Providing for artificial intelligence in facilities, for artificial intelligence use by insurers and for artificial intelligence use by MA or CHIP managed care plans; imposing duties on the Department of Health, the Insurance Department and the Department of Human Services; and imposing penalties.
HB1925 would create a new Pennsylvania framework governing the use of artificial intelligence in three health-related settings: health care facilities, health insurers, and MA or CHIP managed care plans. In each setting, the bill requires disclosure when AI is used for clinical decision making or utilization review, including plain-language notices to patients, covered persons, enrollees, and participating providers, plus website postings. It also requires that human professionals retain independent judgment and not be penalized for disagreeing with an AI-generated recommendation, and it exempts AI used only for administrative tasks such as scheduling, billing, scribing, or other non-decision-making functions.
The bill imposes responsible-use standards on AI systems and the data that support them. Covered entities would have to ensure AI does not supersede human clinical judgment, does not discriminate, is fairly and equitably applied, is reviewed and updated at least quarterly, and does not create foreseeable material risks of harm. The bill also limits use of patient or enrollee data to the intended purpose unless otherwise permitted by law or informed consent, while allowing de-identified or aggregated data for research, development, or improvement of AI tools. Facilities, insurers, and managed care plans would need internal governance policies, validation processes, annual AI compliance statements filed with the relevant department, and retention of related records for at least five years or longer if required by existing law.
The bill would amend Titles 35, 40, and 67 of the Pennsylvania Consolidated Statutes and assign oversight to the Department of Health, the Insurance Department, and the Department of Human Services, respectively. Each department would be authorized to issue regulations or guidance, request additional information, and enforce compliance. The bill also requires annual aggregated, deidentified reports to the General Assembly and public posting of those reports. Violations could trigger civil penalties, plans of correction, injunctions, and, in some cases, treatment as unfair or deceptive practices under consumer protection or insurance law. The bill also allows temporary restrictions on new enrollments for violators in the insurance and managed care contexts.
The overall sentiment reflected in the committee action appears supportive, with the bill advancing out of the House Communications & Technology Committee on a 18-7 vote after amendments were adopted. The amendment votes were largely noncontroversial, including one unanimous 25-0 vote, suggesting the committee was willing to refine the measure rather than reject it. At the same time, the final committee vote shows some opposition, indicating concerns remained about the scope or regulatory burden of the proposal.
The main points of contention are likely to be the breadth of the disclosure and reporting requirements, the compliance burden on facilities and insurers, and the extent to which the bill regulates third-party AI vendors and internal governance processes. Another likely issue is the bill’s strong human-oversight requirement, which could be viewed as necessary patient protection by supporters but as potentially limiting operational use of AI by opponents. The bill’s penalties, enforcement authority, and application to utilization review and clinical decision making are also likely to be debated, especially by health care and insurance stakeholders concerned about administrative complexity and confidentiality of submitted materials.
HB1925 would add new chapters to Titles 35, 40, and 67 of the Pennsylvania Consolidated Statutes, creating state-level AI disclosure, governance, reporting, and enforcement requirements for health care facilities, insurers, and MA/CHIP managed care plans. It would require the Department of Health, Insurance Department, and Department of Human Services to promulgate implementing regulations or guidance, coordinate with one another, and oversee compliance. The bill would also create new duties for covered entities and their third-party AI vendors, while authorizing civil penalties, plans of correction, injunctions, and related enforcement actions under existing consumer protection and insurance laws.
The committee record suggests generally favorable sentiment toward the bill’s goals, with amendments adopted and the measure reported from committee on an 18-7 vote. The unanimous amendment vote indicates broad agreement on at least some refinements, while the final split vote shows that a meaningful minority remained unconvinced. Overall, the discussion history points to cautious support for regulating AI in health care and insurance, paired with concern about implementation details and regulatory reach.
The likely areas of contention are whether the bill goes too far in regulating AI use in clinical and utilization-review settings, how burdensome the annual compliance statements and quarterly review requirements will be, and whether the disclosure rules are sufficiently clear and workable. Health care providers and insurers may also object to the bill’s restrictions on AI-driven recommendations, the potential exposure to penalties, and the requirement to maintain internal governance structures. Supporters, by contrast, are likely focused on transparency, anti-discrimination protections, and preserving human judgment in decisions affecting patient care and coverage.