Abstract

Designers are increasingly using Generative Artificial Intelligence (GenAI) in design processes; however, knowing how designers use GenAI – especially in professional design practice – is under-explored. This paper presents an ethnographic study of an early-stage design team at NASA that explores the natural variation of GenAI use across team members during a speculative design workflow. We aimed to uncover when, how, and why GenAI tools were or were not employed using ethnographic observations to map the team’s speculative design process and follow-up interviews to provide deeper insights into team members’ interactions (or lack thereof) with GenAI. Through inductive qualitative coding, our analysis revealed three strategies of GenAI use observed among professional engineers and designers – intimate co-design with GenAI, selective delegation to GenAI, and minimal use of GenAI – as well as factors that appeared to influence their decisions whether or not to use GenAI. This study proposes new theory in human-AI collaboration that sheds light on the strategies, rationale, and circumstances under which design professionals do and do not use GenAI. These strategies and factors tied to GenAI use offer insights into when, how, and why professionals use GenAI in design and how GenAI could be built to better accommodate designers.