SB 503: Health care artificial intelligence: bias in clinical decision-making and resource allocation

Latest recorded action

August 30, 2026. Enrolled and presented to the Governor at 6:00 PM.

What the bill does

Addresses artificial intelligence in health care services: identification of AI systems used for clinical decision-making or health-care resource allocation that carry known or reasonably foreseeable risks of biased outputs, with duties described in Senate floor material as falling on developers and deployers, including mitigation, monitoring and reporting.

What it does not do

Not law. The enrolled text has NOT been read at this snapshot: the developer and deployer duties, the definition of clinical-decision AI and resource allocation, the treatment of protected populations, and local deployment responsibility are stated here only as the floor material described them. Exact analysis is withheld until the enrolled text is read.

KPSGILL position

No position taken, and no duty-by-duty analysis is published yet — the enrolled text has not been read. When it is, it will be assessed against the KPSGILL AI Governance Readiness Framework, which treats developer and deployer accountability, bias monitoring and model change control as separate dimensions.

Open verification items

Named here rather than answered with an assumption.

• Enrolled text of SB 503 — full read required before any duty-by-duty analysis
• Governor action by 30 September

What happens next

Governor action deadline: September 30, 2026, computed from this bill’s own presentment rather than applied mechanically. Every qualifying measure is tracked on the Governor’s Health Bill Desk; scheduled dates sit on the decision calendar.

Primary sources

• California Senate — floor analysis material, SB 503

Register: data/ca-bills.json · internal identity CA|2025-2026|SB503 · verification state VERIFIED · tracker: California Legislature 2026.