Book a demo
Topics · GOVERNANCE-AGENTS

Insurance Agent Kill Switch NAIC | Moring

The NAIC Model Bulletin requires insurers to adopt, implement, and maintain a documented AIS Program, and a kill switch is what makes the program enforceable at runtime. The bulletin expects the rigor applied to cybersecurity, financial controls, privacy, and operational risk to be applied to AI as well, because the technology does not change the underlying legal obligation.

On this page
  1. Short answer
  2. How it works
  3. FAQs
  4. Related reading
  5. Sources

Short answer

The NAIC Model Bulletin requires insurers to adopt, implement, and maintain a documented AIS Program, and a kill switch is what makes the program enforceable at runtime. The bulletin expects the rigor applied to cybersecurity, financial controls, privacy, and operational risk to be applied to AI as well, because the technology does not change the underlying legal obligation.

How it works

The NAIC Model Bulletin on the Use of Artificial Intelligence Systems by Insurers is now adopted by roughly 24 states, and it requires insurers to adopt, implement, and maintain a documented AI program, an AIS Program, to support responsible AI use and mitigate the risk of inaccurate or discriminatory decisions, particularly when AI is used in regulated processes like underwriting, pricing, and claims. The bulletin's foundational position is that the technology does not change the legal obligation, and the same rigor that governs cybersecurity, financial controls, privacy, and operational risk belongs on AI.

The kill switch is what makes the AIS Program enforceable at runtime. The NAIC bulletin expects a written program with governance, risk management, and internal controls, and a kill switch that has never been tested is a document, not a control. The bulletin's seven workflows where insurers deploy agentic AI, from submission intake to servicing after the sale, each get the same five rows: the decision the agent influences or takes, the data it touches, its failure modes, the controls that contain them, and the regulatory exposure. The highest-exposure action is the silent decline: a submission scored out of appetite never reaches an underwriter, so no person sees the error.

For a carrier, the kill switch is what makes the AIS Program enforceable, and it is the runtime expression of the NAIC expectation of internal controls. A lightweight sidecar can add policy, human approval, decision lineage and a kill switch, with no change to the agent's code.

An enrolled agent cannot run until a policy allows it, and new agents start in a sandbox. The bulletin's Section 4 provides guidance on the kinds of information and documents that the Department may request in the context of an AI-focused investigation, including a market conduct action, and the kill switch is the control that produces the evidence.

The failure modes are predictable. The switch lives inside the agent and cannot be trusted. Stopping the agent requires a redeploy. Policy is not managed centrally, so a change takes too long to reach every enforcement point. Or the response is a whole-system shutdown that halts claims queues and creates a secondary incident, so the agent keeps running at full access while the team figures out what is wrong. The remedy is graduated containment: a dimmer switch, not only a kill switch, with read-only on certain data first, sensitive tools dropped next, and higher approval thresholds for anything above a defined size, each adjustment reversible and logged.

Frequently asked questions

What does the NAIC Model Bulletin require of an insurance carrier's kill switch?

The NAIC Model Bulletin, adopted by roughly 24 states, requires insurers to adopt, implement, and maintain a documented AI program, an AIS Program, to support responsible AI use and mitigate the risk of inaccurate or discriminatory decisions. The kill switch makes the AIS Program enforceable at runtime, and it is the runtime expression of the NAIC expectation of internal controls. A kill switch that has never been tested is a document, not a control, and the bulletin expects records sufficient to show the program operates and to produce to regulators on request.

How does the NAIC bulletin frame the AIS Program in insurance terms?

The bulletin expects a written program with governance, risk management, and internal controls, and it directs insurers to bring the same rigor they apply to cybersecurity, financial controls, privacy, and operational risk to AI, per the bulletin. The bulletin's foundational position is that the technology does not change the legal obligation. The seven workflows where insurers deploy agentic AI, from submission intake to servicing after the sale, each get the same five rows: the decision the agent influences or takes, the data it touches, its failure modes, the controls that contain them, and the regulatory exposure.

Why is the silent decline the highest-exposure action for an insurance intake agent?

Because a submission scored out of appetite never reaches an underwriter, so no person sees the error. The kill switch stops the agent when its behavior changes, and the NAIC bulletin expects a documented AIS Program with the same discipline applied to cybersecurity, financial controls, privacy, and operational risk. A lightweight sidecar can add policy, human approval, decision lineage and a kill switch, with no change to the agent's code, and an enrolled agent cannot run until a policy allows it, with new agents starting in a sandbox.

What evidence does the NAIC bulletin expect a carrier to produce in a market-conduct examination?

Section 4 of the bulletin provides guidance on the kinds of information and documents that the Department may request in the context of an AI-focused investigation, including a market conduct action. The kill switch is what produces the evidence: the audit trail of agent actions and the termination event itself, the tamper-evident record of agent actions and the termination event itself, supporting post-incident review and regulatory documentation. The bulletin expects records sufficient to show the AIS Program operates, and the kill switch is what proves it.

Sources

  1. content.naic.org
  2. content.naic.org
  3. vijil.ai
  4. compassmsp.com
  5. lumenova.ai
  6. criticalbridge.ai
  7. forvismazars.us
  8. fiddler.ai
  9. - artificialintelligenceact.eu
  10. - eur-lex.europa.eu