Generative AI as a Catalyst for Participatory Journalism: A Scoping Review of Opportunities, Risks, and Emerging Practices

Abstract

Background: The public release of large language models (LLMs) since late 2022 has repositioned artificial intelligence from a back-office newsroom instrument to a public-facing interface through which audiences increasingly encounter, interrogate, and co-create news. Yet scholarship remains predominantly newsroom-centric, leaving the audience-facing dimension of generative AI comparatively under-theorised. Objectives: This research paper conducts a scoping review to identify all theoretical frameworks and empirical studies about generative AI usage in participatory journalism. The study answers four research questions which investigate theoretical frameworks and core research domains and organisational implementation methods and the complete set of ethical and legal and credibility risks. The research team implemented Arksey and O’Malley’s six-stage framework together with PRISMA-ScR reporting guidelines to search five databases which included Scopus and Web of Science and EBSCO/LISTA and ProQuest Central and Google Scholar for peer-reviewed English-language publications that appeared between 2022 and 2025 and had open or hybrid access. The data collection process involved creating charts which showed author details and publication date and country of origin and theoretical framework and research method and target industry and AI system used and core study topics and results about public involvement. Results: The charted corpus reveals theoretical plurality spanning participatory journalism theory, uses and gratifications, UTAUT, disruptive innovation, and cultural political economy alongside a methodological reliance on qualitative, newsroom-centric designs. The applications group into five main categories which include conversational news interfaces and AI anchors and personalisation and immersive storytelling and accessibility tools but the system faces four major security threats which consist of hallucination and opacity and authorship ambiguity and North South resource asymmetries. Conclusion: Generative AI enables user participation through a conversational interface which allows them to generate content while the system maintains the decision-making power. The shift from user-generated content to AI content requires research about audience preferences and developers need to enforce disclosure systems which prove AI content authenticity.