A recent article by Dr. Iain Armstrong and Peter Winter examines a critical blind spot in digital health ethics known as “misrecognition”, which occurs when artificial intelligence systems evaluate patients solely through easily quantifiable metrics. By prioritizing standardized data over the complex, subjective realities of living with a chronic illness, AI tools inadvertently foster epistemic injustice, treating patient and caregiver insights as lower-status knowledge. This rigid approach forces deeply human, embodied experiences into narrow clinical boxes where they fit poorly, ultimately failing to capture the full scope of a patient’s health reality.
Grounding their argument in a case study of pulmonary hypertension (PH), the authors demonstrate how AI applications can act as risk multipliers. While pulmonary hypertension involves highly measurable clinical data points, its daily toll on independence, energy, and family life is profoundly qualitative—aspects captured well by tools like the emPHasis-10 questionnaire but often ignored by algorithms. When AI systems institutionalize narrow proxies like medical costs and hospital utilization rates, they inadvertently dictate what a healthcare system validates as legitimate knowledge, potentially misinterpreting emotional distress or systemic barriers as patient non-compliance.
To resolve this, the paper outlines a six-step ethical framework designed to reshape future digital health tools through collaborative co-design. This approach urges developers to involve patients early in defining clinical problems, prioritize patient-reported outcome measures (PROMs), and systematically audit algorithms for harms related to misrecognition. Ultimately, the authors argue that the true benchmark of an ethical AI system in healthcare should not merely be its mathematical precision or technical fairness, but rather its capacity to keep genuine human experiences visible and valued within the care process.
Read More at this link on the European Heart Journal – Digital Health
Citation
Armstrong, I., & Winter, P. (2026). The ethical application of artificial intelligence in digital health:
patients’ knowledge, pulmonary hypertension, and the problem of misrecognition. European Heart
Journal – Digital Health, 7(6), ztag087.
Summary of article by Deger Kesimoglu, volunteer for the Alliance for Pulmonary Hypertension
