Abstract: GPT (Generative pre-training transformation) language models, represented by ChatGPT, show the ability to simulate human mind. Their continuous optimization and technological iterations makes it possible for them to become an indispensable part of people’s daily work and life. Once GPT language models start to have the human-like mind and obtain industrial application, they will bring new productivity like never before, and even reshape the basic pattern of human society. However, GPT language models not only promote productivity, but also incur such problems as job replacement and derivative risks resulting from extremely rapid technological upgrading. The integration of GPT language models and their derivative applications into social production and social life should be a gradual and steady process, so as to create buffer time and avoid excessive impact on society

Key words: GPT, language model, human-like mind, employment, risk