AI adoption and the innovation of Chinese hidden champion firms: From the perspective of dynamic capabilities

Summary:  View help for Summary As the underlying technology of the latest round of scientific and technological revolution and industrial transformation, artificial intelligence has been widely integrated into various fields of economic and social development, greatly promoting rapid development, extensive radiation and comprehensive penetration of the digital economy and bringing unprecedented opportunities and challenges to numerous Chinese hidden champion firms. By constructing panel data of 547 Chinese hidden champion listed companies in the A-share market from 2016--2022 and using the fixed effect model, the impact of AI adoption on corporate innovation performance and the mediating role of dynamic capabilities have been empirically tested. Moreover, by introducing managerial autonomy as a key contextual factor, the moderating role of managerial autonomy in the corporate innovation process has been explored. The research findings are as follows: (1) AI adoption can significantly promote innovation performance; (2) innovative capability, absorptive capacity and absorptive capacity play partial mediating roles between AI adoption and innovation performance; and (3) organizational autonomy and individual autonomy within managerial autonomy significantly moderate the promoting effect of AI adoption on dynamic capabilities. Robustness tests further support the above findings and reveal the time lag of the impact of AI adoption on innovation performance. In this way, it provides novel empirical evidence for further understanding the relationship between AI adoption and corporate innovation performance and offers both theoretical references and practical implications for accelerating the adoption of AI by many SMEs.

Geographic Coverage:  View help for Geographic Coverage China

Time Period(s):  View help for Time Period(s) 2016 – 2022

Data Type(s):  View help for Data Type(s) aggregate data; other

Data Source:  View help for Data Source This study initially considered over 98,000 Chinese hidden champion firms identified by the Ministry of Industry and Information Technology as of December 2023. The data collection process involved several steps: first, only listed companies were selected; second, firms under special treatment (ST, *ST) or delisting were excluded; and finally, only those with complete data from 2016--2022 were retained. Patent application data were sourced from the CNRDS database, AI word frequency data from the Sens database, and basic firm and industry data from the CSMAR database. After data processing, the final effective sample consisted of 547 enterprises, resulting in 3,589 enterprise-year nonequilibrium panel observations.