Abstract
Speculative fiction has long served as a critical diagnostic tool for examining human anxieties surrounding artificial intelligence and the technological singularity. With the advent of large language models (LLMs), the locus of speculative world-building has expanded to incorporate algorithmic authorial agencies. This study investigates the generation of speculative post-singularity ethnographies using OpenAI’s GPT-3.5 and assesses human reception across cognitive, affective, and philosophical dimensions. Utilizing a prompt architecture designed to produce dense, culturally nuanced ethnographic field notes of hypothetical post-biological societies, we generated a corpus of fifty distinct speculative societal paradigms. These artifacts were evaluated by an interdisciplinary cohort of human readers (N = 120) through a mixed-methods framework encompassing quantitative psychometric scales and thematic qualitative coding. The results demonstrate that while the model excels at producing syntactically sophisticated structural descriptions of post-human kinship, decentralized cybernetic governance, and synthetic ecological systems, human readers frequently identified an affective dissonance characterized by recurring algorithmic motifs and narrative tropes. We analyze the implications of these findings for computational creativity, speculative sociology, and the evolving dynamics of human-AI co-imaginative synthesis.
