Operationalising Conflicting Values in AI-enabled Healthcare: A Reflection on Current Challenges – Alessandro Bozzon (TU Delft)

The deployment of AI in healthcare presents a unique set of challenges when compared to other sectors. These challenges are due in part to the many values that are in conflict when AI is used in healthcare decision-making. In this presentation – and in the following discussions – we will reflect on the current state of AI in healthcare and the challenges that need to be addressed in order to operationalise conflicting values in AI-enabled healthcare. We highlight some key issues that emerge from a purely technical conceptualisation of the AI design, development, evaluation, and auditing processes, where the situated nature of (clinical) decision making, especially when supported by AI systems, is ignored. We will argue that design research, design methodologies, and human-computer interaction are key disciplines that can help to address the challenges of operationalising conflicting values in AI-enabled healthcare.