Skip to content
Mementropy

12 — Safe User Interactions with Generative AI Applications

12 — Safe User Interactions with Generative AI Applications

Section titled “12 — Safe User Interactions with Generative AI Applications”

Part 3 Introduction

Section titled “Part 3 Introduction”
  • Welcome

Module Introduction

Section titled “Module Introduction”
  • Overview
  • Purpose of Content Input Filtering
  • First line of defense
  • Input Filtering Techniques in Generative AI

Amazon Bedrock Guardrails

Section titled “Amazon Bedrock Guardrails”
  • Introduction
  • Amazon Bedrock Guardrails
  • Addressing key risks
  • Bedrock Guardrails core components and design principles
  • How it works
  • Amazon Guardrails best practices

Aws Step Functions And Lambda For Custom Moderation Workflows

Section titled “Aws Step Functions And Lambda For Custom Moderation Workflows”
  • Introduction
  • Benefits of AWS Step Functions for custom moderation workflows

Prompt Injections And Real-time Validation Mechanisms

Section titled “Prompt Injections And Real-time Validation Mechanisms”
  • Introduction
  • Prompt Injections Overview
  • Best practices for prompt injection prevention

Model Response Filtering Purpose And Techniques

Section titled “Model Response Filtering Purpose And Techniques”
  • Introduction
  • Model Response Filtering
  • Benefits of model response filtering
  • Techniques for model response filtering
  • Implementing model response filtering with Amazon Bedrock
  • Best practices for model response filtering

Output Response Filtering With Amazon Bedrock Guardrails

Section titled “Output Response Filtering With Amazon Bedrock Guardrails”
  • Introduction
  • Model response filtering with Amazon Guardrails
  • Content Filters
  • Denied Topics
  • Word Filters
  • Sensitive information filters
  • Contextual grounding check
  • Automated Reasoning Checks
  • Safeguard tiers with protection levels and options

Content Moderation And Toxicity Detection

Section titled “Content Moderation And Toxicity Detection”
  • Introduction
  • Content moderation and toxicity detection
  • Toxicity Detection API Overview
  • Toxicity categories and detection
  • Implementing toxicity detection
  • Best practices for content moderation

Text-to-sql Transformations To Prevent Harmful Outputs

Section titled “Text-to-sql Transformations To Prevent Harmful Outputs”
  • Introduction
  • Understanding text-to-SQL transformations
  • Use cases
  • Implementing steps for text-to-SQL

Grounding And Fact-checking Responses With Amazon Bedrock Kb

Section titled “Grounding And Fact-checking Responses With Amazon Bedrock Kb”
  • Introduction
  • Contextual grounding checks
  • Components of contextual grounding checks
  • Implementing contextual grounding checks with Bedrock Knowledge Base

Semantic Similarity Searches For Verification

Section titled “Semantic Similarity Searches For Verification”
  • Introduction
  • Semantic similarity search for verification
  • Benefits of semantic similarity searches for verification

Enforce Structured Outputs With Json Schema

Section titled “Enforce Structured Outputs With Json Schema”
  • Introduction
  • Benefits of structured output enforcement

Benefits Of Multilayer Content Safety Architectures

Section titled “Benefits Of Multilayer Content Safety Architectures”
  • Introduction
  • Multilayer content safety architectures
  • Pre-processing filters with Amazon Bedrock Data Automation and Amazon Comprehend
  • Model-based guardrails with Amazon Bedrock
  • Post-processing validation with AWS Lambda
  • Comprehensive content safety strategy

Module Summary

Section titled “Module Summary”
  • Recap and next steps
  • Resources