Treating governance as entirely separate from the model development and deployment tooling would be a mistake. What is the reason?
Select an answer to reveal the explanation.
Short Explanation
Keep governance walled off from where the work happens and it has nothing to watch. You're left with records that depend on what someone remembers to tell it, so evidence and asset status go stale.
Full Explanation
The correct answer is the option stating that governance would have no way to collect evidence and monitor assets as they move through development and deployment, leaving its records incomplete and out of date. Evidence collection and asset monitoring depend on a connection to the environments where models are actually built and run; without that connection, governance records fall behind or go missing. The licensing-terms option is incorrect because there is no general licensing rule that ties integration to permitted use of development tools. The prediction-accuracy option is incorrect because how governance tooling is connected has no effect on model prediction accuracy. The developer-code option is incorrect because developers can write code with or without governance integration; the issue is what gets recorded, not whether work can happen. For the exam, tie 'governance separate from tooling' directly to a loss of evidence collection, not to unrelated technical effects.