Abstract for: Feature balance of scale and scope of data in AI platform firms

This study focuses on AI platform firms that, since the commercialization of deep learning with big data, have positioned and used the computational power of artificial intelligence (AI) as a core business function or mainstream product or service. This study argues for a cyclical structure that increases the scale and scope of data, enabling the exponential growth of AI platform firms. Therefore, we develop qualitative and dynamic models based on the scale and scope of data and investigate the mechanism of the exponential growth of AI platform firms. First, the simulation of AI platform firms was executed using a set of Julia packages, and the reproducibility of the execution results was verified using Vensim, a system dynamics development environment. Second, the sensitivity analysis of the dynamic model of AI platform firms was performed using the data network effect strength and the data sharing rate as parameters, and contour plots of the data boundary rate values as indexes of the scale and scope of the data were generated. Furthermore, through linear/nonlinear regression estimation that approximates the results of sensitivity analysis, we attempt to gain a qualitative and quantitative understanding of the feature balance between the scale and scope of data.