This paper aims to address the limitations of traditional Named Entity Recognition (NER) approaches in the task of knowledge triplet extraction, proposing an innovative architecture based on the Artificial Intelligence Generated Content (AIGC) model. Traditional NER schemes typically extract contiguous words or phrases from provided text as outputs. However, they often fail when a substantial number of the extracted entities do not appear explicitly in the original text. To tackle this issue, we leverage the strengths of the AIGC model and devise a novel input processing strategy.The AIGC model we propose harnesses the characteristics of rare Chinese characters from a native Chinese pre-trained model, coupled with an auxiliary list of relational terms, to concatenate them with the original text, forming a new input sequence. This approach enables the model to handle the extraction of knowledge triplets in generative tasks, utilizing delimiter words to separate different parts of the triplet. To construct a viable model, we selected a pre-trained model that has been specifically trained and optimized for this purpose.To evaluate the performance of the model, we conducted extensive experiments. The results indicate that our AIGC model performs well in the task of knowledge triplet extraction, particularly in scenarios where the output entities do not appear in the original text.