Zhaoqian Xue

Characterizing User-Reported Risks across LLM Chatbots

As Large Language Models (LLMs) become increasingly integral to daily life, users are engaging with multiple LLM chatbots for various needs; however, prior research on LLM risks often remains lab-based or focuses on single LLMs like ChatGPT or singular risks like privacy. To gain a multi-risk, cross-chatbot understand…