AI Frontier Institute

Creates the Louisiana Artificial Intelligence Insurance Fairness Act (OR +$1,029,533 SG EX See Note)

View primary source →

HB 880 (2026 Regular Session) has not been enacted. It was prefiled, referred to the House Committee on Insurance, and has gone no further. If enacted, it would create the Louisiana Artificial Intelligence Insurance Fairness Act (R.S. 22:1401 through 1424). The Act would regulate how insurers use AI and algorithmic decision systems (ADS) in underwriting, rating, and claims for homeowners, commercial property and casualty, auto, health, and life insurance (§1403, §1404). Insurers would stay responsible when they use third-party models (§1404(C)). Main provisions: (1) Fairness standards (§1405). Insurers could not intentionally use protected-class membership. They could not use proxy variables that cause disparate impact unless each is actuarially justified and is the least discriminatory alternative. They could not make arbitrary decisions. Credit scores, certain geographic variables, coverage gaps, education or occupation, social media data, and retail purchasing data would be presumed proxies that get heightened scrutiny. (2) Governance (§1406). Insurers would need written AI governance policies, a designated senior officer, and staff training. (3) Annual disparate impact audits (§1407). A qualified independent actuary or data scientist would do the audit. A certification would be filed with the Commissioner of Insurance by June 30 each year, and the Commissioner would keep a public registry of certifications. (4) Prohibited variables (§1408). For homeowners insurance, credit scores, protected-class characteristics, social media and online behavioral data, and retail purchasing data would be banned. Commercial lines would ban only protected-class characteristics and protected-class membership. The Commissioner could add more banned variables by rule. (5) Claims (§1409). A licensed claims professional would have to review and approve each final AI-assisted claims payment determination. Algorithmic lowball settlement offers would be barred. (6) Consumer rights. Insurers would have to disclose AI use, and chatbots would have to identify themselves as automated (§1410). Consumers could request a written explanation of an adverse action (§1411) and human review (§1412). Consumers could also dispute and correct data (§1413). (7) Oversight (§1414–§1417). The Commissioner would get rulemaking authority and market conduct examination powers. Insurers would have to file notice 60 days before deploying a new or materially modified system, and the Commissioner could stay or block deployment. The Commissioner would also publish an annual report. (8) Enforcement (§1418–§1422). Civil penalties would range from $5,000 to $250,000 per violation, and certain filing failures would be penalized per day. Consumers would have a private right of action with actual, statutory ($1,000–$10,000), and punitive damages plus attorneys' fees (§1419). Class actions would be allowed (§1420). A dedicated fund (§1421) and a 15-member advisory board (§1422) would be created. (9) Relation to other laws and implementation. The Act would supplement existing anti-discrimination laws (§1423). The text includes a severability clause (§1424). A phased effective-date section (§1425) is also included, although the bill's enacting clause and digest refer only to §1401–1424.

← Back to the tracker