Standard
Semantic Continuity in Migration and Retirement Standard
When one thing replaces another, the organization tends to verify that the move was complete: the records transferred, the checksums match, the new system is readable and populated. Completeness answers whether the bits arrived; it does not answer whether the successor decides, classifies, or behaves as the predecessor did. A migration can be complete and lossless at the level of data and still change the meaning of the outputs: a re-mapped code list, a re-trained model, a replacement control that admits a different population. When the predecessor is then retired, that semantic change becomes invisible and irreversible.
- Type
- Standard
- Version
- 0.0.8
- Date
- 2026-10-27
- Owner
- GRC Programme Manager
- Approving authority
- Governance Library Maintainer
- Classification
- Public
- Confidentiality
- Public
- Category
- Governance
- Review frequency
- Annual and upon material change to migration, lifecycle, or retirement practice
- Licence
- CC BY-SA 4.0
Read the full text on GitHub (opens in a new tab)
Outline
- 1. Purpose and scope
- 2. The semantic-continuity gate
- 3. Comparison methods by predecessor class
- 4. Pass criteria and tolerance
- 5. Authorization on a partial pass
- 6. Fail path
- 7. Evidence and retention
- 8. Relationship to existing controls
- 9. Framework alignment
- 10. Limitations
Frameworks it maps to
- ISO/IEC 42001:2023
- NIST AI RMF 1.0
- ISO/IEC 20000-1:2018
- NIST SP 800-53 Rev. 5
- CSA CCM v4.1
- ISO/IEC 27001:2022
Related documents
- AI Model Succession and Identity Continuity Standardai
- Foundation Model Lifecycle Procedureai
- AI Model Lifecycle Management Procedureai
- Cloud Exit and Data Portability Standardsupply-chain
- Supplier Exit and Data Return Proceduresupply-chain
- Change Management and Configuration Control Procedureoperations
- Asset Inventory Registeroperations
- Records Retention and Destruction Standardgovernance
Next
All Documents