AI inventory
Definition
A maintained register of the AI systems and models an organization develops, buys or uses, with owner, purpose, data, risk level and status for each.
An AI inventory is the starting point of any AI governance program: you cannot classify, assess or audit systems you do not know exist. The NIST AI RMF (Govern 1.6) calls for mechanisms to inventory AI systems, and an ISO/IEC 42001 management system depends on a clear scope of the AI systems it covers.
Each entry usually records the business owner, the purpose and context of use, whether the system is built in-house or supplied by a vendor, the data it uses (including personal data), the role the organization plays under the EU AI Act (provider or deployer), the risk classification, links to documentation such as model cards, and the lifecycle status. Intake procedures keep it current, and it pairs naturally with the record of processing activities kept for data protection.
Example: an insurer surveys business units, finds a claims triage model and a generative AI assistant nobody had registered, and adds both with owners and review dates.
Where this term is assessed
Assessed in the Responsible AI Practitioner (RAIP), domain 1: AI system lifecycle and inventory.
Also covered in: Responsible AI Leader (RAIL)
Primary sources
- ISO/IEC 42001:2023
- NIST AI RMF 1.0 (Govern 1.6)