Procedure
Foundation Model Lifecycle Procedure
This procedure governs the lifecycle of foundation and frontier models consumed by the organization: evaluation, supplier selection, contractual integration, deployment, ongoing monitoring, version-transition management, and exit. It complements the AI model lifecycle management procedure (which is generic) with the specific obligations and risk vectors of foundation-model consumption.
- Type
- Procedure
- Version
- 0.0.20
- Date
- 2026-10-29
- Owner
- AI Governance Approver
- Approving authority
- Governance Library Maintainer
- Classification
- Public
- Confidentiality
- Public
- Category
- AI Governance
- Review frequency
- Annual and upon material change to foundation-model providers, regulatory expectation, or risk posture
- Licence
- CC BY-SA 4.0
Read the full text on GitHub (opens in a new tab)
Outline
- Purpose
- Scope
- Procedure
- Provider systemic-risk expectations (GPAI models with systemic risk)
- Specific risk vectors and mitigations
- Generative-AI risk classes
- Operating expectations
- Framework alignment
- Limitations
Frameworks it maps to
- ISO/IEC 42001:2023
- EU AI Act
- EU GPAI Code of Practice (Safety and Security Chapter, July 2025)
- NIST AI RMF 1.0
- NIST AI 600-1
- ISO/IEC 5338:2023
- ISO/IEC 23894:2023
- OECD AI Principles
- G7 Hiroshima Process International Code of Conduct (2023)
- DORA
- CSA AI Model Risk Management Framework (2024)
Related documents
- AI Governance and Risk Frameworkai
- AI Security and Risk Standardai
- AI Model Risk Standardai
- AI Model Lifecycle Management Procedureai
- AI Evaluation Procedureai
- Model Registryai
- AI Vendor Security Questionnaire Templateai
- Third-Party AI Due Diligence Proceduresupply-chain
- Cloud Exit and Data Portability Standardsupply-chain
- AI Model Succession and Identity Continuity Standardai
- Semantic Continuity in Migration and Retirement Standardgovernance
Next
All Documents