AI Change Management Procedure
Procedure for managing changes to AI systems including substantial modification assessment, change approval, and implementation.
35 min
Read Time
Level
STD-AI-009: Quality Management Standard, STD-AI-002: Risk Management Standard
[To be filled]
[To be filled]
Purpose
This procedure establishes the requirements for managing changes to AI systems throughout their lifecycle, ensuring that modifications do not adversely affect compliance with the EU AI Act, system performance, or risk levels. It ensures all changes are properly assessed, approved, implemented, and documented.
Applies To
- All changes to deployed AI systems (high-risk, limited-risk, minimal-risk)
- Model updates, retraining, and fine-tuning
- Data changes (training data, validation data, reference data)
- Algorithm and parameter changes
- Infrastructure and configuration changes
- Integration and interface changes
- Documentation changes
14
Procedure Steps
7
Roles Defined