Accelerating Innovation through LLM-enabled Prototypes

LLMs have become a strong force multiplier, enabling researchers to bring complex ideas to fruition that previously would have been deemed infeasible. This talk will introduce some areas of focus in security research where LLMs are assisting greatly including fuzzing, triage, and security analysis. It will also provide a case study where industry researchers identified a long-standing issue in kernel code which then motivated the use of an LLM to enable a CPU-based kernel shadow stack security mitigation that required deep kernel modifications. Overall, the talk will share examples of how LLMs are enabling new research to be performed in ways that were previously considered intractable.
Co-sponsored by: Santa Clara University
Bldg: Benson Memorial Center 101BC (Parlors BC), Santa Clara University, 500 El Camino Real, Santa Clara, California, United States, 95053