The “Meta-intermediary” of News Access: The Reconstruction of Journalistic Authority in the Age of Generative AI

Abstract

Generative AI, through a four-fold mechanism of “query translation-cross-source retrieval-content integration-discursive reformulation,” packages multi-source information into coherent dialogues with itself as the narrative subject, thereby ascending to the position of a “meta-intermediary” situated above existing platforms. This process accomplishes the “re-intermediation” of already-mediated information. It not only reshapes the pathways of information access but, more profoundly, reconstructs journalistic authority. Beyond the traditional “institutional authority” and “platform authority,” a new “model authority” emerges, based on anthropomorphic interaction and cognitive outsourcing. These three form a dynamically contested triangular structure, leading to structural impacts such as the narrowing of public agendas (agenda circuit-breaking), the stripping of news brand value (white-labeling), and the blurring of accountability. Accordingly, this paper argues that the governance paradigm must evolve from content control to the holistic regulation of cognitive structures. It proposes collaborative governance pathways, including constructing an “auditable chain of meaning generation,” designing ecosystem-level accountability and value compensation mechanisms, and empowering news organizations to become dominant interpreters. This study seeks to move beyond the technical discourse of “channel substitution,” offering new theoretical perspectives and policy insights for understanding the reorganization of power within the news ecosystem and the safeguarding of publicness in the era of intelligent communication.