Human–AI Collaboration under Constraint: Mediatized Translation Production and Cross-Cultural Risks - A Case Study of iQIYI International

Amid expanding global cultural circulation and rapid AI-driven transformation, human–AI collaboration has become increasingly central to the translation and localization of Chinese audiovisual content for overseas markets. Drawing on mediatization theory and a socio-technical perspective, this article examines how AI, as a technological medium, restructures audiovisual translation production and how this restructuring shapes cross-cultural communication. Based on a qualitative case study of iQIYI International combining seventeen semi-structured interviews with technology managers and frontline practitioners, internal production documents, and public materials, the study identifies three interconnected dimensions of mediatized restructuring: spatiotemporal, agentive, and institutional. Their interaction produces an efficiency-oriented production architecture characterized by accelerated production, bounded human intervention, and technically encoded rules. This configuration generates three recurrent production-side risks for cross-cultural communication: semantic thinning, contextual and affective fragmentation, and reduced interpretive flexibility. The study argues that cultural discount in the AI era may be shaped not only by linguistic and cultural distance, but also by production-side mechanisms embedded in the institutional organization of translation. It therefore proposes a differentiated translation workflow architecture, stronger human interpretive authority at culturally sensitive points in the workflow, and more negotiable rules and culturally sensitive feedback mechanisms to embed cultural judgment within AI-assisted translation systems. Received: 11th May, 2026; Revised: 13th July, 2026; Accepted: 17th September, 2026