When can the use of AI, machine learning, or robotic process-based insurance models result in adverse action under the FCRA? | Newsletters | Legal news: Insurance

As insurers consider augmenting the quote process with algorithmic predictive models, including those aided by artificial intelligence, machine learning, and/or robotic process automation (“Models“”FCRA“) dictates the distribution of a Notice of Adverse Action when a model is not implemented for the purpose of rendering “hedging and rating decisions” (determining whether to accept or reject a particular risk or the premium charged), but rather for the purpose of determining whether other measures can be taken with regard to consumers, such as directing applicants to certain payment methods or other designations unrelated to coverage and pricing decisions (“administrative decisions”).

Under the FCRA, “adverse action” can mean different things in the context of different industries or uses. As part of the insurance, a “adverse action“is defined as meaning “a denial or cancellation, increase of any charge or reduction or other adverse or unfavorable change in the conditions of cover or the amount of any insurance, whether existing or applied for, in connection with the purchase of insurance. “1 Under another FCRA section, “If a person takes adverse action against a consumer that is based in whole or in part on information contained in a consumer report,” that person must, among other things, providing adverse action notice to the consumer.2

A “consumer report” is defined as “any written, oral or other communication of any information by a consumer reporting agency relating to credit worthiness, solvency, creditworthiness, character, general reputation, personal characteristics or mode of life of a consumer that is used or likely to be used or collected in whole or in part for the purpose of serving as a factor in establishing the consumer’s eligibility for . . . (A) credit or insurance to be used primarily for personal, family or household purposes; Where . . . (C) any other permitted purpose [as a permissible purpose of consumer reports]”3 The “permitted purposes“Consumer reports include, in relevant part, the provision of a consumer report by a consumer reporting agency “to a person whom it has reason to believe. . . intends to use the information in the context of underwriting insurance involving the consumer.”4

1. “Adverse action”

First, insurers should consider whether an administrative decision could be considered “[1] an increase in any charge for . . . or any other unfavorable or unfavorable change in the terms of coverage. . . applied for, [2] in the context of insurance underwriting.

1.A “An increase in any charge for . . . or any other unfavorable or unfavorable change in the terms of cover . . . applied for”

An administrative ruling could be viewed as an increase in coverage charges, as applicants subject to an administrative ruling might place more value for the same level of coverage in one way or another. This additional value could be so small as to appear nominal, but could theoretically be interpreted as an “increase”.

An administrative decision could be considered an unfavorable or unfavorable change in the terms of coverage, as the burden of having to pay the premium in a different way or obtaining or interacting with their coverage in a different way could be interpreted as “unfavorable or unfavourable” the applicant’s point of view. In many circumstances, especially those involving applicants with fewer resources, paying more at once or in a different way could mean that the applicant has fewer funds to contribute towards other needs. An administrative decision could therefore be considered “adverse” or “adverse”.

1.B “As part of taking out insurance”

Depending on the nature of the administrative decision, it could be interpreted as being taken in the context of taking out insurance. The only permitted purpose for which a consumer report may be provided to an insurer is “to use the information in connection with purchasing insurance”. Further, it seems counterintuitive that the FCRA’s legislative intent is to permit the provision of consumer reports without attaching related restrictions and obligations such as the FCRA’s adverse action requirements.

2. Adverse Action Notice

As noted above, according to the FCRA, if a person takes an adverse action with respect to a consumer that is based in whole or in part on information contained in a consumer report,” the person must, among other , provide adverse action notice to the consumer.5 Insurers must therefore consider whether an administrative decision can be interpreted as being (1) based in whole or in part on (2) any information contained in a consumer report.

2.A “Based in whole or in part on”

The expression “based in whole or in part on” has been interpreted to apply only in the case of a causal relationship “in the absence of”. An adverse action is not considered to be based in whole or in part on the consumer’s report unless “the report is a necessary condition” of the adverse action.6

In some case law, the baseline or benchmark for determining whether there has been an adverse rate increase (and, therefore, adverse action requiring notice to the plaintiff) has been interpreted as “what the plaintiff would have had if the company had not taken its own[/her] credit score into account.”7 It may be that the sole purpose of a model’s use of a consumer report is to determine whether an administrative decision will be made. In this case, the “baseline” could be considered as the absence of an outcome of the administrative decision. In other words, without the use of the model that incorporates the consumer report, there would be no possibility of the administrative decision having an impact on the claimant.

2.B “Any information contained in a consumer report”

An insurer should analyze whether particular information used in a model was obtained from a consumer reporting agency based on the authorized purpose of the insurer. An insurer should also analyze whether the information is: (i) a written communication of information from a consumer reporting agency; (ii) affecting the creditworthiness, solvency, creditworthiness, character, general reputation, personal characteristics or lifestyle of a consumer; (iii) that is or should be used or collected in whole or in part for the purpose of serving as a factor in determining consumer eligibility for insurance to be used primarily for personal, family or household purposes.

3. State Insurance Rating Laws and NCOIL Model Law

Finally, an insurer should determine whether the above analysis would differ or whether additional considerations arise from state insurance rating laws enacted based on the National Council of Insurance Legislators Model Law on Insurance Rating. use of credit information in personal insurance (“Model NCOIL”). The NCOIL model defines what constitutes an “assurance score” (which is similar to the FCRA’s consumer report definition), what constitutes an “adverse action” with respect to these assurance scores (which is similar to the FCRA’s definition of an adverse action), and when an adverse action notice should be sent with respect to such adverse action (the trigger language of which is similar to the FCRA’s trigger language). This analysis will depend on the state-specific implementation of the NCOIL Model (if any) or other related state laws and regulations dealing with this subject (for states that have not adopted some form of the NCOIL Model ).

Of course, when analyzing these issues, insurers should consult thoroughly with insurance and federal regulatory guidance on the specific nature of administrative decisions, how models are created and used, and the impact of these administrative rulings and models on claimants and consumers. .



1 15 USCA § 1681a(k)(1)(B)(i).

2 15 USCA § 1681m(a).

3 15 USCA § 1681a(d)(1)(A) and (C).

4 15 USCA § 1681b(a)(3)(C).

5 15 USCA § 1681m(a).

6 Safeco Ins. Co. of Am. vs. Burr, 551 US 47, 63, 127 S. Ct. 2201, 2212, 167 L. Ed. 2d 1045 (2007). This case is also sometimes called Geico versus Edo.

Sherry J. Basler