In a significant move toward regulating artificial intelligence (AI) in health care, Assembly Bill (AB) 3030 will require health facilities, clinics, and physician practices in California to provide disclaimers to patients about their use of generative AI (GenAI) in communications regarding patient clinical information.

The legislation, which takes effect on January 1, 2025, is among more than a dozen AI-related bills that California Governor Gavin Newsom signed into law in September 2024. By mandating disclosure about the usage of GenAI, AB 3030 aims to promote transparency, informed decision-making, and accountability — key themes in the emerging regulatory framework for AI.

Under the new statute (as codified in Health & Safety Code §1339.75), all health facilities, clinics, and physician practices that use GenAI to generate patient communications regarding patient clinical information must include both of the following in those communications:

  1. Disclaimer of AI Generation: A disclaimer indicating that the communication was generated by GenAI.
  2. Instructions for Human Contact: Clear instructions on how to contact a human health care provider; employee of the health facility, clinic, or physician practice; or other appropriate person.

The disclaimer must be formatted according to the medium of communication:

  • Written communications (e.g., letters, emails): The disclaimer must appear prominently in physical or digital format at the beginning of each communication.
  • Audio communications: The disclaimer must be provided verbally at the start and end of the interaction.
  • Video communications and written communications involving continuous online interactions (e.g., chat-based telehealth): The disclaimer must be displayed prominently throughout the interaction.

These requirements apply only to patient communications pertaining to “patient clinical information,” which means information related to a patient’s health status. Such information does not include administrative matters, such as appointment scheduling and billing. Notably, if an AI-generated communication is “read and reviewed” by a licensed or certified human health care provider, the disclaimer and instruction requirements do not apply.

Health facilities and clinics that fail to comply will be subject to penalties, including fines and licensure actions. Similarly, physicians who fail to comply will be subject to disciplinary action against their licenses.

To avoid such penalties, providers should evaluate how they use GenAI in their patient communications and take actions to ensure compliance with AB 3030, including:

  • Technology Updates: Collaborate with GenAI technology vendors to implement necessary updates for including disclaimers and instructions.
  • Communication Templates: Develop templates with consistent messaging for disclaimers and instructions across different communication platforms (e.g., email, chat boxes, video).
  • Policies and Procedures: Establish policies and procedures for using GenAI in patient communications, including:
    • Protocols for when human review is necessary to bypass the disclaimer and instruction requirements.
    • Policies and procedures to safeguard the privacy and security of patient information that is processed with GenAI technology.
  • Training: Train staff on integrating the new requirements into their workflows.
  • Patient Education: Educate patients about the use of GenAI in their care and the purpose of the disclaimers and instructions to maintain transparency and trust.