Safeguarding Generative AI Applications: Navigating Security Challenges and Best Pract...
summary
ChatGPT hits 100 million users in two months (versus 16 years for smartphones) and Gartner predicts 80% of enterprises using generative AI by 2026 up from 5% in 2023, which drives urgent threat modeling across three layers: AI usage, AI application, AI platform. Concrete attacks include prompt injection and document/email poisoning, wallet abuse, data/model theft; mitigations are prompt validation, red‑teaming, scoped access, logging, grounding outputs with citations, and vendor tools like Azure Content Safety and Prompt Flow.
AI-written summary · about
chapters (12)
- 0:00Securing generative AI apps
- 1:49Generative AI security risks
- 3:34Three threat layers in AI apps
- 6:26How AI apps are built
- 8:51Prompt injection attacks
- 12:07Best practices for AI security
- 16:41Red teaming & prompt validation
- 18:46Mitigating AI over-reliance
- 21:59UX and interface security
- 23:02Continuous AI security monitoring
- 24:06Discover protect comply framework
- 27:02Q&A
related talks
31mGenAI among us
46mGenerative AI: Cyber resiliency and risk
44mGenerative AI Security — A Practical Guide to Securing Your AI Application
45mOverview of Risks, Vulnerabilities and Paths to Safer Use of Generative AI
27mChallenges of Generative AI
40mSecuring Generative AI: Is it all an Illusion?
45mSecuring and Governing Generative AI: Learnings from Microsoft
44mThreat Modeling to Protect Your Generative AI Applications