【Purpose/significance】To construct a new cross-modal knowledge service model based on big language model, which is

helpful to quickly adapt to the new paradigm of knowledge interaction led by ChatGPT and improve the work and learning efficiency of teachers and students in colleges and universities.【Method/process】This paper focuses on serving the personalized needs of users, sorts out the characteristics and limitations of the knowledge service model of colleges and universities at different stages, builds a set of digital life knowledge service system with learning ability, and explores a question-and-answer knowledge service model, and evalu⁃ates the service effect of the model in personalized needs through research.【Result/conclusion】The question-and-answer knowledge service model with digital life system as the core improves the personalized knowledge acquisition efficiency of teachers and students in colleges and universities, and provides a decision-making reference for the virtualization and intelligent knowledge service system.【Innovation/limitation】Digital life defines the two-way reshaping mechanism between users and platforms, and provides the possibil⁃ity for self-supervised learning of digital life through multimodal knowledge services. However, there are still many technical chal⁃lenges, such as the cognitive ability of language models, the common interface of experimental tools, and the regulatory constraints of information security.