Abstract
Human-Centered Design (HCD) often requires different types of expertise, ranging from the specialised knowledge of designers and developers (interdisciplinarity) to the domain expertise of stakeholders (transdisciplinarity). Communication between team members is paramount, but often achieving mutual understanding is challenging. HCD also requires mitigation of potential risks related to the implementation of black-box Generative AI tools, such as Large Language Models (LLMs), which provide both interesting opportunities and critical risks. To the best of our knowledge, little attention has been paid to how LLMs can scaffold conceptual alignment among HCD team members in a way that minimises such risks. This paper presents recent work on fostering conceptual alignment among HCD team members by observing misalignment between team members and LLMs. To do that, the paper reports the case study of an interdisciplinary workshop, highlighting that reflecting critically on LLMs can represent a valuable way to foster mutual understanding in HCD processes. Based on the workshop’s results, the paper also describes a graphical user interface based on 2D projections of LLM-generated sentence embeddings, to foster discussion among HCD team members and progressively attain in-group conceptual alignment. Finally the paper discusses implications and future work, aiming to foster discussion on how, in HCD teams, critical reflection can be leveraged to both minimise risks related to LLMs and how conceptual alignment as a strategy for ensuring LLMs’ human-centricity and trustworthiness.