Abstract: Based on the four-dimensional classification framework of Information Resources Management (IRM) research questions, this study targets core literature in the IRM field both from the CNKI and WoS between 2016 and 2026. From three dimensions—conceptual autonomy, topic autonomy, and framework autonomy—it preliminarily explores the distributional characteristics of discourse autonomy in China's IRM discipline under AI penetration. Employing methods such as Large Language Model (LLM) classification, cross-lingual word embeddings, and discourse framework coding, the study draws the following conclusions. Firstly, structural differences in AI penetration: the technological-methodological type experiences the deepest penetration, while the resource-organizational type exhibits the strongest local self-sufficiency. Secondly, conceptual autonomy presents the characteristics of "technological isomorphism and governance differentiation" with the governance-institutional and resource-organizational types constituting the high ground of local concepts. Thirdly, topic convergence shows an unbalanced pattern where three fields are highly synchronized, whereas the resource-organizational field is significantly heterogeneous. Fourthly, framework autonomy emerges as the core advantage of Chinese IRM discourse, with Chinese and English literature forming a composite socio-technical paradigm and a one-dimensional technological-instrumental paradigm, respectively. This study reveals the dialectical unity law of discourse autonomy in China's IRM discipline under the impact of AI, providing a methodological exploration for the quantitative study of disciplinary discourse autonomy, and offering preliminary empirical references for understanding the differentiated discourse construction logic of local disciplines under AI penetration in IRM.

Key words: Information resources management, Independent knowledge system, AI4IRM, IRM4AI, Disciplinary discourse autonomy