Stop Mimicking Human Empathy: Designing Empathy as a Mediated, Bounded, and Relational Characteristic of an AI Companion for Emergency Departments
Klein, Jacobe; Sörries, Peter; Mutlugil, Yasemin; Müller-Birn, Claudia – 2026
Emergency departments (EDs) are high-pressure care environments where empathy is needed but structurally constrained. Patients often wait in pain and uncertainty, while clinical staff must prioritize urgent cases within time and resource constraints. In this position paper, we examine how artificial intelligence (AI) can support empathetic ED care without simulating or replacing human empathy. Drawing on prior research, we propose a patient-facing, ephemeral AI companion designed to support low-acuity patients. Rather than expressing emotions, the companion mediates empathy by providing informational support and offering emotionally attuned interactions when staff is unavailable. We frame empathy as a sociomaterial practice by embedding technology in clinical workflows: foregrounding temporality and intentional positioning, we argue, rather than mimicking human empathy, we should design for relational AI that helps to preserve human relationships, and, through this case, engage with the workshop themes of materiality, situatedness, and responsible use of empathy in AI.
