microsoft/AI-Red-Teaming-Playground-Labs
AI red-teaming training course with hands-on challenges for security testing of generative AI systems.

This repository contains interactive challenges for teaching AI red-teaming methodology to security professionals. The curriculum, originally delivered at Black Hat USA 2024, covers adversarial machine learning and Responsible AI failure modes. The playground environment is built on Chat Copilot, and the repository includes Jupyter Notebooks demonstrating how to use Microsoft’s PyRIT framework to automate red-teaming tasks including prompt injection testing and risk identification in generative AI systems.