Secure Generative AI with Amazon Bedrock

Course details

While the promise of generative AI is boundless, it comes hand-in-hand with unique security considerations. In this course, AWS-certified software programmer and DevOps advocate Dipali Kulshrestha explores generative AI as a tool to revolutionize customer experiences, innovate uncharted applications, and elevate organizational productivity, while also discussing the corresponding security considerations. Learn about Amazon Bedrock and the very convenient way it provides to simplify generative AI development, with foundation models from Amazon and leading AI startups. Explore the architectures, data flows, and security-related aspects of model fine-tuning, as well as the prompting and inference phases. Plus, find out how Amazon Bedrock uses AWS security services and capabilities, such as AWS KMS, AWS CloudTrail, and AWS Identity and Access Management (IAM).

Concepts

Introduction

Security concerns in generative AI

Introduction to Generative AI

Importance of security in generative AI development
Addressing ethical challenges in generative AI development
Ethical best practices for generative AI
Introduction to generative AI models and techniques
Overview of AWS services used in generative AI development

Securing Generative AI Models Using Amazon Bedrock

Introduction to Amazon Bedrock
Amazon Bedrock features and its workflow
Foundation models (FMs) supported by Amazon Bedrock
Knowledge base for Amazon Bedrock
Agents for Amazon Bedrock
Datasets and vector embeddings
Architecture patterns with Bedrock
Amazon Bedrock Dashboard overview (Demo)
Integrating FMs into your code with Amazon Bedrock (Demo)

Network Security for Generative AI Applications

Client connectivity with Bedrock
Guardrails for Amazon Bedrock

Security of the GenAI Model

Data privacy and localization
Data security - Data in transit
Data security - Data at rest

Security in the GenAI Model

Use and misuse of GenAI models
GenAI security scoping

Securing Data Flows

Single-tenancy vs. multi-tenancy
Multi-tenancy inference
Single-tenancy inference
Model fine-tuning

Secure Deployment of Generative AI Models

Deploy GenAI models with SageMaker JumpStart
Deploy GenAI models on Amazon EKS

Compliance and Governance in Amazon Bedrock-Based Generative AI

IAM
Implementing auditing, logging, and compliance mechanisms
Infrastructure security
Cross-service confused deputy prevention
Configuration and vulnerability analysis in Amazon Bedrock
Configurable security controls

Emerging Trends in Generative AI Security with Amazon Bedrock

Monitor Amazon Bedrock
Quotas for Amazon Bedrock

Conclusion

Learning more about generative AI in AWS