Can LLMs Keep a Secret? - Privacy and Security in Large Language Models

Explore privacy and security concerns surrounding Large Language Models (LLMs) in this guest lecture delivered by Niloofar Mireshghallah, examining the critical question of data confidentiality in AI systems. Delve into the challenges and implications of protecting sensitive information when working with LLMs, understanding potential vulnerabilities and safeguards needed to maintain data privacy. Learn about cutting-edge research and methodologies for ensuring secure AI implementations while gaining insights into the balance between model performance and privacy preservation.