AI-IRF Documentation Center
Operational AI incident-response documentation, machine-readable controls and practitioner guidance.
AI-IRF v1.0
AI-IRF is ODA3 Institute's operational incident-response framework for AI systems. Version 1.0 defines 79 controls across 19 domains, supported by public machine-readable AIIS artifacts and practitioner publications.
Start with the framework
Control Catalogue
Search all canonical control records by ID, domain, requirement and evidence type.
AIISImplementation Schema
Read the complete public schema documentation as accessible HTML.
EXBExecutive Brief
Understand purpose, operating model, audiences and adoption value.
FAQFrequently Asked Questions
Practical guidance on scope, assessment, licensing and roadmap.
Sector guidance
Publication and access model
| Artifact | Access | Use |
|---|---|---|
| Executive Brief, FAQ and sector notes | Public | Discovery and practitioner orientation |
| AIIS schema, control records and reference JSON | Public under GEL | Machine-readable implementation subject to GEL restrictions |
| Technical Report | Controlled Publication | Authoritative controlled specification; request access |
| Audit workbook, CB operations and commercial instruments | Restricted | Authorised programme use only |
Future conformance work
The v1.0 schema reserves an optional conformance_test structure. Test vectors and automated conformance objects are not populated in v1.0.0 and remain future work.