Differentiated Coherence in Chinese News Media

Abstract

This study examines how sociotechnical imaginaries of artificial intelligence (AI) in cultural heritage are organised across institutionally differentiated Chinese news media and reproduced over time. Drawing on nine outlets from 2022–2025, it combines exploratory computational mapping, thematic analysis, relation-focused lexical indicators, independent double coding, and temporal modelling. The primary analysis uses a 232-case high-precision semantic core, checked against a 357-case sensitivity set. The findings identify a recurring normative orientation in which AI is positioned as culturally beneficial through two overlapping configurations: cultural revival and memory infrastructure. Cultural revival is widely shared across analytical positions, whereas memory infrastructure shows the clearest recurrent position-linked differentiation. Public actors are often visible, yet no case meets the strict criterion for explicitly attributed public definitional authority. Temporal tests show persistence in the broader orientation alongside uneven plasticity across articulation modes. The study develops differentiated coherence as a communication-level concept for explaining how a sociotechnical imaginary can remain recognisable through structured selective rearticulation, visibility–authority decoupling, and discursive reproduction without requiring uniform messages or implying causal institutional effects.