AI Regulation Tracker / Patient communications and disclosure
California Requires an AI Disclaimer and a Human Contact on GenAI Patient Messages
Effective January 1, 2025. California AB 3030 requires health facilities, clinics, physician offices, and group practices to add a prominent disclaimer and clear human-contact instructions whenever generative AI is used to communicate clinical information to a patient, unless a licensed provider reviews the message first.
California AB 3030 has been in force since January 1, 2025, and it answers a question that a lot of practices started asking the moment generative AI got good at writing patient messages: if the software drafted the note explaining a lab result, does the patient get to know that? In California, the answer is yes, unless a human clinician reviewed it first. The law applies to health facilities, clinics, physician offices, and group practices that use generative AI to generate written or verbal communications conveying clinical information about a patient's health.
The statute requires two elements. The first is a disclaimer. In the words of the law, the communication must carry "a disclaimer that indicates to the patient that the communication was generated by generative artificial intelligence." The second is a path back to a person. The message must include "clear instructions describing how a patient may contact a human health care provider, employee of the health facility, clinic, physician's office, or office of a group provider, or other appropriate person." Label plus lifeline. That is the design.
Where does the disclaimer have to appear?
The placement rules track the format so the disclosure is actually seen. For written communications like letters and emails, the disclaimer has to appear prominently at the beginning. For continuous online interactions such as chat-based telehealth, it has to be prominently displayed throughout the exchange. For audio, it has to be given verbally at the start and again at the end. For video, it has to be displayed throughout. The common requirement across every channel is prominence. A disclosure buried in a footer or spoken once too fast to register is not what the statute is asking for.
What is the human-review exemption?
This is the provision that shapes most practices' compliance strategy, so it is worth stating precisely. If an AI-generated communication is read and reviewed by a licensed or certified health care provider before it goes to the patient, the disclaimer requirement does not apply. That gives practices a genuine choice. Either a qualified human reviews the message, in which case it is treated as a clinician's communication, or the message goes out with the AI label and the human-contact instructions attached. What you cannot do is send an unreviewed, AI-generated clinical message to a patient with no disclosure at all.
What practices and vendors should do now
For physician practices and health systems, map where generative AI touches patient communication. Portal message drafts, appointment and result summaries, chat-based intake, and any patient-facing chatbot are the usual candidates. For each one, decide the model: human review or disclosure. If you rely on review, make the review real and document who did it and when, because that is what converts an AI draft into a compliant clinician message. If you rely on disclosure, build the disclaimer and the human-contact instructions into the template so they cannot be dropped, and match the placement to the format. For messaging and chatbot vendors, this is a product requirement, not a nice-to-have, and California customers will expect it configured out of the box. AB 3030 does not stop you from using AI to talk to patients. It makes you tell them, or put a clinician in front of it first.
Questions professionals are asking
When does AB 3030 require a disclaimer?
When a health facility, clinic, physician office, or group practice uses generative AI to communicate clinical information about a patient's health, and no licensed or certified provider reviewed the message before it was sent. In that case the communication needs a prominent GenAI disclaimer plus instructions to reach a human.
Is there a way to avoid the disclaimer requirement?
Yes. If a licensed or certified health care provider reads and reviews the AI-generated communication before it goes to the patient, the disclaimer requirement does not apply. Practices generally choose per channel between real human review and attaching the disclosure.
Where does the disclaimer have to appear?
Placement tracks the format. At the beginning of written messages like letters and emails, throughout chat-based and video interactions, and verbally at the start and end of audio. The consistent requirement is that the disclosure be prominent, not buried.
Does AB 3030 apply to patient-facing chatbots?
Yes, when the chatbot uses generative AI to convey clinical information about a patient's health. For continuous chat interactions the disclaimer must be displayed throughout, and patients must be told how to reach a human, unless a licensed provider reviews the content first.
RELATED BRIEFINGS
Browse the full AI Regulation News tracker
Informational analysis for working professionals, not legal advice. Confirm how any statute or requirement applies to your situation with qualified counsel licensed in the relevant jurisdiction.