AI Unlocked: Responsible AI & Azure AI Content Safety [English]

Learn conceptual and technical approaches to implementing LLM applications and handling data responsibly, whether you're using OpenAI or OSS models. In this session, Amanda Wong addresses how to evaluate your LLM application while protecting sensitive information using tools such as Azure AI Content Safety and the Presidio De-identification Toolbox. We will also cover best practices for data security in Azure, including encryption and access control.
Agenda
- Overview on MSFT Responsible AI Principles on Generative AI
- Overview on MSFT built-in LLM data security and trust
- Walk-through of LLM + RAI architecture diagram
- Introduction to importance of content safety
- DEMO: Content AI Safety Tool
- Introduction to importance of de-identifying data
- DEMO: Presidio De-identification toolbox
- Next steps: resources
For more Microsoft AI resources and events, please visit the Microsoft AI & ML Partner Prep page at aka.ms/AIMLPar...