AI in Life Insurance Underwriting and the New Regulatory Test

Life insurance underwriting is where predictive modeling meets mortality. For decades, the industry built its risk selection around medical exams, lab tests, prescription history, and actuarial tables. The process worked, but it was slow: the timeline from application to issued policy could run to a few months, and insurers came to treat that wait as an obstacle to the sale ¹. Accelerated underwriting was supposed to fix that. It uses external data and predictive models to shorten the timeline from weeks to hours, sometimes without a medical exam or fluids.

The trade-off is governance. Every input that replaces an exam or a question is a new potential source of unfair discrimination. Regulators have been waiting for this. The NAIC has long warned that the use of big data and external consumer data in underwriting can produce outcomes that correlate with race, income, geography, or disability even when no protected class is an explicit input. New York’s Insurance Circular Letter No. 7 (2024) answered it with a procedure rather than a ban: measure the correlation, then defend what survives it ². Colorado has gone further with its own rules, making the issue bigger than one state.

Life underwriting is now the test case for whether AI in insurance can be both fast and fair.

What accelerated underwriting does

Accelerated underwriting is the practice of issuing life insurance without the traditional medical exam, blood work, and urine sample, relying instead on data from external sources and predictive models. The NAIC describes it as a response to consumer demand for faster digital service, with insurers using prescription drug history, motor vehicle records, credit attributes, and other data to process applications in hours ¹.

The business case is straightforward. A shorter application process reduces the drop-off rate, increases sales, and lowers administrative costs. Consumers get coverage faster. Carriers get more policies on the books. LIMRA’s 2024 read of the industry put the hard part not in the technology but in the process: adopting AI is “not only about adopting new technology, but also reengineering business processes to fully leverage its capabilities,” and the technology is “doing more than just enhancing existing workflows” ³. Accelerated underwriting is one of the clearest examples of that reengineering. The broader underwriting AI governance framework, including claims and P&C dimensions, is covered in AI in underwriting insurance.

But the same external data that makes the process fast also makes it opaque. An applicant may not know that a prescription history record, a traffic violation, or a third-party risk score influenced the decision. The underwriter may not know exactly how the model combined those inputs. The examiner’s question is whether the carrier can explain the decision and prove that the explanation is fair.

The data sources that matter

The inputs in accelerated underwriting fall into four categories. Each carries its own regulatory risk.

Prescription history is the most common and usually the least controversial. It is directly related to the risk being insured. But it can still embed disparities if the data is incomplete, outdated, or influenced by factors unrelated to mortality, such as geographic access to care.

Motor vehicle records are used because they correlate with risky behavior. A history of serious traffic violations is plausibly related to mortality. But the same record may also correlate with income, geography, or disability, depending on how it is scored.

Credit attributes and financial data are more sensitive. The relationship between credit and mortality is statistical, but the mechanism is disputed. Regulators have long questioned whether credit-based scoring is a proxy for race or income. New York asks insurers to evaluate that correlation and justify what they keep; Colorado requires the testing by regulation and wants the results reported.

Third-party risk scores and black-box models are the highest-risk category. A carrier may license a score that combines dozens of data sources and produces a single mortality risk indicator. The carrier may not know how the score was built or what data it uses. That is a problem when the regulator asks what the score does to protected classes and who checked. The letter treats that as the carrier’s question to answer: an insurer “may not rely solely on a third-party’s claim of non-discrimination or a proprietary third-party process to determine compliance with anti-discrimination laws” ².

The rule is the same across all four categories. If the data is not grounded in the actual risk being insured, it is a proxy discrimination risk waiting to happen.

One New York wrinkle cuts across that list, and it is easy to get backwards in either direction. The proxy assessment applies to external consumer data and information sources, and the letter names four things it does not include: “an MIB Group, Inc. member information exchange service, a motor vehicle report, prescription drug data, or a criminal history search” ². Two of the four categories above therefore sit outside that assessment in New York. That is not the same as a pass on fairness work. A model built on prescription history is still an artificial intelligence system, and the letter’s second test, the one that looks at outputs rather than inputs, reaches external data and AI systems alike. Criminal history searches carry a separate New York obligation of their own under Executive Law § 296(16) ².

Where the proxy question enters a life decision

New York separates an input question from an outcome question. Its ECDIS proxy assessment concerns external consumer data as the letter defines it. Its assessment of disproportionate adverse effects reaches both ECDIS and AI systems.² The exact sequence, legal wording, notice boundary, and annual alternative search belong to the Circular Letter 7 guide, and the New York state page records where the letter and the NAIC model bulletin diverge.

Life teams still need to explain the local evidence seam. What traditional evidence did the accelerated program replace? Why is the new source relevant to mortality? Which applicants fall out of the accelerated path? Does referral to full underwriting change access, time, or outcome by cohort? Those questions cannot be answered by copying the New York procedure into a life workflow.

Colorado’s insurance rules create a separate filing and governance context for covered lines.⁴ That regime should be mapped as a state delta, not merged into New York’s test, and the Colorado state page shows how far it runs from the NAIC model bulletin. The governance hub explains how to separate portable controls from state-specific filings.

Three predictable failure points

Accelerated underwriting creates risk in three predictable places.

The model that no one owns. A predictive model licensed from a vendor may sit between the application and the underwriter without a clear internal owner. If the model’s accuracy drifts, or if its inputs change, the carrier may not notice until a regulator points it out. The NAIC Model Bulletin expects the insurer, not the vendor, to be accountable for the AI systems it uses ⁵.

The override that is not real. A model may be labeled as decision-support, but if the underwriter accepts its recommendation almost every time, the human is not really reviewing. The regulator will look at the override rate, the reasons given, and whether the underwriter had enough information to disagree. A system that is theoretically overridable but practically determinative is the highest-risk configuration.

The use case that expands. A model built for accelerated underwriting of healthy applicants may be reused for broader risk classes without revalidation. A model trained on one distribution of applicants may fail when applied to a different age band, geography, or product. Monitoring must catch this before the regulator does.

Four life-specific records

A life program should connect four records that generic governance language cannot supply:

• the traditional evidence replaced and the reason the substitute is relevant to mortality;
• the accelerated-program eligibility and referral rules, including what sends a case to full underwriting;
• the model or score version used for the applicant and the underwriter’s final action;
• cohort outcomes for accelerated, referred, and traditionally underwritten applicants.

Fairness concepts and alternatives analysis stay in the disparate-impact guide, and vendor diligence stays in the vendor playbook.

The decision evidence pack shows how to connect those records for one application without recreating either control here.

Before the next regulatory request

Select one accelerated program and trace an applicant through accelerated issue, referral, or decline. Confirm that the source, score version, rule, underwriter action, notice, and final outcome can be connected. Then compare the three paths at cohort level. The inventory playbook identifies the systems; the life team supplies the mortality and program logic that makes the record meaningful.

Life as the precedent line

The line-by-line version of these questions lives in AI use cases in insurance by business line.

Life underwriting offers a mature set of questions about external data, accelerated decisions, mortality rationale, and outcome differences. Those questions are useful when another line begins its review. They do not establish compliance elsewhere: each product, decision, jurisdiction, and governing rule still needs its own analysis.

Footnotes

1. NAIC, “Accelerated Underwriting,” Insurance Topics, updated April 2026: https://content.naic.org/insurance-topics/accelerated-underwriting ↩ ↩²

2. New York Department of Financial Services, “Insurance Circular Letter No. 7 (2024): Use of Artificial Intelligence Systems and External Consumer Data and Information Sources in Insurance Underwriting and Pricing,” July 11, 2024: https://www.dfs.ny.gov/industry-guidance/circular-letters/cl2024-07 ↩ ↩² ↩³ ↩⁴ ↩⁵

3. LIMRA, “The AI Industry Today: Understanding the Current State of Play,” 2024: https://www.limra.com/globalassets/limra-loma/trending-topics/ai-governance-group/the-ai-industry-today---understanding-the-current-state-of-play.pdf ↩

4. Colorado Division of Insurance, “SB21-169: Protecting Consumers from Unfair Discrimination in Insurance Practices,” 2025: https://doi.colorado.gov/for-consumers/sb21-169-protecting-consumers-from-unfair-discrimination-in-insurance-practices ↩

5. NAIC, “Use of Artificial Intelligence Systems by Insurers,” Model Bulletin adopted December 4, 2023: https://content.naic.org/sites/default/files/inline-files/2023-12-4%20Model%20Bulletin_Adopted_0.pdf ↩