AWS AI Practitioner Responsible AI Practices

1. Which of the following is a best practice when deploying AI systems in regulated industries?

Document model decisions and maintain audit trails

Use black-box models exclusively

Minimize stakeholder communication

Avoid fairness assessments

2. Which principle ensures that organizations can explain and justify AI system decisions to stakeholders?

Explainability and interpretability

Model optimization

Data compression

Feature selection

3. How does AWS recommend addressing identified bias in production models?

Document findings and develop mitigation strategies

Ignore minor bias issues

Deploy without remediation

Increase model complexity

4. What is the primary goal of fairness assessment in AI models?

To ensure models perform equally well across different demographic groups

To maximize model complexity

To reduce training time

To increase model parameters

5. Which of the following stakeholders should be involved in responsible AI governance?

Data scientists only

Technical teams, business leaders, and ethics experts

Marketing departments exclusively

External consultants only

6. What is the relationship between transparency and accountability in responsible AI?

Transparency enables accountability by making decisions understandable

They are unrelated concepts

Accountability makes models faster

Transparency reduces model accuracy

7. Which responsible AI practice involves regularly testing models for performance degradation?

Model monitoring and maintenance

Feature normalization

Hyperparameter tuning

Data augmentation

8. What is the purpose of bias detection in pre-training data?

To identify and mitigate potential unfairness before model training

To increase data volume

To reduce inference time

To improve data compression rates

9. Which AWS tool provides feature importance analysis to identify which inputs drive model predictions?

Amazon SageMaker Autopilot

Amazon SageMaker Clarify

AWS Data Exchange

Amazon Forecast

10. What does 'explainability' mean in the context of machine learning models?

The ability to understand why a model makes specific predictions

The speed at which a model processes data

The number of features used in the model

The percentage of training data used

11. Which AWS service provides automated bias detection and fairness metrics for machine learning models?

Amazon SageMaker Clarify

AWS Lambda

Amazon QuickSight

AWS Glue

12. What is model governance in the context of responsible AI?

The process of tracking and managing ML model lifecycles

Reducing model inference latency

Compressing model parameters

Automating feature scaling

13. AWS Clarify can detect bias in which of the following model stages?

Pre-training only

Post-training only

Both pre-training and post-training

Only during inference

14. Which principle emphasizes that AI systems should be designed to benefit all users equally?

Inclusivity

Transparency

Robustness

Scalability

15. What is the role of fairness metrics in responsible AI?

To measure how equitably a model treats different groups

To increase model speed

To reduce model size

To improve data storage efficiency

16. Which AWS service helps monitor model performance and detect data drift over time?

Amazon CloudWatch

AWS Model Monitor

Amazon DevOps Guru

AWS Systems Manager

17. Algorithmic bias in machine learning models can result from which of the following?

Biased training data

Flawed feature engineering

Imbalanced class distributions

All of the above

18. What does SHAP stand for in the context of model interpretability?

Scalable Hierarchical Analysis Platform

SHapley Additive exPlanations

Shared Adaptive Prediction Handler

Statistical Hypothesis Analysis Protocol

19. Which of the following is a key component of AWS's responsible AI framework?