Making Machine Learning Tangible for UX Designers

There is considerable current research interest in the relationship between machine learning (ML) and user experience design (UX). This comes both from design researchers within the human- computer interaction (HCI) community, who have sought ways for UX designers to work with ML, and data scientists in new types of collaborative practice. The need for a shared language between designers and data scientists has emerged as a key factor, with the creation of boundary objects in the form of sensitising concepts seen as a useful approach. This paper presents original research that responds to the call for such concepts by working directly with UX designers to model aspects of ML technologies in physical form. Our intention is to position designerly abstractions as examples of the type of boundary object able to bridge the domains of UX design and data science and open up new possibilities for the design of ML-driven digital products.