In the report AI-assisted analysis of war-related content on grey zone domains, researchers of Lund University Psychological Defence Research Institute examines under-resesearched digital environments that play a role in contemporary information influence operations. In line with earlier conducted research exploring non-traditional information influence dissemination through for example video game platforms, this report focuses on so-called “grey zone domains”: web platforms of dubious legality that host and aggregate violent content including murders, torture, suicides, terrorism and accidents. Often sourcing material from private messaging apps, closed forums, social media, and the dark web, these sites enable content to be rapidly reshared across the wider digital ecosystem, repeatedly challenging platform moderation systems. Despite their apparent importance—particularly in relation to disturbing content from conflicts such as Russia–Ukraine and Israel–Palestine—grey zone domains remain a poorly understood component of information influence infrastructure, with limited empirical data on their scale, actors, and impact.
To address this gap, the study develops and tests an AI-assisted methodology designed to analyse grey zone domains while minimising analyst exposure to harmful material. The system combines automated web crawling, AI-based content classification, and a customisable dashboard that allows detailed filtering and behavioural analysis. As a proof of concept, the pilot study examines one grey zone domain hosting both gore and pornographic content, finding that around 17% of classified posts related to ongoing conflicts and that some individual posts achieved hundreds of thousands of views through cross-platform sharing. The results demonstrate both the scale of the phenomenon and the potential of AI-enabled approaches to support research, counter-FIMI efforts, and law enforcement, while significantly reducing human exposure to traumatic content.