This study aims to examine how generative AI (GenAI) reshapes revenue logics, newsroom collaboration, ethical risks and audience trust in six Indonesian hyperlocal newsrooms operating under tight financial and governance constraints.
The authors use an explanatory-sequential mixed-methods design that combines a 24-month newsroom–month financial panel analysed with staggered difference-in-differences; workflow logs and ethnographic shadowing of human–AI collaboration; a survey experiment on GenAI disclosure and procedural transparency (n = 600); and a structured ethical content audit.
GenAI adoption is associated with modest gains in advertising and membership income and substantial reductions in routine production time, mainly through a dominant “draft-augment” pattern in which journalists revise and localise model outputs. Hyperlocal outlets use GenAI to open narrow but non-trivial diversification “side doors” such as data-as-a-service products, rather than transforming structural precarity. The audit identifies recurring representational bias, partial plagiarism and factual errors, especially in low-stakes genres, but also indicates that simple safeguards can mitigate some risks. In the audience experiment, procedural transparency that briefly explains human oversight for GenAI-assisted stories performs as well as or better than, no disclosure in sustaining trust and willingness-to-pay.
The study draws on a small number of outlets in one country and relies on proxies for complex constructs; findings are context-bound rather than universal.
The study offers a concrete roadmap for small newsrooms to integrate GenAI without losing control of ethics or trust. It highlights the need to prioritise “draft-augment” workflows, second-pair-of-eyes review and simple internal dashboards that track revenue, risk and audience responses together. Donors, regulators and platform partners are encouraged to fund measurement infrastructures and capacity building, rather than only headline “AI innovation” projects.
The findings show that GenAI can modestly strengthen hyperlocal information ecosystems when combined with human oversight and context-sensitive transparency. In settings marked by news deserts, information scarcity and scepticism towards national media, procedural disclosure about AI use can support, rather than erode, community trust. At the same time, uneven capacity and governance frameworks risk widening gaps between well-resourced and fragile outlets, underscoring the need for equity-focused AI policy in the Global South.
By embedding GenAI into the Digital Business Model Canvas and linking financial, workflow, ethical and audience layers, the article offers rare Global South evidence and a “revenue and risk cockpit” heuristic for capacity-sensitive GenAI governance in hyperlocal journalism.
