Abstract: [Purpose/Significance] The large models of artificial intelligence pose challenges and transformations to the basic theories of archival science, and it is urgent to explore how the large models can drive the innovation and advancement in the basic theories of archival science. [Design/Methodology] By comparing the key points of the basic theories of archival science in the traditional era and the era of artificial intelligence, this paper elaborates the response of the basic theories of archival science to the challenges of the large models, and further explains the question "How do the large models of artificial intelligence affect the basic theories of archival science? And how should the basic theories of archival science respond to the challenge posed by the large models of artificial intelligence?" [Findings/Conclusion] The large models of artificial intelligence have brought challenges to the basic theories of archival science, such as the narrow concept of the principle of provenance, the extension of the records life cycle theory, the transformation of archival value, the lack of credibility in archival appraisal theory, the pin but not destroy, and the crisis of the subject of the construction of the archival memory concept. In response to these challenges, the basic theories of archival science need to undergo self-adjustment and innovation: rediscover the concepts and contexts of the new provenance perspective, emphasize the dynamic continuity of the records life cycle, focus on the credibility assessment in archival appraisal theory and the right to be forgotten, regulate the power of algorithmic, and assert control over the discourse of memory. [Originality/Value] From a theoretical perspective, this article re-examines the challenges posed by large models to the basic theories of archival science, and how the basic theories of archival science should self-correct to respond appropriately, so as to provide theoretical guidance for archival work in the era of artificial intelligence.

Keywords: Artificial intelligence, Large models, Basic theory of archival science, Archival work