企业人工智能发展水平测度与分析OA

Artificial Intelligence(AI)is widely recognized as a core engine of economic growth.However,in-depth research on its development patterns and heterogeneous characteristics at the micro-enterprise level remains scarce due to challenges in data acquisition and measurement.Innovatively,this study employs web-scraping techniques to collect annual reports of listed companies from 2007 to 2020,which are further processed into sentence-level coropora for meticulous manual annotation.Subsequently,pre-trained language model BERT is utilized for training and supervised fine-tuning to construct a multidimensional evaluation system.This system assesses the development of AI across the entire enterprise spectrum,as well as specifically within the infrastructure,technology,and application layers,with model performance validated via a confusion matrix.The findings indicate that at the enterprise level,the development of AI in the technology layer lags relatively behind,with significant internal disparities among firms.Specifically,smaller,long-established,and non-state-owned enterprises(non-SOEs)exhibit higher levels of AI development.Geographically,enterprises in five major eastern provinces and cities,including Beijing and Shanghai,lead in AI advancement,while regional development patterns across provinces demonstrate a phased characteristic of initial divergence followed by convergence.In terms of sector-wise analysis,the enterprises in information transmission,software,and IT service industries shows the highest penetration of AI.Manufacturing firms are the most mature in AI application,whereas the enterprises in agriculture,forestry,animal husbandry,and fishery industries remain relatively weak in both the breadth and depth of AI integration.This study reveals the current status and trends of corporate AI development,providing a robust empirical basis for governments to formulate precise AI policies and for enterprises to advance their digital transformation.