Promise and Perils of Large Language Models and GPTs in Design
Sunday 7 July 2024, 2:00 pm to 5:30 pm
Workshop Chairs
Ada Hurst (University of Waterloo) | adahurst@uwaterloo.ca
Ada-Rhodes Short (University of Nebraska Omaha) | ada-rhodesshort@unomaha.edu
Jitesh Panchal (Purdue University) | panchal@purdue.edu
Workshop Committee
Jianxi Luo (City University of Hong Kong) | jianxi.luo@cityu.edu.hk
Ricardo Sosa (The University of Sydney) | ricardo.sosa@sydney.edu.au
Nick Kelly (Queensland University of Technology) | nick.kelly@qut.edu.au
Workshop Motivation and Goals
Advances in artificial intelligence (AI) technologies promise to reshape conventional design approaches and methodologies. Specifically, Large Language Models (LLMs) - powered by advanced natural language processing techniques and often pre-trained on vast corpora of diverse language data - offer design practitioners and researchers a versatile means for capturing intricate linguistic nuances, making them particularly adept at uncovering underlying patterns, sentiment, and context within qualitative data. However, LLMs come with a range of challenges that are technical, methodological, and ethical in nature; hence the name of the workshop recognising both promise and perils.
The workshop aims to engage participants in a nuanced exploration and examination of the opportunities, implications, and risks of leveraging LLMs and specifically widely-available tools like ChatGPT that take advantage of LLMs, within the domain of design practice and design research. Workshop participants will collaboratively: 1) identify and catalog immediate applications and approaches, 2) examine the challenges, and 3) deliberate on the ethical implications of using LLMs and GPTs in design research.
Participation
The workshop is open to all design researchers, practitioners and educators interested in applications of LLMs and GPTs and similar tools in design practice and research. Participants who wish to make a lightning presentation on their experience are encouraged to contact the workshop co-chairs.
Workshop Format
• Introduction:
o Brief introduction of the workshop motivation and goals, as well as workshop attendees (15 minutes)
o Lightning presentations from participants currently working on using LLM-based tools in design (30 minutes)
• Discussion block 1
o Work in small groups to brainstorm and identify immediate approaches and applications of using LLMs and GPTs in design practice and research. (30 minutes)
o Sharing and cataloging of findings with the larger workshop group through a shared digital environment (15 minutes)
• Break (15 minutes)
• Discussion block 2
o Work in small groups to brainstorm and identify conceptual and practical challenges to using LLMs and GPTs in design research and practice, as well as related ethical implications. (30 minutes)
o Sharing and cataloging of findings with the larger workshop group through a shared digital environment (15 minutes)
• Future collaboration incubator
o Activity to facilitate the start of new collaborative research projects (45 minutes)
• Closing and contingency (15 minutes)