Which set of stages should a lifecycle-tracking system for models and generative AI assets cover?
Select an answer to reveal the explanation.
Short Explanation
Skip a stage and you get a partial story - the whole point is that every stage is covered, not just the exciting or the regulated ones.
Full Explanation
The correct answer is the full six-stage set: intake, development, validation, deployment, monitoring, and retirement. Each stage carries information the others do not: intake sets purpose and ownership, development and validation set how it was built and tested, deployment and monitoring show how it behaves in use, and retirement closes the record. The development-and-deployment-only answer is incorrect because omitting intake and validation loses ownership context and evidence that the model was tested before use. The intake-and-retirement-only answer is incorrect because skipping the middle stages loses exactly the detail an investigation would need. The validation-and-monitoring-only answer is incorrect because narrowing to regulator-focused stages leaves the rest of the lifecycle ungoverned. For the exam, name all six stages when asked what lifecycle tracking covers.