The Open Standards
Two published open specifications. One engine-live standard in active development. The assessment infrastructure carriers, reinsurers, and security leaders are building on.
AI Insurance Readiness Score
Version 1.1 — PublishedCarrier-grade scoring across five domains: Model Integrity, Operational Lifecycle, Supply Chain & Substrate, Risk & Controls, and Surface & Reach. The scoring layer the insurance industry is underwriting against.
Read the AIRS SpecificationVulnerability Scoring System
Version 1.0 — PublishedPublic severity scoring for AI-specific vulnerabilities. Reproducible, vendor-neutral, and engineered to sit alongside CVSS for the categories CVSS cannot address — prompt injection, model exfiltration, training-data poisoning, agent escape.
Read the VSS WhitepaperThreat Scoring System
Open Standard · Coming v1.1Adversarial threat severity for AI systems. Captures attacker intent, capability, and accessibility across the AI attack surface — providing a common language for threat intelligence teams, red teams, and the carriers underwriting them. The open specification is in active development; the scoring engine is live in the ASI Market Index.
Explore the Threat Scoring EngineThe Open Spec Process
Every ASI standard is developed in the open: published specifications, versioned releases, public comment cycles, and a documented lineage from threat model to scoring formula. We publish what we measure and why — so carriers, regulators, and security leaders can audit the work, not just the score.
Scoring Methodology → Market Taxonomy → Disclosure Policy → Responsible AI Principles →