Following completion of its 2022 educational paper, the NAIC’s Accelerated Underwriting (A) Working Group (AUWG) moved to the second part of their charge — draft guidance for the states. During their call on February 22, the AUWG discussed their:

The AUWG Draft Regulatory Guidance Document

The AUWG Draft Regulatory Guidance Document contains two sections:

The regulatory guidance is composed of three parts:

1. Nine “regulatory considerations,” which the draft also refers to as “regulatory expectations.”

Four expectations relate to consumer data, algorithms, or predictive models used in accelerated underwriting. Among other things, these considerations require that the data being used is evaluated for “unfair bias,” that the data used and the decisions made are based upon sound actuarial principles, and that the algorithms and predicative models “accurately assess and price risk” and achieve an outcome that is not “unfairly discriminatory.”

Three expectations relate to the insurer’s policies and procedures regarding the consumer data being collected and used, including the ability of consumers to correct data, receive notices, and opt out of data sharing or otherwise restrict the use of data.

One expectation is that consumers are provided with all information upon which the insurer based an adverse underwriting decision.

One expectation is that the insurer will produce information to regulators upon request or in connection with market conduct examinations.

2. Five proposed “regulatory actions.”

Four proposed regulatory actions relate to requests for information that a regulator may make. These include information on the data sources, predictive models, and algorithms, including how they are used in an accelerated underwriting program and whether they are audited by the insurer to “ensure they are accurate, reliable, and do not result in unfairly discriminatory outcomes.”

One proposed regulatory action relates to the review of a life insurer’s “initial submission of policy filings to confirm the proper use of data elements.”

These five proposed regulatory actions raise a number of questions, including:

3. Fourteen example “regulatory questions and requests” to life insurers.

Seven questions relate to the insurer’s auditing of the data sets, predictive models, and machine learning algorithms to ensure accuracy, reliability, and outcomes that are not unfairly discriminatory and that the models “are based on sound actuarial principles” and “ensur[e] that external consumer data’s correlation to risk is not outweighed by any correlation to a protected class(es).”

Four questions relate to disclosures to consumers and consumers’ rights in the event of an adverse underwriting decision, including the right to correct mistakes in the external data. 

Three questions relate to the external data or information being used by the accelerated underwriting program and how such data is utilized, stored, and destroyed after the completion of the underwriting process.

The concluding paragraph of the regulatory guidance makes clear that additional guidance will be needed from other NAIC committees and working groups.

AUWG Draft Referral to Market Conduct Examination Guidelines (D) Working Group

The draft referral includes a recommendation that the NAIC’s Market Regulation Handbook include a new standard in Chapter 23 — Conducting the Life and Annuity Examination — related to a life insurer’s use of big data, artificial intelligence, and machine learning to underwrite life insurance. According to the AUWG, the applicable standard should address how accelerated underwriting programs are fair, transparent, and secure. Under the AUWG’s recommendation, examiners would review “policy rates and forms, accelerated underwriting models and/or summaries of those models, information about source data used as part of the accelerated underwriting program, consumer disclosures, and testing and/or auditing policies and procedures of the models.”

We will continue to monitor AUWG activities, and other NAIC groups with respect to their review of insurers’ use of consumer data, algorithms, and machine learning. 

DISCLAIMER: Because of the generality of this update, the information provided herein may not be applicable in all situations and should not be acted upon without specific legal advice based on particular situations.

© Carlton Fields | Attorney Advertising

Copyright © JD Supra, LLC