Lenny's PodcastSander Schulhoff: Why AI guardrails fail every red team test
How prompt injection and jailbreaks bypass state-of-the-art guardrails; agents that send emails or touch databases turn every bypass into real damage.
Sander SchulhoffguestLenny Rachitskyhost
CHAPTERS
- 0:00 – 5:14
Introduction to Sander Schulhoff and AI security
- 5:14 – 11:42
Understanding AI vulnerabilities
- 11:42 – 17:55
Real-world examples of AI security breaches
- 17:55 – 19:44
The impact of intelligent agents
- 19:44 – 21:09
The rise of AI security solutions
- 21:09 – 23:44
Red teaming and guardrails
- 23:44 – 27:52
Adversarial robustness
- 27:52 – 38:22
Why guardrails fail
- 38:22 – 44:44
The lack of resources addressing this problem
- 44:44 – 55:49
Practical advice for addressing AI security
- 55:49 – 59:06
Why you shouldn’t spend your time on guardrails
- 59:06 – 1:09:15
Prompt injection and agentic systems
- 1:09:15 – 1:11:47
Education and awareness in AI security
- 1:11:47 – 1:17:52
Challenges and future directions in AI security
- 1:17:52 – 1:21:57
Companies that are doing this well
- 1:21:57 – 1:32:40
Final thoughts and recommendations
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome