We are witnessing a paradigm shift in AI, transitioning from deep learning models to the era of Large Language Models (LLMs). This shift signifies a transformative advancement in AI, enabling it to be applied to diverse real-world safety-critical applications. Despite these impressive achievements, a fundamental question remains: are LLMs truly ready for safe, and secure use?
In this talk, I will demonstrate how my research integrates a computer security mindset to address this question. To build secure and safe LLMs, I will explore two core system perspectives: (1) examining the lifecycle of LLMs, and (2) analyzing the role of LLMs along with the information flows within agentic systems. I will discuss how to develop principled frameworks to systematically assess and enhance LLM safety across stages—from inference to fine-tuning. Furthermore, I will highlight why securing LLMs requires more than just a model-level focus and introduce key security vulnerabilities from a systems perspective. Ultimately, I will share my vision for securing LLMs through both model-level and system-level approaches.
Bio:
Chaowei Xiao is an Assistant Professor at the University of Wisconsin–Madison. His research focuses on building secure and trustworthy AI systems. He has received several prestigious awards, including the Schmidt Science AI2050 Early Career Award, the Impact Award from Argonne National Laboratory, and various industry faculty awards. His work has been recognized with best paper awards including the USENIX Security Distinguished Paper Award (2024), ACM Gordon Bell Prize Finalist (2024), ACM Gordon Bell Special Prize for HPC-Based COVID-19 (2023), the Best Paper Award at the International Conference on Embedded Wireless Systems and Networks (EWSN) (2021), and the MobiCom Best Paper Award (2014).
Dr. Xiao’s research has been cited over 14,000 times according to Google Scholar and has been featured in multiple media outlets such as Nature, Wired, Fortune, and The New York Times. Additionally, one of his research outputs was exhibited at the London Science Museum.