Ontologies in Design: How Imagining a Tree Reveals Possibilites and Assumptions in Large Language Models

TL;DR

This paper reframes the analysis of sociotechnical AI systems by foregrounding ontologies—the underlying assumptions about what exists and what is possible—as a design concern alongside traditional axiology. It introduces four ontological orientationsPlurarism, groundedness, liveliness, and enactmentto guide practice-based analyses and demonstrates their utility through two probing exercises on LLMs: prompting multiple chatbots and examining the Generative Agents architecture. The analyses reveal that ontological diversity is often hidden within training data and architectures, with Western, fixed, and prescriptive representations tending to predominate, limiting imaginative possibilities. The work argues for integrating ontological inquiries throughout the LLM development pipeline, highlights methodological pathways for surfacing alternatives, and outlines future directions for designing sociotechnical systems that better reflect plural, situated, dynamic, and enacted visions of intelligence and agency.

Abstract

Amid the recent uptake of Generative AI, sociotechnical scholars and critics have traced a multitude of resulting harms, with analyses largely focused on values and axiology (e.g., bias). While value-based analyses are crucial, we argue that ontologies -- concerning what we allow ourselves to think or talk about -- is a vital but under-recognized dimension in analyzing these systems. Proposing a need for a practice-based engagement with ontologies, we offer four orientations for considering ontologies in design: pluralism, groundedness, liveliness, and enactment. We share examples of potentialities that are opened up through these orientations across the entire LLM development pipeline by conducting two ontological analyses: examining the responses of four LLM-based chatbots in a prompting exercise, and analyzing the architecture of an LLM-based agent simulation. We conclude by sharing opportunities and limitations of working with ontologies in the design and development of sociotechnical systems.