Red Teaming AI Systems: The Path, the Prospect and the Perils
summary
Panelists from Google, Microsoft and MITRE discuss why red teaming for AI is growing, what concrete value organizations can get from it, and whether it’s necessary or just a trend. They examine questions such as whether practical attack tools (a “mimikatz for machine learning”) exist and recommend simple steps organizations can take to secure AI systems.
AI-written summary · about
chapters (9)
- 0:00Introduction to AI security
- 4:44Defining AI red teaming
- 7:03Real-world threat landscape
- 11:20Role of red teams in industry
- 14:00Audience Q&A: Practicality
- 20:37Anatomy of an AI red team
- 25:44Formulating the red team
- 30:13Audience Q&A: Mitigations
- 37:41The future of AI security
related talks
41mHow Microsoft Approaches AI Red Teaming
44mRed Teaming AI: Getting Started with PyRIT
25mProtecting AI Infrastructure
50mSecurity as Part of Responsible AI: At Home or At Odds?
45mHardening AI/ML Systems - The Next Frontier of Cybersecurity
34mLLM Security 101 - An Introduction to AI Red Teaming
18mCommit Secure AI: Embedding Security in Code, Pipelines, and Production
44mAI Meets Cybersecurity: Boosting Speed Without Losing Control