My research investigates how unstable, contested concepts can be formally defined and represented as semantic spaces—using "cult" as a primary case study. The project originated from asking whether AI psychosis could be understood as a form of sectarian drift.