Send message to Author

Paper short abstract

What happens when AI is imagined not as a chatbot, but as a room? Through a digital ethnography of the Backrooms – an internet horror legend – I trace how users employ the language of procedural generation, faulty copies and synthetic memory to interpret its endless liminal spaces.

Paper long abstract

What if artificial intelligence is imagined not as a chatbot or robot, but as a room? Drawing on a small digital ethnography of reception, I ask how users and critics employ the language of generation, model training, faulty copies and synthetic memory to make sense of the Backrooms (2019–) – an internet horror legend built around endless, unstable liminal spaces. The analysis proceeds in three steps. First, users describe the Backrooms as a procedurally generated environment whose incomplete rules govern movement. Second, they interpret its interiors as a faulty copy: familiar objects are reproduced, but their relations no longer form a usable whole. Third, users extend this account of faulty copying to the memories and human forms that can be traced throughout the Backrooms. These readings translate the opacity of computational procedures into the image of a terrifying system that produces distorted copies of reality in ways users associate with generative AI. This shifts the anthropological question from whether AI can possess a conscious mind to how agency is attributed to systems that reproduce traces of experience and cannot be addressed as persons or held answerable as moral subjects. After these three steps, I discuss how this internet horror form that did not originate in AI imagery can help us understand both how people make sense of AI and what this spatial model clarifies about AI itself.