Human or machine? discursive constructions of authenticity in AI-generated news and public communication in Indonesia

Abstract

Generative artificial intelligence is reshaping news production and public communication by complicating established relations among authorship, credibility, transparency, and institutional responsibility. This study investigates how Indonesian governmental, regulatory, and media institutions discursively construct authenticity in documents concerning AI-assisted journalism and public communication. Using a qualitative corpus-assisted design, this study analyses fifteen public documents published between 2023 and 2026 through Fairclough’s critical discourse analysis, van Leeuwen’s social-actor framework, and multimodal discourse analysis, with metadata, textual evidence, and coding stored in linked datasets. Findings show that authenticity is constructed primarily through institutional authority, human verification, source traceability, ethical disclosure, and professional accountability. Human-centred authenticity dominates journalism-related documents, whereas public-communication texts more frequently combine innovation, partnership, official-data authorization, and risk framing. AI is widely granted productive agency as a tool, collaborator, presenter, or intermediary, but responsibility remains assigned to journalists, media organizations, government institutions, and platform operators. Contextual variation further indicates that news discourse protects epistemic authority, while governmental communication more readily legitimizes AI as an administrative resource. This study contributes a relational model of authenticity that moves beyond binary human–machine authorship by integrating discursive legitimation, distributed agency, institutional context, and multimodal representation within an Indonesian communication setting amid rapid technological institutional change.