Jaybee S Cabañeros
The use of artificial intelligence (AI) in diagnosis, prognosis, triage, and clinical decision-making has intensified concern that patients may be reduced to data profiles, risk scores, and probabilistic classifications. Current ethical approaches emphasize accuracy, privacy, bias, explainability, and human oversight but do not fully address the authority granted to algorithmic representations. This conceptual article develops the principle of non-totalization through a constructive dialogue between Emmanuel Levinas and Pope Leo XIV's Magnifica Humanitas. Drawing on Levinas's accounts of totality, the face, vulnerability, responsibility, and the third, together with the encyclical's theological anthropology of dignity, communion, vulnerable embodiment, accountability, participation, and data justice, the article argues that even an accurate and explainable AI representation becomes ethically problematic when treated as a morally exhaustive account of the patient. Non-totalization therefore requires representational humility, meaningful patient voice, contestability and remedy, attributable human responsibility, and data justice. Three illustrative scenarios involving palliative care prediction, rural triage, and a mental-health chatbot show how statistically useful outputs can foreclose attention to patients' goals, circumstances, relationships, and spiritual or existential concerns. In the Philippine context, loob and kapwa serve as contextual interlocutors that illuminate relational interiority and shared personhood without being equated with Levinasian alterity or treated as universal descriptions of Filipino patients. The framework further attends to unequal access, multilingual communication, digital literacy, workforce limitations, religious plurality, and unequal control over health data. The article concludes that AI may inform clinical judgment without acquiring morally exhaustive authority over the person to whom healthcare remains answerable.