Who Gets Believed When Machines Read Pain? AI, Testimony and Nursing Judgement

Pain remains a persistent and ethically consequential problem in healthcare because recognizing another person's suffering requires judgements about testimony, credibility, and appropriate response. Yet, nursing responses to pain continue to be shaped by uncertainty about what pain is, how it is known, and how it should be responded to. Although nursing scholarship contains extensive work on pain knowledge, attitudes, barriers, and assessment tools, it offers less conceptual clarity on a central question raised by emerging technologies: what happens to nursing judgement when artificial intelligence (AI)-assisted systems claim to recognize pain? This paper argues that AI-assisted pain recognition should be treated as a fallible adjunct to nursing judgement rather than as an objective arbiter of pain. Drawing on nursing philosophy, epistemic injustice, contemporary pain theory, and recent debates on nonverbal assessment and AI, I argue that pain is not merely a signal to be measured but a personal, embodied, and interpretively mediated experience disclosed through testimony, context, and clinical judgement. I further argue that epistemic humility, while necessary, is not sufficient for nursing practice. In response, I develop a nursing framework of relational discernment, grounded in presumptive respect for pain testimony, contextual triangulation, bias-aware credibility judgement, relational accountability, and contestability of algorithmic outputs. This framework clarifies the nurse's role in recognizing and responding to pain under conditions of technological mediation and offers a philosophically grounded basis for more humane and ethically defensible pain assessment and management.