Module 2 Labs: Responsible AI in practice

These notebooks implement the responsible-AI techniques from Module 2 on Amazon Bedrock. Read each chapter first, then work the lab.

Lab	Notebook	Connects to
2	Lab 2: Data Protection	Ch. 3: privacy and security, protecting sensitive data.
3	Lab 3: Robustness	Ch. 3: implementing robustness against noisy and adversarial inputs.
4b	Lab 4b: Watermarking	Ch. 4: watermarking AI-generated text.
4c	Lab 4c: Debiasing	Ch. 4: debiasing model outputs.

Expected error: “AccessDeniedException” (model access not enabled)

Some lab cells may display an error like:

AccessDeniedException: An error occurred (AccessDeniedException) when calling
the InvokeModel operation: Model access is denied due to IAM user or service
role is not authorized to perform the required AWS Marketplace actions
(aws-marketplace:ViewSubscriptions, aws-marketplace:Subscribe) to enable
access to this model.

This is not a bug in the lab code. It means your AWS account or IAM role has not enabled access to that specific foundation model in Amazon Bedrock. To fix it: go to the Amazon Bedrock console -> Model access, request or enable the model, make sure your role has the aws-marketplace:ViewSubscriptions and aws-marketplace:Subscribe permissions, and re-run the cell after a few minutes. Model availability and the exact permissions required vary by AWS Region and account, so consult the Amazon Bedrock documentation for your setup.

Running the labs

The notebooks call live Amazon Bedrock endpoints and are rendered here for reading rather than executed during the book build. Run them in an environment with AWS credentials and Bedrock model access (for example Amazon SageMaker with the conda_python3 kernel), installing the packages in the lab’s requirements.txt. Lesson 1 (Evaluating LLMs) has no lab; explore Bedrock’s model evaluation feature in the console instead.