UCLIC Seminar, 24 June 2025, Nava Haghighi

Описание: Title: Ontologies in Design: How Imagining a Tree Reveals Possibilities & Assumptions in LLMs

Speaker: Nava Haghighi (Stanford)
Date: Tuesday 24 June 2025, 15:00-16:00

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 - are a vital but under-recognised dimension in analysing 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 large language model (LLM) development pipeline by conducting two ontological analyses: examining the responses of four LLM-based chatbots in a prompting exercise, and analysing 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.

Bio:
Nava Haghighi is PhD candidate in computer science at Stanford University, advised by Professor James Landay. Her research examines the ontological assumptions - the boundaries of what we allow ourselves to imagine - embedded in the design of sociotechnical systems. A critical technical designer, Nava develops theories and methods for surfacing ontological assumptions in current technological systems and builds technical systems (AI and sensing) that expand the presumed defaults. Nava holds a Bachelor of Architecture from California Polytechnic State University, San Luis Obispo, and a dual Master of Science in Computer Science and Integrated Design and Management from MIT.