Analysis of Grounded System Theory to Dilemmas of Artificial Intelligence

Abstract

This paper aims to uncover hidden feedback loops in the dilemmas of artificial intelligence. Applying the grounded system theory method to interviews with ChatGPT, this paper performed open and axial coding, and conducted analysis using loop coding instead of selective coding as in traditional grounded theory. Through this analysis, a circular paradigm model of “dilemmas of artificial intelligence” is constructed. Overall, it is confirmed that artificial intelligence sufficiently understands dilemma situations. Additionally, it is pointed out that there is a vicious cycle where “bias” is reinforced through the “learning” of artificial intelligence. Furthermore, AI is concerned about a loop where decision delays under a dilemma exacerbate the dilemma. And a causal loop diagram with clearer causal relationships is constructed based on the circular paradigm model. This study proposes that the circular paradigm model constructed through the procedures of grounded system theory can help uncover hidden feedback structures in interview texts.