CA SB503

Requires developers and deployers of AI systems used for clinical decisionmaking or health care resource allocation to identify systems with known or foreseeable risks of biased impacts on patients based on protected characteristics

Mandates reasonable efforts to mitigate bias risks and requires deployers to regularly monitor AI systems and take proportionate steps to address any bias that occurs

Beginning January 1, 2027, developers must submit compliance reports to the Department of Public Health before making AI systems commercially available to deployers, including reports for substantial updates

Beginning January 1, 2027, deployers must annually report their compliance efforts to the department, which will publish all reports on its website

Defines "biased impact" as unintended adverse effects on individuals' health care access, quality, or outcomes based on protected characteristics under California civil rights law