This panel discussion podcast focuses on generative AI security, specifically discussing the OWASP Top 10 for LLM Apps and GenAI 2025 release. The podcast begins with introductions and informal Thanksgiving plans before diving into the recently released top 10 list, highlighting the two new categories: system prompt leakage and vector/embedding flaws. The hosts then discuss the shift towards RAG (Retrieval Augmented Generation) architectures and the emerging field of agentic applications, acknowledging the challenges and future research needed in these areas. Finally, they mention an upcoming roadmap detailing future projects and initiatives, including a deepfake threat intelligence guide planned for release in early 2024, accessible at GenAI.OWASP.org/roadmap.
