In the fields of healthcare and health research, there is particular interest using machine learning (ML) to generated novel or refined diagnostic, prognostic, risk, and treatment categories. This talk interrogates the nature of these categories and their implications for the people thus (re)categorised. It approaches these questions through the lens of the philosophical idea of ‘human kinds’. It asks to what extent health-related categories generated by ML might function as human kinds and, if so, whether they might differ, in ethically significant ways, from socially-originating kinds. In doing so, it suggests that our understanding of responsible ML categorisation practices need to look beyond technical capabilities and clinical utility to consider wider personal and social impacts.

Please note limited seats are available at Inspace for in-person audiences, so please book tickets in advance. For those joining online please visit the online event page for the Zoom joining link and password.

*Important Notice* This event will be photographed, recorded and live streamed and the data published online and used for research, promotional and reporting purposes by BRAID UK and the Institute for Design Informatics based at the University of Edinburgh. For further information please contact the organisers.