AI Security Interview Series: Pushing the Boundaries of Prompt Injection Attacks
HTML-код
- Опубликовано: 5 окт 2024
- In this installment of our AI security interview series, we bring you a conversation between AI security researcher Kai Greshake and Kojin Oshiba, Co-Founder of Robust Intelligence. He has pioneered research in discovering vulnerabilities in generative AI, with a focus on indirect prompt injections, . During this session, he sheds light on the complexities of AI security, shares examples of indirection prompt injections, discusses the implications for organizations, and provides insights into a tool he developed for injecting prompt injections.
Interesting hacks /exploits, thanks for sharing. And Congrats Kai on your nVidia gig! Would be nice if the video description contained links to Kai's twittter, blog and his indirect prompt injection paper.