Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Health Ai topic
No spam. Unsubscribe anytime.
Committee advances SB 503 to require bias identification and monitoring for AI used in clinical decision making
Summary
SB 503 would require specified AI tools used for clinical decision support or health-resource allocation in health care facilities to be identified, mitigated for biased impacts, and regularly monitored. The committee voted the bill out to privacy and consumer protection with amendments.
SB 503, authored by Senator Weber Pearson and described to the Assembly Health Committee on July 8, would require developers and deployers of certain artificial intelligence tools used in clinical decision-making or health-care resource allocation to take steps to identify, mitigate and monitor biased impacts.
"AI tools do not operate in a vacuum," Weber Pearson told the committee, adding that bias can be embedded at each step of tool development and deployment and that bias in health care algorithms can worsen existing disparities. Physician and organizational witnesses said SB 503 takes a practical first step to build safeguards while sustaining innovation.
Dr. Steven Wang, a radiologist with the Southern California Permanente Medical Group and chair of the California Medical Association’s health IT subcommittee, testified in support and described his employer's efforts to adopt and monitor AI tools responsibly. "SB 503 is a measured and practical first approach in ensuring AI tools are safe and reliable," he said. George Sorey of the California Medical Association also urged support, saying physicians need transparent evaluation and accountability when AI tools are used.
Support came from health systems, physician groups, hospitals and specialty organizations. The committee noted outstanding definitional issues between developers and deployers and said those would be clarified in later amendments. Chair and members said transparency from developers would aid deployers in addressing biased outputs.
Committee action and next steps
The committee accepted author amendments and passed SB 503 as amended to the Committee on Privacy and Consumer Protection for further refinement of definitions (notably, developer vs. deployer) and to align implementation details with the privacy committee’s review.
Why this matters
Health care organizations are increasingly piloting and deploying AI for tasks ranging from image interpretation to resource triage. The committee’s action establishes expectations that tools used in clinical decision-making be subject to bias mitigation efforts and ongoing monitoring rather than unregulated deployment.
Votes at a glance
Action: Motion to pass SB 503 as amended to Privacy and Consumer Protection — Passed (committee roll call recorded)
