AWS Certified AI Practitioner
Tools for Model Transparency
The artifacts that make a model accountable: SageMaker Model Cards for the models you build, AWS AI Service Cards for the ones you buy, Clarify reports, Bedrock evaluations, and what a model's license and openness actually tell you.
Intermediate 22 minutes 5 Learning Objectives
- Describe what a SageMaker Model Card records, including risk ratings and its versioning behavior
- Explain what an AWS AI Service Card contains and when to read one
- Distinguish a model card from an AI service card by author, scope, and the question each answers
- Explain how Bedrock evaluations and Clarify reports supply evidence rather than documentation
- Separate open weights, open training data, and permissive licensing when assessing how transparent a model is