Generative artificial intelligence is reshaping how newsrooms produce and distribute visual storytelling, yet its application in non-English, resource-constrained media environments remains largely unexamined. This study asks how Kazakh-language journalists can use generative AI to reconstruct socially significant events that conventional reporting cannot fully capture — particularly under conditions of censorship and information scarcity. Drawing on Uses and Gratifications Theory and Sundar and Limperos's MAIN model (modality, agency, interactivity, navigability), we conduct a qualitative case study of Qantar Qursauynda (In the Grip of January), a media project reconstructing Kazakhstan's January 2022 events. Using RunwayML, Midjourney, DALL·E, and Stable Diffusion, the project rebuilt the event from fact-checked open-source photographs, eyewitness interviews, and archival footage, applying a principle of "extreme transparency" that logged every tool, prompt, and disclaimer. Notably, all prompts were composed in English, because generative models perform markedly worse in low-resource languages such as Kazakh. Findings show that generative AI lowers the technical and financial barriers to immersive, emotionally resonant reconstruction and offers an alternative channel for reporting sensitive events. However, audiences frequently could not distinguish synthetic from authentic footage, sharpening the ethical stakes and underscoring the need for standardized disclosure. The study argues that AI journalism is emerging in Kazakhstan as both an opportunity and a governance challenge, and calls for ethical standards, digital literacy training, and culturally and linguistically adapted tools — contributing an empirically grounded, non-Anglophone case to global debates on AI in journalism.

Authors:
Askhat Yerkimbay, SDU University, Kazakhstan
Danagul Naukhan, SDU University, Kazakhstan
Alisher Altay, SDU University, Kazakhstan

About the Presenter(s)
Askhat Yerkimbay, journalism lecturer at SDU and Program Manager at Internews. Interested in AI policy and media literacy in non-English media. Current project: Rapid Media Mapping of Kazakh-language regional media.

Connect on Linkedin
https://www.linkedin.com/in/askhat-yerkimbay-052187a/

See this presentation on the full schedule – Tuesday Schedule

Bookmark