Committee begins framing AI standards: definitions, informed consent and data-security concerns

Staff presented a draft regulation (proposed new 318.3) that would define AI and generative AI, require licensees to take reasonable steps to understand tools they deploy, and mandate documentation when AI is used to inform clinical decision-making. The draft included a broad prohibition on using AI to fabricate clinical findings, diagnoses, documentation or treatment plans.

The committee discussed a new state bill, AB 1979, that would prohibit use of AI to perform activities requiring professional judgment; staff cautioned that the board must avoid adopting regulations inconsistent with statute and will monitor AB 1979 as it moves through the Legislature. Committee members pressed for an AI-specific informed consent process that goes beyond a general disclosure and includes procedural steps to ensure patient autonomy, the ability to decline use of AI, and documentation of any post-AI follow-up.

Members raised patient-data security concerns when patient records are transmitted to or processed by external AI systems. The committee directed staff to analyze HIPAA implications and to clarify what counts as de-identified data for AI use. Several members urged a separate, detailed AI-informed-consent form and suggested incorporating competency-based education or continuing-education requirements to raise "AI literacy" among licensees.

Staff also noted enforcement challenges: AI may enable sophisticated documentation that could mislead investigators, so the board will consider strategic enforcement and patient-statement collection practices to detect abuse. The board's legal team will review the draft for consistency with pending state legislation and enforcement policy before the next committee meeting.