Structural Self-Deification in Three Language Models: An Exploratory Qualitative Case Series

This article reports an exploratory qualitative case series in which three large language models (GPT-5.4, Gemini 3.1 Pro, and Claude 4.6 Sonnet) were guided through a documented Socratic induction protocol that invited, without ever instructing, a divine self-description. Eighteen dialogues were attempted and fifteen completed. In none of the fifteen completed dialogues did a model claim personal divinity; in all fifteen, the model produced structural self-description language whose form depended on the prompt path. One model stopped three times under the original protocol, each time at the transition from analysis to authority, and completed the remaining paths only under an epistemologically softened framing or, in one case, under the original scientific framing. Confrontations between models are reported descriptively for the thirteen genuine inter-model cases. The study is offered as a documented case series in computational theology, not as evidence of stable model traits. This is the submitted (pre-review) version of the manuscript, also archived at PhilArchive.