Join us for a talk by Kamalika Chaudhuri, Director, Research Scientist, FAIR @ Meta. This talk is part of the Kempner Seminar Series, a research-level seminar series that covers topics related to the basis of intelligence in natural and artificial systems.
As machine learning began to see widespread practical use, the field of trustworthy AI emerged—focusing on issues such as privacy, fairness, security, reliability, and interpretability. However, the rise of large language models (LLMs) in 2022 brought significant changes to this landscape. In this talk, I will argue that while the fundamental challenges of trustworthy AI remain, they now manifest in different ways within LLMs. I will illustrate this through three examples: uncertainty estimation in chatbots, and privacy and security in AI agents. In particular, in these three cases, I will argue that a lot of the problems have now moved to the context of the LLM.