RELATING TO THE USE OF ARTIFICIAL INTELLIGENCE IN HEALTH CARE.
SB2281 creates a new part in chapter 321, Hawaii Revised Statutes, to regulate the use of artificial intelligence in health care. It requires health care providers that use AI in patient-facing remote communication to disclose that the patient is interacting with AI, and to identify how the patient can reach a human health care provider. For AI used in making or substantially influencing a “consequential decision” affecting a patient’s physical or mental health, the bill requires advance written notice, a plain-language description of the system and its purpose, and an opt-out from profiling used for legally or similarly significant decisions.
The bill also gives patients post-decision rights when AI is involved in consequential decisions: a written explanation of the decision and the AI’s role, an opportunity to correct inaccurate health or personal data, and an opportunity to appeal with human review when technically feasible, except where delay would not be in the patient’s best interest. Health care providers must maintain qualified human AI oversight personnel to review, validate, or override AI outputs before they are used in consequential decisions, and must monitor systems, evaluate bias and safety risks, address deficiencies, and keep records on system design, training data, oversight, and corrective actions. The Department of Health, in coordination with the Department of Business, Economic Development, and Tourism, must adopt implementing rules. The bill’s operative provisions take effect July 1, 2028, while the overall act takes effect January 30, 2050.
SB2281 would add a new regulatory framework to Hawaii health law governing AI use by health care providers, expanding chapter 321 with disclosure, notice, patient-rights, oversight, monitoring, and recordkeeping requirements. It would affect providers using AI for telehealth-style interactions and for clinical, administrative, or other decisions that significantly affect a patient’s health, while also implicating the handling of individually identifiable health information and personal data. The bill would require state rulemaking to define oversight qualifications and evaluation frequency, and would likely impose compliance obligations on hospitals, clinics, telehealth vendors, and other providers using AI systems.
The available voting history suggests broad support in committee, with unanimous passage in both the Senate Health and Human Services Committee and the Senate Labor and Technology Committee, each passing the bill with amendments. The amended version then advanced through second reading and referral to the next committees, indicating generally favorable sentiment toward regulating AI in health care rather than opposition to the concept. No committee transcript was provided, so there is no recorded floor or hearing debate to indicate broader public controversy in the materials supplied.
The main points of potential contention are the scope of the disclosure and notice obligations, the breadth of the definition of a “consequential decision,” and the operational burden of requiring human oversight personnel, performance evaluations, and detailed recordkeeping. Health care providers and technology vendors may be concerned about compliance costs, workflow delays, and how to apply the opt-out and appeal rights in time-sensitive or emergency settings. On the other hand, patient advocates are likely to favor the bill’s transparency, correction, and human-review protections, especially where AI could affect diagnosis, treatment, access to care, or other significant health outcomes.