Large language models for digital product passports

The launch of advanced Artificial Intelligence (AI) tools like ChatGPT has broadened the availability for applying Large Language Models (LLMs) within user research in the manufacturing sector. This presentation examines the potential of using LLMs to mimic the roles and responsibilities of individuals involved in the Digital Product Passport (DPP) lifecycle. By simulating the decision-making processes and duties of crucial stakeholders such as Technicians and Materials Scientists, LLMs attempt to provide researchers with realistic responses that emulate expert behaviour under specific parameters and conditions.

This augmented user research approach facilitates an ongoing research dialogue, offering the potential to bridge the gaps caused by limited workforce availability and generating insights typically acquired from direct expert interactions. Building on our experience using LLMs to simulate roles, events, and decision-making processes within the DPP lifecycle, this study offers insights into how these strategies might be adapted by start-ups and SMEs to expedite their development cycles.

The presentation aims to discuss the potential of LLMs to supplement user research within the manufacturing sector, focusing on their utility in enhancing the R&D processes for start-ups and SMEs, while critically assessing LLMs inherent constraints and potential gaps of their training knowledge.