Question 18
Domain 2 — AI Operations, Lifecycle, and Control EnvironmentModel cards are primarily used to do which of the following?
Correct answer: C
Explanation
Model cards are used to summarize a model’s purpose and behavior, including its intended use, performance, limitations, and risks. This matches the standard definition of model cards as documentation that helps users understand how a model should be used and what ethical considerations apply.
Why each option is right or wrong
A. Store model weights and deployment metadata in the model registry.
B. Track training experiment hyperparameters and run metrics.
C. Document a model's intended use, performance characteristics, limitations, and ethical considerations.
Model cards are the standard documentation artifact introduced by Mitchell et al. (2019) to record a model’s intended use, evaluation results, limitations, and ethical considerations such as bias and fairness. In practice, they are meant to accompany a trained model so users can assess performance across relevant metrics and understand constraints before deployment or reuse.
D. Define feature pipeline versioning and data transformation rules.