Workspaces / Demonstration library / P18
AI Model Assessment, Pre-Production Review
File corpus.json
Type internal policy
Classification internal
Department risk
Permitted roles analyst, manager, executive, admin
Version 1.0 (effective 2026-01-01)
Owner Synthetic demonstration policy
Status active
SHA-256 80ad6b830ad567a7ec81badb2bbae0ec91d165f3ebaa5eefbb658c8db3105b8e
Uploaded 2026-10-31T06:37:16+00:00
Chunks (1)
Each chunk is a citable unit. Answers cite the chunk id.
[P18]
An AI model assessment should document the model purpose, intended users, data sources, performance measures, known limitations, risk classification, security and privacy considerations, human oversight, monitoring requirements and approval ownership. Material AI systems require appropriate validation and risk review before production deployment. Assessment evidence must be retained and material model changes must trigger reassessment.
Answers come only from the governed library and are decision support, not legal advice. Verify regulatory content against the regulator's current publication.