AIF-C01 Question 225
Select 3Your organization has deployed a machine learning model using Amazon SageMaker to predict customer churn. As part of your governance protocols, you need to ensure the model complies with your organization's policies for ethical AI usage and data privacy. Which processes should you follow to maintain governance compliance for the deployed model?
- A
Conduct regular reviews of the model's performance and fairness metrics.
- B
Implement monitoring to detect data drift and retrain the model frequently.
- C
Ensure access to the model and its predictions is restricted using AWS Identity and Access Management (IAM) policies.
- D
Document and review the model's decision-making process to ensure transparency.
- E
Disable logging in Amazon CloudWatch to reduce storage costs.
Show answer and explanation
Correct answers: A, C, D
Explanation
Governance protocols for machine learning models emphasize fairness, transparency, and security. Regular reviews of performance and fairness metrics ensure the model adheres to ethical standards. Restricting access using IAM policies safeguards data privacy, and documenting the decision-making process ensures transparency. These practices align with governance frameworks and policies, making them essential for compliance.
- A. Correct.
Regular reviews of performance and fairness metrics are essential to ensure that the model adheres to governance protocols, such as ethical AI standards and fairness in predictions.
- B. Incorrect.
While monitoring and retraining are good practices for maintaining model accuracy, they are not explicitly part of governance protocols like ethical AI and data privacy compliance.
- C. Correct.
Restricting access using AWS IAM policies ensures that only authorized users can access the model and its predictions, which is crucial for governance and data privacy.
- D. Correct.
Documenting and reviewing the model's decision-making process promotes transparency, which is a core aspect of AI governance frameworks.
- E. Incorrect.
Disabling logging in Amazon CloudWatch may reduce costs but violates governance practices, as logging is critical for auditing and maintaining compliance protocols.