From principles to practices: the intertextual interaction between AI ethical and legal discourses

Abstract

Abstract The ascendancy and ubiquity of generative AI technology, exemplified by ChatGPT, has resulted in a transformative shift in the conventional human–AI interaction paradigm, leading to substantial alterations in societal modes of production. Drawing on CDA approach, this study conducts a thematic intertextuality analysis of 29 AI ethical documents, and delves into the restructuring of the human–AI relations catalysed by ChatGPT, as well as the complex ethical and legal challenges it presents. The findings indicate that the thematic intertextuality between AI ethical discourse and legal discourse promotes the connection and convergence of narrative-ideological structures, which in turn primarily creates new meaningful texts and ethical frameworks that promote a holistic approach to a good AI society. This research also identifies the importance of integrating law-making efforts with substantive ethical analysis and appropriate discursive strategies to promote the responsible and ethical development of generative AI that benefits society as a whole.