An open standard for quantifying AI security risk in entities subject to insurance underwriting. Five weighted scoring domains. Twenty-five discrete factors. Five rating tiers — T1 AI Insurance Ready through T5 Uninsurable — that map directly to underwriting decisions. Free to read, cite, and implement.
The AI Insurance Readiness Score (AIRS) is an open standard specification for quantifying artificial intelligence security risk in entities subject to insurance underwriting. AIRS defines five weighted scoring domains encompassing twenty-five discrete risk factors, a composite scoring methodology producing scores on a 0–100 scale, and five rating tiers that map directly to underwriting decisions. The specification provides the complete methodology, scoring rubrics, rating tier definitions, assessment process requirements, and conformance criteria necessary for institutional adoption by carriers, reinsurers, regulators, and enterprise risk teams.
Published by the AIRS Standards Body · April 2026 (v1.0) · Revised May 2026 (v1.1) · Editorial stewardship and methodology administered by AI Security Intelligence LLC · Jurisdiction: United States of America
Every component of AIRS v1.1 is openly published and free to use. Choose where to start, or move between them as your evaluation requires.
The complete 70-page open standard. Five weighted domains, twenty-five risk factors, five rating tiers (T1 AI Insurance Ready through T5 Uninsurable), the domain-floor rule, full conformance criteria, and the assessment process. The canonical reference.
Read the SpecificationDomain architecture, tier logic, the evidence model, and the scoring approach behind defensible institutional output. How AIRS was constructed, and why each design choice produces underwriting-grade results.
Explore the MethodologyAIRS controls mapped factor-by-factor to the NIST AI Risk Management Framework, ISO/IEC 42001, the EU AI Act, and the NAIC Model Bulletin. For organizations evaluating AIRS alongside the frameworks they already comply with.
View the CrosswalksEvaluate your organization against all twenty-five AIRS factors. Receive your composite score, tier classification, and prioritized remediation guidance. Free. No registration required.
Begin the Self-AssessmentHow AIRS is designed for carrier and reinsurer adoption — pricing input, exclusion triggers, mandatory remediation thresholds, and integration patterns with existing underwriting workflows.
Read the Carrier BriefThe market thesis behind the standard — why AI-dependent insureds require a uniform readiness framework, how AIRS aligns with NIST AI RMF and ISO/IEC 42001, and the rationale for an open, publicly maintained specification.
Read the WhitepaperEach domain captures a distinct dimension of AI insurability and is composed of five discrete risk factors, evaluated on a five-point maturity scale. Domain weights reflect the relative contribution of each area to overall AI insurability.
Training data provenance, adversarial robustness testing, model versioning & rollback, poisoning detection, drift monitoring.
Human-in-the-loop oversight, hallucination controls, bias & fairness testing, output auditability, content provenance.
Model card & vendor documentation, third-party AI risk, API key governance, foundation-model dependency mapping, sub-processor controls.
NIST AI RMF alignment, ISO/IEC 42001 conformance, EU AI Act readiness, state AI law monitoring, GDPR/CCPA AI-specific controls.
AI incident response plan, recovery time objectives, AI-specific tabletop exercises, inter-system failure modes, business continuity for AI-dependent workflows.
AIRS is designed to be useful at the underwriting desk, in the regulator's review file, and inside the enterprise risk function — each with a different entry point and a different application.
AIRS produces a defensible, auditable score that maps directly to underwriting decisions — broadest coverage, conditional terms, mandatory remediation, or denial. Built to integrate with existing risk engineering and underwriting workflows, not replace them.
An openly published, vendor-neutral methodology aligned with the NAIC Model Bulletin on Artificial Intelligence, the NIST AI Risk Management Framework, and ISO/IEC 42001 — usable as a reference standard in market conduct review, examination programs, and rulemaking.
Use AIRS to benchmark internal AI security maturity against the same criteria carriers use to underwrite. Identify the factors most likely to drive coverage decisions before renewal, and prioritize remediation against an institutionally legible standard.
AIRS v1.1 is openly published under attribution terms designed to support institutional adoption. Organizations may freely reference, implement, and build upon the methodology for internal risk assessment, underwriting, regulatory compliance, and academic research. The specification is aligned with — and explicitly crosswalked against — the prevailing AI governance frameworks.
70 pages. Full methodology, scoring rubrics, five rating tiers (T1 AI Insurance Ready through T5 Uninsurable), the domain-floor rule, assessment process, and regulatory crosswalk. Free. No form. No email required.
Download AIRS v1.1 (PDF, 194 KB)