Bill > SB503

This bill requires developers and deployers of artificial intelligence (AI) systems used for clinical decision support, meaning AI that aids in diagnosis or treatment decisions, to take reasonable steps to identify and mitigate potential biases that could negatively impact patients based on protected characteristics. Developers must provide deployers with information about the system's intended uses, risks, training data, evaluation methods, and bias mitigation efforts. Deployers must regularly monitor these AI systems for bias and take action to address it. Both developers and deployers are responsible for reporting their compliance efforts to the state starting in 2027, and these reports will be made public. The bill defines key terms like "developer" (those who create or significantly modify AI for health use) and "deployer" (healthcare entities that use these AI systems), and clarifies that this law is in addition to existing anti-discrimination laws.