Automated Pain Recognition for Non-clinical Applications: Objectivity, Intelligent Habits, and Epistemic Injustice

Automated pain recognition technology (APR) promises to monitor patients with severe communication impairments on a continuous basis, thereby providing better pain diagnosis and treatment. Increasingly, APR is used in non-clinical contexts, in which users are very much capable of verbalizing their own pain experiences. In this paper, we assess the viability of transferring the technology to the terrain of workplace health management, health insurance benefit programs, fitness, or other well-being applications. In such a setting, not only epistemic conflicts between the self-reports of patients and the automated pain recognition tools may arise, which could result in forms of epistemic injustice. At the same time, automated pain recognition tools could form and digitally support intelligent habits that may reduce cognitive load and enhance bodily awareness. We suggest a dialogic design procedure to ensure these ethical considerations during application design.