Senate considers bill to protect clinicians who override AI recommendations in patient care

The Senate considered AB 2575, a proposal that would establish protections and transparency requirements when health care facilities deploy AI tools in direct patient care.

Senator Padilla presented the bill on behalf of Assemblymember Ortega, arguing that AI should support clinical judgment but not replace it. On the floor he outlined three core provisions: prohibit retaliation against clinicians who exercise professional judgment to follow or override an AI recommendation; require facilities to provide information about risks and intended uses of AI tools; and prevent developers or employers from shifting liability onto clinicians when AI errors harm patients.

Supporters, including Senator Gonzales, said the bill protects patient safety and clinician autonomy as AI becomes more common in hospitals. The sponsor said the author narrowed the bill in amendment rounds to limit covered tools and avoid disrupting care teams. Debate acknowledged both the growing use of AI in clinical settings and concerns about liability and transparency.

Floor action moved to a roll call; members signaled support and the measure advanced for the next steps. Implementation and the scope of tools covered were noted as follow‑up items for oversight and rulemaking.