Making Large Language Models More Reliable and Beneficial: Taking ChatGPT as a Case Study

This article suggests practical ways to make large language models more reliable and beneficial by taking ChatGPT as a case study. Specifically, we describe ChatGPT’s workflow and promised services and highlight the perils requiring the immediate attention of ChatGPT stakeholders.