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.
Pennsylvania Senate Bill 1113 (2026) has not been enacted. It was referred to the Senate Institutional Sustainability and Innovation committee on January 9, 2026. If enacted, it would regulate AI use in three health settings. (1) Health care facilities would be covered by new 35 Pa.C.S. Chapter 35, enforced by the Department of Health. (2) Health insurers would be covered by new 40 Pa.C.S. Chapter 52, enforced by the Insurance Department. (3) Medical Assistance (MA) and CHIP managed care plans would be covered by new 40 Pa.C.S. Chapter 53, enforced by the Department of Human Services. Facilities that use AI-based algorithms for clinical decision making would have to disclose that use to patients (§ 3502). Disclosure would be required in related written communications and on the facility's website. AI-generated patient communications about clinical information would need a clear disclaimer and instructions for reaching a human. Administrative communications and communications reviewed by a human provider are excepted. Under § 3503, the algorithms must: - not supersede provider clinical decision making; - not discriminate; - be applied fairly; - be periodically reviewed; - not use patient data beyond the stated purpose; - not create foreseeable, material risks of harm. Insurers and MA/CHIP plans that use AI in utilization review would have parallel duties (§§ 5202–5203, 5302–5303). They would have to disclose the use to participating network providers and to covered persons or enrollees, and post it on their websites. Determinations must consider the person's clinical history and circumstances and may not rest solely on a group data set. Before a provider issues or upholds a denial, reduction or termination of benefits, including a denied prior authorization, the provider must review individual clinical records, document that review, and exercise judgment independent of the AI's recommendations (§§ 5205, 5305). Each regulated entity would file an annual AI compliance statement (§§ 3504, 5204, 5304). It would have to: - summarize the function and scope of the algorithms; - give a logic or decision tree; - describe each training data set and its source; - attest to compliance with the responsible-use section and give evidence; - describe the entity's oversight process. Each department would issue an annual aggregated, deidentified report. It would post the report online, and the insurance and human services reports would also go to the General Assembly (§§ 3505, 5206, 5306). Each department would also set record retention policies, may request more information, and would issue regulations or guidance. It would review them every three years (§§ 3506–3508, 3513, 5207–5208, 5214, 5307–5308, 5314). Third-party vendors that supply AI algorithms or services would be subject to the chapters (§§ 3508, 5209, 5309). Tools used for administration, scheduling, scribing, clinical calculators or validated static decision-support are exempt from the facility chapter (§ 3509). AI used for administrative or scheduling purposes unrelated to benefit denials is exempt from the insurer and MA/CHIP chapters (§§ 5210, 5310). Enforcement would include: - civil penalties up to $5,000 per violation, with each nondisclosure counted separately; - annual aggregate caps of $500,000 for a facility, insurer or plan, and $100,000 for any other person (§§ 3510, 5211, 5311); - injunctions; - plans of correction in lieu of fines (§§ 3511, 5212, 5312); - for insurers and MA/CHIP plans, treatment of violations as Unfair Insurance Practices Act violations; - for insurers and MA/CHIP plans, possible temporary bars on new enrollment. Appeals go to Commonwealth Court (§§ 3512, 5213, 5313). The act would take effect in one year (Section 3).
Status history
Current status as of 2026-01-09
In committee
2026-01-09
observed 2026-07-26
Impact areas
- Enterprise Adoption
- Quality Assurance
- AI Policy
- Privacy & Data Protection
- Public Sector Use
- Health
- Algorithms & Automated Decisions