AI at war on screen: Record politics and mediated memory in cinematic futures

Abstract

Rather than treating popular feature films as predictions of technological futures, this article approaches them as widely circulating cultural texts through which ideas about AI, authority, violence, memory, and evidence become available for public interpretation. It analyses films that place artificial intelligence (AI) in war, violent conflict, or post-conflict recovery, asking how such films shape what counts as memory and evidence in imagined futures. Using an IMDb search, we build a high-visibility corpus and select a balanced core sample of ten films with explicit rules on conflict salience, franchise concentration, and diversity. We code each title for: future configuration and normative closure, AI’s dominant role in conflict (belligerent agent, infrastructure, mediator, or victim), anticipatory mood and control imaginaries, and the memory practices foregrounded (archives and logs, biometrics, testimony, simulation, erasure, and mind upload). Most films depict dystopian futures, yet they often extend present-day security and platform arrangements rather than invent wholly new orders. Across the sample, AI most often appears as infrastructure, and control is repeatedly framed as control of the record: who can collect, authenticate, edit, and replay logs, surveillance feeds, biometrics, and synthetic reconstructions. These narrative choices reshape evidence, accountability, and the imagined conditions of repair after violence. Endings commonly restore governability through selective visibility and strategic silence, narrowing what must be publicly acknowledged. We argue that AI–conflict cinema links anticipation to struggles over proof, making certain forms of exceptional rule and post-conflict settlement feel plausible.