The large language model GPT has experienced several iterations and has the advantages of multi-modal text generation, strong reasoning ability and multi-domain generality. The release of ChatGPT and its iteration, Multimodal GPT-4, overturned human perception of Artificial intelligence-generated Content (AIGC) and broke the situation that only some humans can produce knowledge. It promotes the integration of multiple knowledge, gives birth to a new knowledge production mode with human-machine two-way enabling, improves the efficiency and quality of knowledge search and knowledge evaluation, expands the space and choice of knowledge element recombination, provides new tools and means for knowledge production and reuse, and has a transformative impact on the process of human knowledge production and reuse. However, the progress of technology will also bring challenges such as the controversy over academic ethics, the reduction of knowledge authority and the intensification of discourse power imbalance. In the face of a new round of challenges, we should further establish and improve the intellectual property protection system, reduce excessive reliance on AIGC, and highlight the core role of humans in knowledge production and reuse to avoid potential risks and threats.