Regulon — LA-HB-880 · LA HB 880 (AI Insurance Fairness)

WHAT THIS BILL REGULATES · 7 REQUIREMENT TYPES

How Is This Bill Enforced

Enforcement Authority

Commissioner of Insurance has primary enforcement authority through market conduct examinations, rulemaking, and civil penalties. Private right of action available to any consumer who suffers harm from a covered insurer's violation, with no exhaustion of administrative remedies required. Class actions are authorized for systematic patterns of discriminatory algorithmic practice.

Private Right of Action

Private right of action available to any consumer who suffers harm from a covered insurer's violation, with no exhaustion of administrative remedies required.

Penalties

Private action: actual damages, statutory damages of $1,000–$10,000 per violation regardless of actual damages, punitive damages for knowing/willful/reckless conduct, injunctive or declaratory relief, and reasonable attorneys' fees and costs. Commissioner civil penalties range from $5,000–$250,000 per violation depending on violation type, with filing violations assessed per day. Two-year statute of limitations from discovery; filing a complaint with the Commissioner tolls limitations.

Verbatim statutory text on the left; plain-language analysis and a per-section checklist on the right. Numbered markers cross-link to the matching checklist row.

Statutory Text

Analysis & Obligations

R.S. 22:1401

Short Title

This Chapter shall be known and may be cited as the "Louisiana Artificial Intelligence Insurance Fairness Act" or the "LAIFA."

Establishes the short title of the chapter as the Louisiana Artificial Intelligence Insurance Fairness Act or LAIFA. Creates no compliance obligations.

R.S. 22:1402

Legislative Findings and Purpose

A–B A. The Legislature finds and declares: (1) Artificial intelligence, machine learning, and algorithmic decision systems are being rapidly deployed across the insurance industry to make decisions about coverage eligibility, premium pricing, claims handling, and policy renewal. (2) These algorithmic systems present significant risks of unfair discrimination. Machine learning models trained on historical data inherit the biases embedded in that historical data, including the effects of past discrimination in housing, lending, insurance, and economic opportunity. When used in insurance without adequate safeguards, such models can perpetuate and amplify discrimination against communities protected under Louisiana and federal law. (3) Black, Hispanic, Indigenous, and other minority communities, communities with lower incomes, communities with older housing stock, and communities located in historically redlined areas are particularly vulnerable to disparate harm from algorithmic insurance practices. (4) Louisiana's homeowners and small business owners within protected classes, already facing a challenged insurance market, are at heightened risk of being further disadvantaged by opaque algorithmic underwriting and pricing systems. (5) Transparency in insurance decision-making is essential to enabling consumers to understand and challenge decisions that affect their access to insurance and their cost of coverage. (6) The responsible use of artificial intelligence in insurance can reduce costs, improve efficiency, and expand coverage if properly regulated. This Act is not intended to prohibit the use of AI in insurance, but to ensure that such use is fair, transparent, and accountable. (7) Louisiana has a compelling interest in regulating the use of algorithmic decision systems in insurance consistent with its longstanding prohibitions on unfair discrimination in insurance practices. B. The purpose of this Act is to: (1) Establish clear standards for the development, testing, and deployment of AI systems used in insurance decisions; (2) Require insurers to conduct and certify annual disparate impact audits of algorithmic systems; (3) Prohibit the use of variables that function as proxies for protected class membership in ways that produce discriminatory outcomes; (4) Ensure that consumers have access to meaningful explanations when algorithmic systems are used in adverse insurance decisions; (5) Establish a private right of action for individuals subjected to discriminatory algorithmic insurance practices; and (6) Position Louisiana as a national leader in the fair and responsible regulation of AI in insurance.

Sets forth the legislature's findings regarding the risks of AI-driven discrimination in insurance — particularly as it affects Black, Hispanic, Indigenous, and lower-income communities — and states the Act's purposes: establishing standards for AI in insurance, requiring disparate impact audits, prohibiting proxy variables, ensuring consumer access to explanations, creating a private right of action, and positioning Louisiana as a national leader in AI insurance regulation. This section creates no independent compliance obligations.

R.S. 22:1403

Definitions

(1)–(13) As used in this Chapter, unless the context clearly requires otherwise: (1) "Adverse action" means a denial of insurance coverage, a cancellation or nonrenewal of an insurance policy, an increase in premium or surcharge, a reduction in coverage limits or benefits, or any other action by an insurer that is materially unfavorable to an insurance consumer. (2) "Algorithmic decision system" or "ADS" means any computational process, including one derived from machine learning, artificial intelligence, statistical modeling, or other data processing, that processes personal information to generate a score, classification, recommendation, or decision that is used by an insurer in making insurance decisions. The term includes both systems developed by the insurer and third-party systems licensed or contracted by the insurer. (3) "Artificial intelligence" means machine-based systems that use statistical and computational methods to perform tasks that would otherwise require human intelligence, including classification, prediction, recommendation, and decisionmaking. (4) "Commissioner" means the Commissioner of Insurance of the State of Louisiana. (5) "Consumer" means any individual who is an applicant for, policyholder of, or claimant under an insurance policy regulated by this Act. (6) "Covered insurer" means any insurer authorized to transact insurance in Louisiana that uses an algorithmic decision system in connection with personal lines homeowners insurance, commercial lines property and casualty insurance, personal automobile insurance, commercial automobile insurance, health insurance, or life insurance issued in Louisiana. (7) "Disparate impact" means a materially adverse and statistically significant difference in insurance outcomes, including denial rates, premium levels, or claims payment rates, between consumers who are members of a protected class and similarly situated consumers who are not members of that protected class. (8) "Explainability" means the capacity of an insurer to provide a specific, intelligible, and actionable explanation of the factors that most contributed to a particular insurance decision made with the assistance of an algorithmic decision system. (9) "Input variable" means any data element or feature used as an input to an algorithmic decision system. (10) "Insurance decision" means any decision by a covered insurer regarding underwriting, rating, policy issuance, renewal, cancellation, or claims handling. (11) "Protected class" means any category of persons protected from discrimination under applicable Louisiana or federal law, including but not limited to race, color, religion, sex, national origin, disability, age, familial status, and sexual orientation. (12) "Proxy variable" means an input variable that is not a protected class characteristic itself but is statistically correlated with membership in a protected class in a manner that causes or contributes to disparate impact on that protected class. (13) "Third-party model" means an algorithmic decision system developed, owned, or operated by a person other than the covered insurer, licensed or contracted to the insurer for use in insurance decisions.

Defines thirteen key terms used throughout the Act, including algorithmic decision system, covered insurer, disparate impact, proxy variable, and adverse action. The covered-insurer definition spans personal and commercial lines across homeowners, auto, health, and life insurance. Notably, the ADS definition is broad — it encompasses any computational process that generates a score, classification, recommendation, or decision used in insurance decisions, including third-party systems.

R.S. 22:1404

Scope and Applicability

Deployer

A–C 1 A. This Chapter applies to all covered insurers transacting insurance in Louisiana. B. The obligations imposed by this Chapter apply to: (1) Algorithmic decision systems developed by the covered insurer; (2) Third-party algorithmic decision systems used by the covered insurer; and (3) Hybrid systems combining algorithmic outputs with human underwriting judgment. C. The use of a third-party model does not relieve a covered insurer of any obligation under this Chapter. Covered insurers shall contractually require third-party model providers to supply the information and access necessary for the insurer to fulfill its obligations under this Chapter.

D D. This Chapter does not apply to: (1) Basic actuarial calculations or rating factor applications that do not involve machine learning or AI components; (2) Fraud detection systems used exclusively to detect insurance fraud, provided such systems are not used to influence coverage or pricing decisions for consumers who are not suspected of fraud; or (3) Internal administrative systems that do not affect insurance decisions regarding consumers.

Defines the scope of the Act: it applies to all covered insurers transacting insurance in Louisiana, covering insurer-developed systems, third-party systems, and hybrid systems combining algorithmic outputs with human judgment. Critically, the use of a third-party model does not relieve an insurer of compliance obligations — insurers must contractually require third-party model providers to supply necessary information and access. The section also carves out basic actuarial calculations without ML/AI components, fraud-detection-only systems, and internal administrative systems that do not affect consumer-facing insurance decisions.

Compliance actions 1 item

Covered insurers must contractually require third-party model providers to supply the information and access necessary for the insurer to fulfill its obligations under this Act.

G-01

R.S. 22:1405

Fairness Standards

Deployer

A(1)–(3) 2 A. No covered insurer shall use an algorithmic decision system in a manner that: (1) Intentionally considers the protected class membership of any consumer as an input variable; (2) Uses input variables that are proxy variables for protected class membership and that produce a disparate impact on a protected class, unless the insurer demonstrates that each such variable is actuarially justified by sound actuarial principles and the least discriminatory alternative sufficient to achieve the actuarial purpose; or (3) Produces decisions that are arbitrary, capricious, or bear no rational relationship to the risk of loss being assessed.

B(1)–(6) 2 B. The following input variables are presumed to function as proxy variables that may produce disparate impact and shall be subject to heightened scrutiny under this Chapter: (1) Credit scores or credit-based insurance scores used in personal automobile or homeowners insurance underwriting or rating; (2) Geographic variables at the census tract, ZIP code, or neighborhood level where those variables are correlated with the racial, ethnic, or national origin composition of the area; (3) Prior insurance coverage gaps or lapses, when used without adjustment for demographic patterns in coverage gaps; (4) Education or occupation variables; (5) Social media data or online behavioral data; and (6) Consumer purchasing patterns or retail transaction data.

C 2 C. Use of a presumptive proxy variable is not automatically prohibited, but an insurer using such a variable bears the burden of demonstrating, in its disparate impact audit required under Section 1407, that the variable does not produce a statistically significant disparate impact on any protected class, or that the disparate impact is justified by actuarial necessity and no less discriminatory alternative exists.

Establishes the Act's core non-discrimination framework for algorithmic decision systems. Prohibits covered insurers from (1) intentionally using protected-class membership as an input variable, (2) using proxy variables that produce disparate impact unless the insurer demonstrates actuarial justification and least-discriminatory alternative, or (3) producing arbitrary or capricious decisions. Designates six categories of presumptive proxy variables — credit scores, geographic variables correlated with racial/ethnic composition, prior coverage gaps, education/occupation, social media data, and consumer purchasing patterns — subject to heightened scrutiny. Use of a presumptive proxy variable is not automatically prohibited but shifts the burden to the insurer to demonstrate either no disparate impact or actuarial necessity with no less discriminatory alternative.

Compliance actions 1 item

Covered insurers must not use an ADS that intentionally considers protected-class membership, uses proxy variables producing disparate impact without actuarial justification and least-discriminatory-alternative showing, or produces arbitrary or capricious decisions. Six enumerated variable categories are presumptive proxy variables subject to heightened scrutiny and burden-shifting.

H-02.1

R.S. 22:1406

Governance Requirements for Algorithmic Systems

Deployer

A(1)–(6) 3 A. Each covered insurer that uses an algorithmic decision system shall adopt written governance policies and procedures for the development, validation, deployment, monitoring, and retirement of such systems. At a minimum, these policies shall address: (1) The roles and responsibilities of persons involved in developing, validating, and overseeing algorithmic systems; (2) testing requirements, including disparate impact testing before any new or materially modified system is deployed; (3) Ongoing monitoring requirements to detect model drift, performance degradation, or emerging disparate impact; (4) Procedures for investigating and remedying identified fairness issues; (5) Documentation requirements to maintain records sufficient to demonstrate compliance with this Chapter; and (6) Escalation procedures for significant fairness concerns identified by staff.

B 4 B. Covered insurers shall designate a senior officer with responsibility for AI governance and compliance with this Chapter.

C 5 C. Covered insurers shall train relevant staff on the requirements of this Chapter and on the principles of fair and responsible AI use in insurance.

Requires each covered insurer using an ADS to adopt written governance policies and procedures covering the full system lifecycle — development, validation, deployment, monitoring, and retirement. Minimum required elements include roles and responsibilities, pre-deployment disparate impact testing, ongoing monitoring for drift and emerging disparate impact, investigation and remediation procedures, documentation sufficient for compliance demonstration, and escalation procedures. Insurers must also designate a senior officer responsible for AI governance and train relevant staff on the Act's requirements.

Compliance actions 3 items

Covered insurers must adopt written governance policies and procedures for the development, validation, deployment, monitoring, and retirement of algorithmic decision systems, addressing roles and responsibilities, pre-deployment disparate impact testing, ongoing monitoring, investigation and remediation procedures, documentation, and escalation.

G-01.1

Covered insurers must designate a senior officer with responsibility for AI governance and compliance with this Act.

G-01.6

Covered insurers must train relevant staff on the requirements of this Act and on the principles of fair and responsible AI use in insurance.

G-01

R.S. 22:1407

Disparate Impact Audits

Deployer

A 6 A. Each covered insurer shall conduct an annual disparate impact audit of each algorithmic decision system used in insurance decisions affecting Louisiana consumers.

B(1)–(5) 6 B. The disparate impact audit shall: (1) Analyze insurance outcomes — including coverage denial rates, premium levels, claims payment rates, and claims denial rates — disaggregated by race, color, national origin, sex, age, disability status, and geographic area; (2) Apply statistical testing appropriate to the size and nature of the insurer's Louisiana book of business to determine whether disparate impact exists at a level of statistical significance; (3) Assess the contribution of each input variable to any identified disparate impact; (4) Evaluate whether less discriminatory alternatives exist that would achieve the insurer's actuarial objectives with reduced disparate impact; and (5) Document the methodology, findings, and any corrective actions taken or planned.

C 6 C. The disparate impact audit shall be conducted by a qualified independent actuary or data scientist. An actuary conducting an audit under this section shall be a Fellow of the Casualty Actuarial Society, a Fellow of the Society of Actuaries, or a Fellow of the American Academy of Actuaries. A data scientist conducting an audit under this section shall have demonstrated expertise in algorithmic fairness methods.

D 6 D. Where an insurer lacks sufficient Louisiana-specific data to conduct a statistically reliable audit, the insurer shall use national data supplemented by Louisiana-specific information to the extent available, and shall note this limitation in the audit report.

E(1)–(4) 7 E. Covered insurers shall file a Disparate Impact Audit Certification with the Commissioner annually, no later than June 30 of each year. The certification shall be signed by the insurer's senior AI governance officer and shall include: (1) A summary of each algorithmic system audited; (2) A statement of whether any disparate impact was identified; (3) If disparate impact was identified, a description of the corrective actions taken or a justification demonstrating actuarial necessity and absence of a less discriminatory alternative; and (4) A certification that the audit was conducted in accordance with this section.

F F. The Commissioner shall maintain a public registry of Disparate Impact Audit Certifications, redacted of confidential proprietary information as provided in Section 1413.

Requires each covered insurer to conduct an annual disparate impact audit of every ADS used in insurance decisions affecting Louisiana consumers. The audit must analyze outcomes disaggregated by race, color, national origin, sex, age, disability status, and geographic area; apply appropriate statistical testing; assess each input variable's contribution to identified disparate impact; evaluate less discriminatory alternatives; and document methodology, findings, and corrective actions. Audits must be conducted by a qualified independent actuary (FCAS, FSA, or FAAA) or data scientist with demonstrated algorithmic-fairness expertise. Insurers must file an annual Disparate Impact Audit Certification with the Commissioner by June 30, signed by the senior AI governance officer. The Commissioner must maintain a public registry of certifications, redacted for proprietary information.

Compliance actions 2 items

Covered insurers must conduct an annual disparate impact audit of each ADS, performed by a qualified independent actuary (FCAS/FSA/FAAA) or data scientist with algorithmic-fairness expertise, analyzing outcomes disaggregated by protected class and geographic area, assessing each variable's contribution, evaluating less discriminatory alternatives, and documenting methodology, findings, and corrective actions.

H-02.1

Covered insurers must file a Disparate Impact Audit Certification with the Commissioner annually by June 30, signed by the senior AI governance officer, including a summary of each system audited, whether disparate impact was found, corrective actions or actuarial-necessity justification, and a compliance certification.

H-02.4

R.S. 22:1408

Prohibited Variables

Deployer

A(1)–(5) 8 A. Notwithstanding any actuarial justification, the following variables shall not be used as input variables in algorithmic decision systems for homeowners insurance underwriting or rating in Louisiana: (1) Credit scores or credit-based insurance scores; (2) Race, color, religion, national origin, sex, disability, familial status, age, or sexual orientation of any consumer; (3) Membership in any protected class as defined in this Chapter; (4) Social media activity, social network connections, or online behavioral data; and (5) Consumer purchasing data from retail transactions.

B 9 B. For commercial lines insurance, use of the variables in subsection A(2) and A(3) is prohibited. Use of other variables in subsection A may be permitted upon demonstration to the Commissioner that the variable is actuarially justified and does not produce disparate impact.

C C. The Commissioner may, by rule, designate additional prohibited variables based on evidence of disparate impact or the absence of actuarial justification.

Categorically prohibits five types of input variables in algorithmic decision systems for homeowners insurance underwriting or rating, regardless of actuarial justification: credit scores, protected-class characteristics, social media/online behavioral data, and consumer purchasing data. For commercial lines, only protected-class characteristics are categorically prohibited; other enumerated variables may be used upon demonstration to the Commissioner of actuarial justification and absence of disparate impact. The Commissioner may designate additional prohibited variables by rule.

Compliance actions 2 items

Covered insurers must not use credit scores, protected-class characteristics, social media or online behavioral data, or consumer purchasing data as input variables in algorithmic decision systems for homeowners insurance underwriting or rating in Louisiana, regardless of actuarial justification.

D-01.5

For commercial lines insurance, covered insurers must not use protected-class characteristics as ADS input variables. Other enumerated presumptive proxy variables may be used only upon demonstration to the Commissioner of actuarial justification and absence of disparate impact.

D-01.5

R.S. 22:1409

Claims Adjustment AI Standards

Deployer

A–B(1)–(3) 10 A. No covered insurer shall use an algorithmic decision system to generate a final claims payment determination without human review and approval of each determination by a licensed claims professional. B. The human reviewer of an AI-assisted claims determination shall: (1) Have sufficient authority to override the algorithmic determination; (2) Have access to the specific factors driving the algorithmic determination; and (3) Document the basis for approval, modification, or override of each determination.

C 11 C. An insurer shall not use an algorithmic claims system to generate initial settlement offers designed to be lower than the expected value of the claim with the purpose of settling claims below their true value. Any algorithmic determination that an insurer intends to use as a settlement offer shall be based on objective property and liability valuation data.

D D. The Commissioner may conduct market conduct examinations of insurers' AI claims systems to assess compliance with this section and the overall fairness of claims outcomes by protected class.

Imposes requirements on AI-assisted claims adjustment. No final claims payment determination may be generated by an ADS without human review and approval by a licensed claims professional who has authority to override, access to the driving factors, and who documents the basis for each decision. Separately prohibits insurers from using algorithmic claims systems to generate initial settlement offers designed to undervalue claims — all algorithmic settlement offers must be based on objective property and liability valuation data. The Commissioner may conduct market conduct examinations of AI claims systems.

Compliance actions 2 items

Covered insurers must not generate a final claims payment determination using an ADS without human review and approval by a licensed claims professional who has authority to override the algorithmic determination, access to the specific factors driving it, and who documents the basis for each approval, modification, or override.

H-01.6

Covered insurers must not use an algorithmic claims system to generate initial settlement offers designed to undervalue claims. All algorithmic settlement offers must be based on objective property and liability valuation data.

CP-01

R.S. 22:1410

Disclosure of AI Use

Deployer

A–B 12 A. Every covered insurer shall disclose to consumers, at the time of application, renewal, or any adverse action, whether an algorithmic decision system was used in making the insurance decision affecting that consumer. B. The disclosure shall be clear, conspicuous, and understandable to a consumer without specialized knowledge, and shall appear in the following language or substantially similar language: "[Insurer name] uses automated or artificial intelligence systems to assist in insurance decisions. Your [application/renewal/claim] has been processed using such a system. You have the right to request an explanation of the factors that most influenced this decision."

C 13 C. When a covered insurer interacts with a consumer through an automated chatbot, virtual assistant, or similar automated system, the insurer shall clearly disclose to the consumer at the beginning of the interaction that they are communicating with an automated system and not a human representative.

Requires covered insurers to disclose to consumers at the time of application, renewal, or any adverse action whether an ADS was used. The disclosure must be clear, conspicuous, and understandable, and the bill prescribes specific model language. Separately, when a covered insurer interacts with a consumer through an automated chatbot, virtual assistant, or similar system, the insurer must disclose at the beginning of the interaction that the consumer is communicating with an automated system rather than a human.

Compliance actions 2 items

Covered insurers must disclose to consumers at the time of application, renewal, or any adverse action whether an ADS was used in the insurance decision, using prescribed or substantially similar clear and conspicuous language that includes notice of the right to request an explanation.

H-01.3

Covered insurers must clearly disclose to consumers at the beginning of any interaction through an automated chatbot, virtual assistant, or similar system that they are communicating with an automated system and not a human representative.

T-01.1

R.S. 22:1411

Right to Explanation

Deployer

A 14 A. Any consumer who is the subject of an adverse action involving the use of an algorithmic decision system has the right to request, within thirty (30) days of the adverse action, a written explanation of the decision.

B(1)–(5) 14 B. Upon receipt of a request for explanation, the covered insurer shall, within twenty-one (21) days, provide the consumer with: (1) A clear statement of the nature of the adverse action; (2) A plain language description of the top factors that most significantly contributed to the adverse action, ranked in order of their relative contribution; (3) A description of the data sources used in the decision; (4) Information about the consumer's right to correct inaccurate data; and (5) Information about the consumer's right to file a complaint with the Commissioner and to seek a review of the decision.

C–D 14 C. The explanation required by this section shall be specific to the individual consumer and shall not consist solely of generic descriptions of the algorithmic system. The explanation shall be sufficiently specific to enable the consumer to understand what, if anything, they could do differently to obtain a more favorable outcome. D. An insurer's assertion of trade secret protection for its algorithmic system does not relieve the insurer of its obligation to provide consumer-specific explanation under this section. The Commissioner shall by rule establish standards for providing meaningful explanation while protecting legitimate trade secrets.

Creates a consumer right to request a written explanation within 30 days of any adverse action involving an ADS. The insurer must respond within 21 days with a clear statement of the adverse action, the top contributing factors ranked by relative contribution, data sources used, and information about the consumer's right to correct data and file a complaint. The explanation must be specific to the individual consumer — generic system descriptions are insufficient. Trade-secret assertions do not relieve the insurer of the obligation to provide consumer-specific explanations.

Compliance actions 1 item

Covered insurers must, within 21 days of a consumer's request, provide a written, consumer-specific explanation of any adverse action involving an ADS, including the top contributing factors ranked by relative contribution, data sources used, the right to correct data, and the right to file a complaint. Trade-secret protection does not relieve this obligation.

H-01.1

R.S. 22:1412

Right to Human Review

Deployer

A 15 A. Any consumer who has received an adverse action based wholly or substantially on the output of an algorithmic decision system may, within thirty (30) days of receiving the adverse action notice or explanation, request human review of the decision.

B(1)–(3) 15 B. Upon receipt of a request for human review, the covered insurer shall: (1) Assign a qualified human underwriter or claims professional to independently review the decision; (2) Complete the review and provide a written determination within thirty (30) days of the request; and (3) Provide the consumer with a written explanation of the human reviewer's determination.

C 15 C. The human reviewer shall have full authority to modify or reverse the original adverse action. The human reviewer's determination shall be based on the specific facts of the consumer's situation and shall not be simply a reendorsement of the algorithmic output.

Grants consumers who receive an adverse action based wholly or substantially on ADS output the right to request human review within 30 days. The insurer must assign a qualified human underwriter or claims professional to independently review, complete the review within 30 days, and provide a written explanation. The human reviewer must have full authority to modify or reverse the adverse action and must base the determination on the consumer's specific facts — not merely re-endorse the algorithmic output.

Compliance actions 1 item

Covered insurers must, upon consumer request within 30 days of an adverse action, assign a qualified human underwriter or claims professional to independently review the decision, complete the review within 30 days, and provide a written explanation. The reviewer must have full authority to modify or reverse the adverse action based on the consumer's specific facts.

H-01.4

R.S. 22:1413

Data Accuracy and Correction

Deployer

A 16 A. Each covered insurer shall maintain reasonable procedures to ensure the accuracy of data used in algorithmic decision systems.

B(1)–(3), C 17 B. Any consumer has the right to: (1) Request that the insurer identify the data sources used in making an insurance decision affecting that consumer; (2) Dispute the accuracy of data used in an adverse action; and (3) Have inaccurate data corrected and the adverse action reconsidered. C. Upon receipt of a dispute regarding data accuracy, the covered insurer shall, within thirty (30) days, investigate the dispute and either correct the data and reconsider the decision or notify the consumer of the insurer's determination that the data is accurate.

Requires covered insurers to maintain reasonable procedures to ensure data accuracy in ADS. Grants consumers the right to identify data sources used in their insurance decision, dispute data accuracy in adverse actions, and have inaccurate data corrected and the adverse action reconsidered. Insurers must investigate disputes within 30 days.

Compliance actions 2 items

Covered insurers must maintain reasonable procedures to ensure the accuracy of data used in algorithmic decision systems.

D-01

Covered insurers must, upon consumer request, identify data sources used in an insurance decision, allow consumers to dispute data accuracy, correct inaccurate data and reconsider the adverse action, and investigate data disputes within 30 days.

D-01.1

R.S. 22:1414

Commissioner Authority and Rulemaking

A–C A. The Commissioner shall have authority to promulgate rules and regulations necessary to implement this Chapter, including rules governing: (1) Standards for disparate impact audits, including acceptable methodologies and statistical significance thresholds; (2) The format and content of Disparate Impact Audit Certifications; (3) Additional prohibited variables; (4) Standards for AI governance policies; (5) The format and content of consumer disclosures and explanations; and (6) Procedures for market conduct examinations related to AI systems. B. In promulgating rules under this Chapter, the Commissioner shall consider guidance from the National Association of Insurance Commissioners, National Council of Insurance Legislators, the Federal Trade Commission, and other relevant regulatory bodies. C. The Commissioner shall consult with the Louisiana Legislative Black Caucus, consumer advocacy organizations, civil rights organizations, and representatives of the insurance industry in developing rules under this Chapter.

Authorizes the Commissioner to promulgate rules implementing the Act, covering audit standards and methodologies, certification formats, additional prohibited variables, AI governance policy standards, consumer disclosure formats, and market conduct examination procedures. The Commissioner must consider NAIC, NCOIL, and FTC guidance, and must consult with the Louisiana Legislative Black Caucus, consumer advocacy organizations, civil rights organizations, and industry representatives. This section delegates regulatory authority and does not impose independent compliance obligations on insurers.

R.S. 22:1415

Market Conduct Examinations

Deployer

A–B(1)–(4) 18 A. The Commissioner may conduct market conduct examinations of covered insurers to assess compliance with this Chapter. B. As part of a market conduct examination, the Commissioner may require a covered insurer to: (1) Provide documentation of its AI governance policies and procedures; (2) Provide access to its algorithmic decision systems and the data used by those systems; (3) Provide complete disparate impact audit reports, not merely the certifications filed with the Commissioner; and (4) Demonstrate the explainability of its algorithmic systems.

C–D 18 C. The Commissioner may retain independent actuarial, data science, or technology experts to assist in market conduct examinations under this Chapter. The reasonable costs of such experts shall be borne by the insurer being examined. D. Market conduct examination reports relating to AI systems shall be confidential to the extent necessary to protect legitimate trade secrets, but the Commissioner shall publish aggregate findings of AI fairness examinations annually.

Authorizes the Commissioner to conduct market conduct examinations of covered insurers for Act compliance. The Commissioner may require documentation of AI governance policies, access to algorithmic systems and underlying data, complete audit reports (not merely certifications), and demonstrations of explainability. The Commissioner may retain independent experts at the insurer's expense. Examination reports are confidential to protect trade secrets, but aggregate findings must be published annually.

Compliance actions 1 item

Covered insurers must, upon request during a market conduct examination, provide the Commissioner with AI governance documentation, access to algorithmic systems and data, complete disparate impact audit reports, and demonstrations of system explainability. Insurers bear the cost of independent experts retained by the Commissioner.

R-02.2

R.S. 22:1416

Filing Requirements for New AI Systems

Deployer

A(1)–(5) 19 A. At least sixty (60) days before deploying a new algorithmic decision system, or a material modification of an existing system, a covered insurer shall file with the Commissioner a notice that includes: (1) A description of the system and its intended use; (2) A summary of predeployment disparate impact testing results; (3) A description of the data used to train and validate the system; (4) A description of the input variables used; and (5) A certification that the system complies with the requirements of this Chapter.

B–C 19 B. The Commissioner may, within sixty (60) days of receiving a notice under this section, request additional information or documentation from the insurer. Deployment of the system shall be stayed pending the Commissioner's response if the Commissioner requests additional information within fifteen (15) days of the notice. C. The Commissioner may prohibit deployment of a system or require modification of a system that the Commissioner determines does not comply with this Chapter.

Requires covered insurers to file a pre-deployment notice with the Commissioner at least 60 days before deploying a new ADS or materially modifying an existing system. The notice must include a system description, pre-deployment disparate impact testing results, training/validation data description, input variables, and a compliance certification. The Commissioner may request additional information within 60 days, which stays deployment if the request is made within 15 days. The Commissioner may prohibit or require modifications to non-compliant systems.

Compliance actions 1 item

Covered insurers must file a pre-deployment notice with the Commissioner at least 60 days before deploying a new ADS or materially modifying an existing system, including a system description, pre-deployment disparate impact testing results, training data description, input variables, and a compliance certification. Deployment is stayed if the Commissioner requests additional information within 15 days.

R-02.3

R.S. 22:1417

Annual Report on AI Insurance Fairness

A(1)–(5), B A. The Commissioner shall prepare and publish an annual report on AI Insurance Fairness in Louisiana. The report shall include: (1) Aggregate findings from Disparate Impact Audit Certifications, including the number of insurers reporting disparate impact and the steps taken to address identified disparities; (2) A summary of market conduct examination findings related to AI systems; (3) An analysis of insurance outcome data by race, color, and national origin for homeowners and commercial lines; (4) Recommendations for regulatory improvements; and (5) An assessment of emerging AI technologies and their potential impact on insurance fairness. B. The Commissioner shall submit the annual report to the Legislature and shall present findings to the House Committee on Insurance and the Senate Committee on Insurance.

Directs the Commissioner to prepare andpublish an annual report on AI insurance fairness, including aggregate audit certification findings, market conduct examination summaries, insurance outcome data analysis by race, color, and national origin, recommendations for regulatory improvements, and assessment of emerging AI technologies. The report must be submitted to the Legislature and presented to relevant committees. This section imposes obligations on the Commissioner, not on covered insurers.

R.S. 22:1418

Civil Penalties

A(1)–(5), B(1)–(5), C A. A covered insurer that violates any provision of this Chapter shall be subject to a civil penalty as follows: (1) For failure to conduct a required disparate impact audit: not less than twenty-five thousand dollars ($25,000) and not more than one hundred thousand dollars ($100,000) per violation; (2) For failure to provide required disclosures or explanations to a consumer: not less than five thousand dollars ($5,000) and not more than twenty-five thousand dollars ($25,000) per violation; (3) For use of a prohibited variable under Section 1408: not less than fifty thousand dollars ($50,000) and not more than two hundred fifty thousand dollars ($250,000) per violation; (4) For failure to conduct required human review of an adverse action: not less than ten thousand dollars ($10,000) and not more than fifty thousand dollars ($50,000) per violation; and (5) For failure to file required certifications or notices: not less than ten thousand dollars ($10,000) and not more than fifty thousand dollars ($50,000) per day of violation. B. In determining the amount of any civil penalty under this section, the Commissioner shall consider: (1) The severity of the violation and the extent to which consumers were harmed; (2) Whether the violation was knowing or willful; (3) The insurer's history of compliance with this Chapter and other insurance laws; (4) The insurer's good faith efforts to correct the violation; and (5) The financial size of the insurer C. Civil penalties collected under this section shall be deposited in the Louisiana AI Insurance Fairness Fund established by Section 1421.

Establishes a tiered civil penalty structure enforced by the Commissioner. Penalties range from $25,000–$100,000 per violation for failure to conduct a disparate impact audit, $5,000–$25,000 per violation for disclosure/explanation failures, $50,000–$250,000 per violation for use of prohibited variables, $10,000–$50,000 per violation for failure to conduct human review, and $10,000–$50,000 per day for failure to file required certifications or notices. Penalty amounts consider severity of harm, willfulness, compliance history, good-faith correction efforts, and insurer financial size. Collected penalties are deposited in the Louisiana AI Insurance Fairness Fund. This section describes enforcement consequences rather than new compliance duties.

R.S. 22:1419

Private Right of Action

A–E A. Any consumer who suffers harm as a result of a covered insurer's violation of this Chapter shall have a private right of action against the insurer in a court of competent jurisdiction in Louisiana. B. A consumer who prevails in an action under this section shall be entitled to: (1) Actual damages, including any economic loss attributable to the adverse action; (2) Statutory damages of not less than one thousand dollars ($1,000) and not more than ten thousand dollars ($10,000) per violation, regardless of actual damages; (3) Punitive damages, where the insurer's conduct was knowing, willful, or reckless; (4) Injunctive or declaratory relief; and (5) Reasonable attorneys' fees and costs C. An action under this section shall be filed within two (2) years of the date the consumer discovered or reasonably should have discovered the violation. D. The filing of a complaint with the Commissioner shall toll the statute of limitations for any related private action for the period during which the Commissioner's investigation is pending. E. Nothing in this section shall be construed to require exhaustion of administrative remedies before filing a private action.

Creates a private right of action for any consumer who suffers harm from a covered insurer's violation. Prevailing consumers may recover actual damages, statutory damages of $1,000–$10,000 per violation regardless of actual damages, punitive damages for knowing/willful/reckless conduct, injunctive or declaratory relief, and reasonable attorneys' fees and costs. Two-year discovery-based statute of limitations; filing a complaint with the Commissioner tolls limitations during the investigation. No exhaustion of administrative remedies is required.

R.S. 22:1420

Class Actions

A A. A named plaintiff in a class action under this section need not show individual actual damages as a prerequisite to maintaining the class action where the claims are based on a systematic pattern of discriminatory algorithmic practice.

Authorizes class actions for systematic patterns of discriminatory algorithmic practice. Named plaintiffs need not show individual actual damages as a prerequisite to maintaining a class action when claims are based on a systematic pattern. This section modifies the procedural framework for enforcement rather than creating a new compliance obligation.

R.S. 22:1421

Louisiana AI Insurance Fairness Fund

A–C(1)–(5) A. There is hereby created in the state treasury a special fund to be known as the Louisiana Artificial Intelligence Insurance Fairness Fund. B. The fund shall receive civil penalties collected under this Chapter and such other appropriations as the Legislature may provide. C. Monies in the fund shall be used to: (1) Fund the Commissioner's AI examination and oversight activities; (2) Provide technical assistance and education to consumers regarding their rights under this Chapter; (3) Provide grants to legal aid organizations for representation of consumers in actions under this Chapter; (4) Fund research on AI insurance fairness; and (5) Cover administrative costs of the AI Fairness Advisory Board established by Section 1422.

Creates the Louisiana Artificial Intelligence Insurance Fairness Fund in the state treasury, funded by civil penalties and legislative appropriations. Monies are used for Commissioner AI examination and oversight activities, consumer technical assistance and education, legal aid grants, AI insurance fairness research, and Advisory Board administrative costs. This section establishes a funding mechanism and creates no compliance obligations for insurers.

R.S. 22:1422

AI Insurance Fairness Advisory Board

A–C A. The Commissioner shall establish an AI Insurance Fairness Advisory Board to advise the Commissioner on the implementation and development of this Chapter. B. The Advisory Board shall consist of fifteen (15) members appointed by the Commissioner, including: (1) Three representatives of the insurance industry; (2) Two consumer advocates; (3) Two civil rights advocates, at least one of whom shall be nominated by the Louisiana Legislative Black Caucus; (4) Two representatives of minority business organizations; (5) Two data scientists or AI fairness experts; (6) One actuary; (7) One legal scholar with expertise in anti-discrimination law; (8) One representative from a Louisiana law enforcement or criminal justice agency; and (9) One member of the general public. C. The Advisory Board shall meet at least quarterly and shall advise the Commissioner on rules, examination priorities, and annual report findings.

Establishes a 15-member AI Insurance Fairness Advisory Board appointed by the Commissioner, including insurance industry, consumer, civil rights, minority business, data science, actuarial, legal, law enforcement, and public representatives. At least one civil rights advocate must be nominated by the Louisiana Legislative Black Caucus. The Board meets quarterly and advises the Commissioner on rules, examination priorities, and annual report findings. This section imposes no compliance obligations on insurers.

R.S. 22:1423

Relation to Other Laws

A–C A. This Chapter supplements and does not supersede existing state and federal anti-discrimination laws applicable to insurance, including but not limited to the federal Fair Housing Act, the Equal Credit Opportunity Act, and Louisiana's unfair trade practices laws. B. Nothing in this Chapter shall be construed to limit or restrict the Commissioner's existing authority to prohibit unfair discrimination in insurance under Louisiana law. C. Where any provision of this Chapter conflicts with a provision of another state law, the provision that affords greater protection to consumers shall apply.

Savings clause providing that the Act supplements and does not supersede existing state and federal anti-discrimination laws (Fair Housing Act, Equal Credit Opportunity Act, Louisiana unfair trade practices laws). Does not limit the Commissioner's existing authority. Where provisions conflict with other state law, the provision affording greater consumer protection applies. Creates no new compliance obligations.

R.S. 22:1424

Severability

If any provision of this Chapter or its application to any person or circumstance is held invalid, the invalidity does not affect other provisions or applications which can be given effect without the invalid provision or application, and to this end the provisions of this Chapter are severable.

Standard severability clause. Creates no compliance obligations.

R.S. 22:1425

Effective Date; Phased Implementation

A–C A. This Act shall take effect January 1 of the year following its enactment. B. The following provisions shall be implemented on a phased schedule: (1) Section 1410 (AI disclosure requirements) and Section 1411 (right to explanation) shall be effective six (6) months after enactment; (2) Section 1407 (disparate impact audits) and Section 1416 (predeployment filing) shall be effective twelve (12) months after enactment; and (3) Section 1408 (prohibited variables) shall be effective eighteen (18) months after enactment, to allow insurers time to modify their systems to comply. C. The Commissioner shall publish implementing guidance and model compliance frameworks for each phase of implementation at least ninety (90) days before the effective date of each phase.

Establishes phased implementation. The Act generally takes effect January 1 of the year following enactment. Disclosure and right-to-explanation provisions (§§ 1410–1411) are effective 6 months after enactment. Disparate impact audits and pre-deployment filing (§§ 1407, 1416) are effective 12 months after enactment. Prohibited variables (§ 1408) are effective 18 months after enactment. The Commissioner must publish implementing guidance at least 90 days before each phase's effective date. This section creates no independent insurer compliance obligations.

Transparency AI identity disclosure · periodic reminders

1 item

T-01.1

Data Controls Data sale restrictions · deidentification

4 items

D-01.5

D-01

Covered insurers must maintain reasonable procedures to ensure the accuracy of data used in algorithmic decision systems.

D-01.1

Deceptive Conduct AI-generated likeness · consent requirements

1 item

CP-01

Human Oversight Decision review · appeal pathways

7 items

H-02.1

H-02.4

H-01.6

H-01.3

H-01.1

H-01.4

Governance Risk management · accountability programs

4 items

G-01

Covered insurers must contractually require third-party model providers to supply the information and access necessary for the insurer to fulfill its obligations under this Act.

G-01.1