H5172 creates a new chapter in Rhode Island insurance law regulating how health insurers use artificial intelligence to manage coverage and claims. The bill’s stated purpose is to ensure transparency, accountability, and compliance with state and federal requirements, including anti-discrimination and privacy laws. It defines key terms such as artificial intelligence, adverse determination, insurer, enrollee, medically necessary care, and third party, and applies to a broad range of Rhode Island health insurers, including Medicaid managed care organizations.
The bill requires insurers to publicly disclose how they use AI in claims and coverage decisions, including the algorithms, data, and resulting determinations, and to retain AI decision documentation for at least five years. It also requires notice to enrollees and providers when AI is used in an adverse determination and a clear appeal process. Importantly, insurers may not rely exclusively on AI to deny, reduce, or alter coverage for medically necessary care; adverse determinations must be reviewed by qualified licensed clinicians with authority to reverse them. The bill further requires ongoing monitoring and audits of AI use, including safeguards for privacy, reasonable administrative burdens, and compliance with anti-discrimination rules.
The bill would also expand enforcement authority for the Office of the Health Insurance Commissioner and the Executive Office of Health and Human Services, with collaboration from other state agencies. It authorizes rulemaking, independent enforcement actions under other laws, and a private right of action for enrollees. Violations could lead to orders limiting AI use, civil fines up to $50,000 per violation, license suspension or revocation, and damages for affected enrollees and providers. The bill states that its provisions supplement existing law and control in the event of direct conflict.
In addition to the AI-specific chapter, the bill amends existing statutes governing the secretary of health and human services and the health insurance commissioner to explicitly include enforcement of the new chapter. Those sections also contain extensive existing duties related to Medicaid oversight, health insurance regulation, prior authorization, claims processing, behavioral health parity, and health system reporting. As drafted, the bill would therefore fit into a broader regulatory framework already giving state health officials substantial oversight of insurers and publicly funded health programs.
The overall sentiment reflected by the bill text and caption is strongly supportive of consumer protection and oversight, with the bill framed as a transparency and accountability measure rather than a restriction on innovation. No committee transcript or vote history was provided, so there is no recorded public debate in the supplied materials. Based on the language alone, likely points of contention would be insurer compliance costs, disclosure of proprietary algorithms and data, the scope of state authority over AI tools, and the private right of action and penalties. Supporters would likely emphasize patient protections, clinician review, and prevention of AI-driven denials of medically necessary care.
This bill would add a new chapter to Title 27 of the Rhode Island General Laws regulating health insurers’ use of artificial intelligence in coverage and claims administration, and it would amend related provisions in Title 42 to make the Office of the Health Insurance Commissioner and EOHHS responsible for enforcement. It would impose new disclosure, documentation, review, audit, notice, and appeal requirements on insurers and their third-party vendors, and it would create a private right of action plus administrative penalties for violations. The bill would affect health insurers, Medicaid managed care organizations, enrollees, healthcare providers, pharmacies, hospitals, and state regulators.
The bill is presented in a clearly pro-consumer, pro-oversight posture, with its stated goal of ensuring transparency, accountability, privacy, and non-discrimination in AI-driven insurance decisions. Because no committee testimony or votes were provided, there is no direct evidence of opposition or support in the record supplied here. The likely general sentiment inferred from the text is favorable among patient advocates and regulators, with insurers and vendors likely to view it as a significant compliance and operational burden.
The main likely points of contention are the breadth of required disclosure of AI systems, including algorithms and data; the prohibition on relying exclusively on AI to make adverse coverage decisions; and the requirement for clinician review with authority to overturn denials. Insurers may object to the administrative burden, potential exposure of proprietary or confidential information, and the risk of litigation and fines under the private right of action. Supporters are likely to argue that these provisions are necessary to prevent inappropriate denials of medically necessary care, protect privacy, and reduce discriminatory outcomes.