HIDA Lecture: Security Challenges Posed by LLMs
Speaker: Sahar Abdelnabi, AI security researcher at the Microsoft Security Response Center
Date: 06.06.2024, 04:30 pm
Title: On New Security and Safety Challenges Posed by LLMs and How to Evaluate Them
Abstract
This online lecture delves into the pervasive integration of Large Language Models (LLMs) within numerous real-world applications, presenting both expansive opportunities for assistance across diverse tasks and formidable challenges in terms of security and safety. In contrast to previous models characterized by static generation, the dynamic, multi-turn, and adaptable nature of LLMs poses considerable difficulties for robust evaluation and control. Join us as we dissect the emergent risks associated with LLMs, explore methodologies for rigorous evaluation, and confront the complex obstacles involved in implementing effective mitigations.
Sahar Abdelnabi
Sahar Abdelnabi is an AI security researcher at the Microsoft Security Response Center (Cambridge, UK). Previously, she was a doctoral researcher at CISPA - Helmholtz Center for Information Security. Her research interests lie in the broad intersection of machine learning with security, safety, and sociopolitical aspects.
