Explore the complexities and pitfalls of applying AI in security through this 44-minute conference talk from BSidesLV. Delve into real-world examples and a live demonstration showcasing the limitations of GenAI in addressing code vulnerabilities. Learn from the speakers' research using OpenAI for auto remediation, which yielded underwhelming results and highlighted the risks of over-relying on AI for complex tasks. Gain insights into a more reliable approach to leveraging GenAI for security using the RAG (Retrieval-Augmented Generation) Framework. Understand the importance of a nuanced approach to AI implementation in security, avoiding the trap of viewing it as a standalone solution. Presented by Kirill Efimov and Eitan Worcel, this talk serves as a valuable cautionary lesson for security professionals considering AI integration.
Overview
Syllabus
Common Ground, Tue, Aug 6, 12:30 - Tue, Aug 6, CDT
Taught by
BSidesLV