalition for Health AI), a public registry of disclosures, and networks of vetted third‑party evaluators to help hospitals with pre‑ and post‑deployment assessment. Panelists agreed the state could accelerate accountability by pairing disclosure mandates with funding and technical assistance to help under‑resourced providers comply.
