SageMaker Clarify Bias Detection Question Explained
This AWS AI Practitioner question is a responsible AI service-selection trap. Bias metrics, demographic fairness, and SHAP explanations point to Amazon SageMaker Clarify.
Short answer
The correct answer is C. Amazon SageMaker Clarify.
Amazon SageMaker Clarify is correct because the scenario asks for fairness analysis and prediction explainability. The lending team wants to know whether approval rates differ across demographic groups before deployment, and it also wants explanations for individual credit-scoring decisions. Clarify is the SageMaker feature built for those signals. It can measure pre-training bias in datasets, post-training bias in model predictions, and feature attributions with SHAP so teams can understand which inputs influenced a prediction. The phrase "before deployment" is an exam clue against SageMaker Model Monitor, which is mainly for monitoring deployed endpoints for drift and quality issues over time. The phrase "explain individual predictions" is a clue against Debugger and Trusted Advisor. Debugger inspects training jobs and tensors; Trusted Advisor checks AWS account best practices. On AIF-C01, choose SageMaker Clarify when a custom ML model question combines demographic groups, fairness metrics, bias detection, pre-training or post-training analysis, and explainability.
The Question
A lending team wants to check whether a credit-scoring model produces different approval rates across demographic groups before deployment, and also wants to explain individual predictions. Which AWS service is purpose-built for this?
Why C is correct
Amazon SageMaker Clarify is correct because the scenario asks for fairness analysis and prediction explainability. The lending team wants to know whether approval rates differ across demographic groups before deployment, and it also wants explanations for individual credit-scoring decisions. Clarify is the SageMaker feature built for those signals. It can measure pre-training bias in datasets, post-training bias in model predictions, and feature attributions with SHAP so teams can understand which inputs influenced a prediction. The phrase "before deployment" is an exam clue against SageMaker Model Monitor, which is mainly for monitoring deployed endpoints for drift and quality issues over time. The phrase "explain individual predictions" is a clue against Debugger and Trusted Advisor. Debugger inspects training jobs and tensors; Trusted Advisor checks AWS account best practices. On AIF-C01, choose SageMaker Clarify when a custom ML model question combines demographic groups, fairness metrics, bias detection, pre-training or post-training analysis, and explainability.
Why the other options are wrong
SageMaker Model Monitor watches deployed endpoints for data quality, model quality, drift, and related production signals. It is not the primary service for pre-deployment bias analysis and individual prediction explanations.
SageMaker Debugger helps inspect and debug training jobs, tensors, gradients, and resource utilization. It is not the responsible AI service for demographic bias metrics or SHAP explainability.
AWS Trusted Advisor checks AWS account best practices such as cost optimization, service limits, security, and fault tolerance. It does not analyze ML model fairness or prediction explanations.
The exam rule for SageMaker Clarify
Amazon SageMaker Clarify is the AWS service to associate with bias detection and explainability for custom machine learning models. It can analyze training data for pre-training bias, evaluate model predictions for post-training bias, and produce SHAP-based feature attribution reports that explain why a model made a prediction. The common trap is confusing Clarify with Model Monitor. Model Monitor is production monitoring: it watches deployed endpoints for data quality, model quality, drift, and related operational changes over time. Clarify is responsible AI analysis: it measures fairness and explainability before or after training. Another trap is confusing Clarify with Bedrock Guardrails. Guardrails are for runtime safety controls on foundation model applications, such as blocking harmful content, denied topics, or sensitive information. Clarify is for statistical fairness and explainability, often with SageMaker models and tabular data. The exam wording matters. "Demographic groups," "approval rates," "bias metrics," "explain predictions," "SHAP," "pre-training," and "post-training" are Clarify clues. "Endpoint drift" and "production data quality" are Model Monitor clues. "Harmful content" and "denied topics" are Guardrails clues.
SageMaker Clarify exam trap table
| Scenario clue | Choose | Avoid | Why |
|---|---|---|---|
| Bias metrics, demographic groups, approval-rate differences, SHAP explanations | Amazon SageMaker Clarify | Model Monitor, Debugger, Trusted Advisor | Clarify is the responsible AI analysis feature for fairness and explainability on ML models. |
| Data drift, model quality drift, endpoint monitoring after deployment | Amazon SageMaker Model Monitor | Clarify | Model Monitor watches deployed endpoints over time. Clarify analyzes bias and explainability before or after training. |
| Training-job debugging, tensors, gradients, resource utilization | Amazon SageMaker Debugger | Clarify | Debugger helps inspect training behavior. It is not the service for demographic fairness metrics or SHAP explanations. |
| Harmful content filters, denied topics, PII controls for a foundation model app | Amazon Bedrock Guardrails | Clarify | Guardrails manage runtime safety for GenAI applications. Clarify measures statistical bias and model explainability. |
Ready to see how you'd score?
Take the free practice quiz and find out which AWS Certified AI Practitioner domains you need to focus on. No signup required.
Practice 5 similar questions
Same cert, same or adjacent domain. Use these after reviewing the explanation.
Related AWS Certified AI Practitioner Practice Questions
Quick FAQ
When should I choose SageMaker Clarify on AIF-C01?
Choose SageMaker Clarify when the scenario asks for bias detection, demographic group fairness, pre-training or post-training bias metrics, SHAP explainability, or individual prediction explanations.
What is the difference between SageMaker Clarify and Model Monitor?
Clarify analyzes bias and explainability before or after training. Model Monitor watches deployed endpoints for data quality, model quality, bias drift, and feature attribution drift over time.
What is the difference between SageMaker Clarify and Bedrock Guardrails?
Clarify is for ML model fairness and explainability. Bedrock Guardrails are for runtime safety controls on foundation model applications, such as denied topics, harmful content, and PII handling.
What does SHAP mean for the AWS AI Practitioner exam?
For the exam, SHAP is the clue for feature attribution and explainability: it helps show which input features influenced a model prediction.