Episode 137: How We Do AI-Assisted Whitebox Review, New CSPT Gadgets, and Tools from SLCyber
Critical Thinking - Bug Bounty Podcast
00:00
Navigating AI Safety and Cybersecurity Challenges
This chapter explores the lucrative realm of AI safety research, focusing on the high bounties offered by companies like Anthropic and OpenAI for discovering vulnerabilities. The speakers discuss technical aspects of AI models, including a Kaggle challenge and the implications of performance and security improvements in AI systems. Additionally, they analyze the evolution of cyber vulnerabilities, showcasing innovative techniques like Client-Side Poisons and Tricks (CSPT) in enhancing industry security.
Transcript
Play full episode
