California Just Set New Ground Rules for AI in Healthcare. Here's What Hawaii Practices Should Know

California's governor signed three healthcare-related AI bills at the end of September 2026. They do not apply directly to a practice in Hawaii, but they show where regulators are heading, and they are a helpful checklist for any practice using AI tools.

The core idea: a licensed human stays accountable

AB 1979 says that when an AI "clinical decision support" tool informs patient care, a licensed provider must still exercise independent clinical judgment. It also bars using AI to independently perform clinical functions reserved for licensed professionals. Administrative tools like scheduling and appointment reminders are exempt. The law also treats businesses that offer consumer-facing "health care chatbots" as providers for purposes of California's medical-information confidentiality law. Effective January 1, 2027.

SB 503 asks developers and users of these tools to look for risks of biased impacts, such as worse access or outcomes based on race, sex, or age, and to take reasonable steps to reduce them. Developers must give users documentation about training data, testing, and intended use. Effective January 1, 2027.

SB 947 covers automated employment decisions. Software output alone cannot justify discipline or termination; a human has to corroborate it. Effective July 1, 2027.

A related law, AB 1609, applies to large businesses (over $500 million in annual revenue). It requires them to disclose when a customer is talking to a chatbot and to connect people to a human within 15 minutes during business hours on request. It exempts certain hospital communications. Effective January 1, 2027.

Two other bills, one limiting AI in psychotherapy and one protecting clinicians who override AI, were vetoed, though the governor praised their goals. Law firm analysts expect lawmakers to revisit them.

Why this matters in Hawaii

These are California laws, so check with a Hawaii healthcare attorney about what applies to you. But the direction is clear across many states: AI can support care, and humans must remain accountable.

A practical AI checklist for any practice

1. List every AI tool you use, from scribes to chatbots to phone assistants, and note who reviews its output.
2. Keep a clinician in charge. Make sure patient-specific clinical decisions are made and documented by a licensed provider.
3. Disclose AI to patients. If a chatbot or voice assistant talks to patients, say so up front and offer a way to reach a person.
4. Ask vendors for documentation about training data, testing, and bias monitoring.
5. Keep AI on the administrative side (scheduling, reminders, intake logistics) unless you have a clear clinical oversight plan.

Quantum Clinical Advisors helps practices evaluate AI tools and set up sensible oversight. This post is general information, not legal advice.