Description
Planned retirement or decommissioning of AI systems with data retention, transfer procedures, and stakeholder communication.
Implementation Guidance
Retirement Planning
Develop a retirement plan covering: timeline, replacement system readiness, data archival requirements, model artifact disposition, stakeholder notification schedule, and regulatory obligations for data retention.
Data and Model Disposition
Determine handling for: training data (archive, anonymize, or delete per retention policy), model artifacts (archive for audit trail or securely delete), prediction logs (retain for compliance period), and user data (transfer or delete per privacy requirements).
Stakeholder Communication
Notify all affected parties: users transitioning to replacement systems, downstream systems losing an integration, regulators requiring notification of material changes, and internal teams adjusting workflows. Provide adequate transition support.
Evidence Requirements
- AI system retirement plan
- Data disposition records
- Model artifact archival or deletion records
- Stakeholder notification records
- Transition completion confirmation