Trustworthy LLMs - Mitigating Issues in Social Bias, Safety and Reliability

Join a thought-provoking lecture by USC researcher Jieyu Zhao exploring the critical challenges and solutions in developing trustworthy large language models (LLMs). Delve into cutting-edge research on auditing models, detecting and mitigating social biases, and understanding LLM decision-making processes. Learn about the dual nature of LLMs - their potential for positive societal impact through enhanced accessibility, communication, disaster response, and public health initiatives, while also examining crucial concerns around accountability, fairness, and transparency. Discover ongoing efforts to create more inclusive and ethically sound LLM practices, contributing to a broader dialogue on responsible AI development and deployment.