The emergence and increased use of large language models (LLMs) in user-facing systems have led to significant privacy concerns. These concerns have been primarily model-centered, exploring how LLMs lead to privacy risks such as memorization or inferring personal characteristics from user content. The authors argue for more research focusing on the human aspect of these privacy issues, such as how design paradigms affect user disclosure behaviors and preferences for privacy controls. The goal is to initiate discussions to outline an agenda for conducting human-centered research on privacy issues in LLM-powered systems.
Publication date: 3 Feb 2024
Project Page: https://doi.org/10.1145/3613905.3643983
Paper: https://arxiv.org/pdf/2402.01994
