Validation of AI

Validation of AI is the common shorthand for the lifecycle management of AI (sub-)systems in the GxP environment. To be technically precise, “AI” itself is not validated: the overall computerized system is validated against its Intended Use; the AI model (or subsystem) itself is verified against its specification and must be fit for purpose for its Context of Use – on an ongoing basis, because it changes with data, training, and context.

Explainability and quality arise from the lifecycle activities (specification, training, verification, operation, monitoring) and their evidence, not from a one-time approval.