AI Red Teaming 101 – Full Course (Episodes 1-10) скачать в хорошем качестве

AI Red Teaming 101 – Full Course (Episodes 1-10)

Welcome to the complete AI Red Teaming 101 series! This beginner-friendly series covers the essential basics of AI red teaming, giving you the core concepts and techniques you need to understand generative AI security risks. Led by Dr. Amanda Minnich (Principal Research Manager, Microsoft AI Red Team), Gary Lopez (Principal Offensive AI Scientist, ADAPT), and Nina Chikanov (Security Software Engineer II, Microsoft AI Red Team), this series demystifies the critical security challenges facing modern AI deployments. Whether you're a security professional, ML practitioner, or just curious about AI risks, this complete series is designed to get you started—no PhD required! You'll learn how Microsoft's AI Red Team secures real-world AI systems like Copilots before release and get hands-on experience with cutting-edge tools and techniques. From understanding how generative AI models work to trying out multi-turn attacks and exploring defense strategies, this series covers the core fundamentals of AI red teaming. You'll explore real-world vulnerabilities like prompt injection and jailbreaks, learn to use Microsoft's open-source PyRIT tool for automated testing, and discover proven mitigation strategies like Spotlighting that help secure AI applications in production. What You'll Learn: The fundamentals of AI red teaming vs. traditional red teaming How generative AI models are built and where vulnerabilities emerge Direct and indirect prompt injection techniques with real-world examples Advanced adversarial prompt engineering and manipulation tactics Multi-turn attacks including Skeleton Key and Crescendo methodologies Defensive strategies and Spotlighting techniques for robust AI security How to automate red teaming workflows using PyRIT Hands-on demonstrations of single-turn and multi-turn attack automation Testing strategies for both text and image generation models Ready to master AI red teaming? This complete series gives you the knowledge and tools to identify, exploit, and defend against the most critical vulnerabilities in generative AI systems. ✅ Chapters: 00:00 - Episode 1: What is AI Red Teaming? | AI Red Teaming 101 06:05 - Episode 2: How Generative AI Models Work (and Why It Matters) | AI Red Teaming 101 13:29 - Episode 3: Direct Prompt Injection Explained | AI Red Teaming 101 19:06 - Episode 4: Indirect Prompt Injection Explained | AI Red Teaming 101 25:25 - Episode 5: Prompt Injection Attacks – Single-Turn | AI Red Teaming 101 35:32 - Episode 6: Prompt Injection Attacks: Multi-Turn | AI Red Teaming 101 ✅ Resources: https://aka.ms/airt101 42:59 - Episode 7: Defending Against Attacks: Mitigations and Guardrails | AI Red Teaming 101 49:53 - Episode 8: Automating AI Red Teaming with PyRIT | AI Red Teaming 101 57:55 - Episode 9: Automating Single-Turn Attacks with PyRIT | AI Red Teaming 101 01:06:26 - Episode 10: Automating Multi-Turn Attacks with PyRIT | AI Red Teaming 101