Most research on AI and conflict focuses on military uses, such as autonomous weapons, predictive targeting, and battlefield simulations. This focus assumes a clear divide between military and civilian spheres, but that divide breaks down when AI is built into everyday digital platforms and media. When the same generative models power both military simulations and content creation, and when optimization functions developed for defense are used to structure social media algorithms, the battlefield becomes part of the very infrastructure of our digital lives.
Conflict increasingly operates through the ambient production and circulation of synthetic media optimized for platform engagement, making digital platforms themselves part of the infrastructure of conflict. Ambient conflict is a mode of warfare that extends to the infrastructures of digital platforms.
A key phenomenon within this is slopaganda: AI-generated short-form video content that stages dystopian futures through platform-native formats, including emerging “POV” video genres that simulate everyday life under speculative conditions. Slopaganda is distinct from traditional propaganda through its lack of explicit ideological content or strategic coordination, operating instead through large-scale circulation and platform feedback rather than identifiable actors. As it is circulated at large scale, it creates expectations in the public about conflict, threat, and social order without requiring a coordinating actor.
This project examines AI-mediated influence across interconnected levels, through analysis of content production and systematic cross-platform tracing across Instagram, TikTok, and YouTube. The project investigates the incentive structures rendering speculative content profitable, and uses computational methods to study circulation patterns, audience uptake, and the futures these media make thinkable.
Top image: AI-generated short-form POV videos ("YouWake Up in 2050" format) from the POV_Sensei ecosystem (Instagram/YouTube Shorts, 2025-2026)