Environmental risk discourse in the digital risk society: A critical analysis of NewsGPT’s AI-generated news

In the digital risk society, risks are reinforced not only by politicians, the media, and expert discourse, but also by algorithms, data visualisations, automated alerts, and continuously updated warning systems. In this context, Beck’s theory of the world risk society becomes further intensified in the digital era, while Lupton’s emphasis on mediated and embodied risk highlights how risks are increasingly produced, circulated, and experienced through digital infrastructures. Despite this, research on risk communication has largely focused on traditional and digital journalism, AI-generated news remains underexplored. Addressing this gap, this study examines how environmental and disaster risks are discursively constructed in NewsGPT, an AI-generated news platform that presents itself as autonomously producing news through artificial intelligence. NewsGPT was selected as a critical case due to its self-declared automation and visibility within algorithm-driven journalism. Using van Dijk’s critical discourse, twenty-five environmental and disaster-themed articles were analysed through purposive sampling. The findings show that risks are predominantly framed through catastrophisation, naturalisation, and technification, while structural and political-economic causes remain marginalised. Corporations frequently remain unnamed, whereas people and nature are positioned as passive objects of risk. By demonstrating how NewsGPT redistributes agency and responsibility, the study extends risk society theory into the domain of algorithmic news production.