GLOSSARY

An international standard defining comprehensive life cycle processes for engineering, deploying, and maintaining artificial intelligence systems.

It enables organizations to embed AI engineering into established software development lifecycles, ensuring traceable processes, quality management, and compliance readiness.

A DevOps engineering organization integrates ISO/IEC 5338 lifecycle stages into its CI/CD pipeline, establishing formal review gates between model prototyping, validation, deployment, and ongoing retirement.

Adopting lifecycle processes requires cultural change and cross-disciplinary collaboration, and process conformance alone does not ensure business viability or algorithmic accuracy.

Extending ISO/IEC/IEEE 12207 software engineering standards, ISO/IEC 5338 details AI-specific lifecycle processes including AI requirements definition, data management, model training, verification and validation, deployment, operational monitoring, and decommissioning. It provides structured guidance on managing iterative model updates and non-deterministic behavior.