AI is rapidly transforming palliative care, enabling earlier identification of patients with unmet needs, improving prognostic estimation, recognising symptom trajectories and supporting clinical decision-making. Much of the contemporary debate has focused on predictive accuracy, explainability, transparency and algorithmic bias. These questions are essential, yet they do not reach the deeper philosophical challenge that algorithmic medicine introduces.
This essay examines whether increasingly powerful predictive systems risk obscuring the nature of clinical judgement itself, arguing that prediction and prudence are distinct but complementary forms of rationality.
Drawing on Aristotle’s concept of phronesis, Pellegrino’s philosophy of medicine and Ricoeur’s hermeneutics of practical wisdom, the essay examines how AI expands medicine’s predictive capacity without replacing the practical reasoning through which clinicians determine what ought to be done for a particular patient. Predictive knowledge remains ethically incomplete until interpreted within the concrete circumstances of an individual’s life, relationships and values.
Building on these philosophical foundations, the essay proposes an Architecture of Prudence for AI-enabled palliative care. Predictive systems should initiate deliberation rather than replace clinical judgement, preserve interpretative space, support the cultivation of practical wisdom and remain accountable to the relational–temporal dignity articulated in the DiRePal model.
The future of palliative care will not be determined solely by the increasing accuracy of algorithms. It will depend on whether healthcare systems remain capable of preserving clinical wisdom as the form of judgement through which predictive knowledge becomes ethically responsible care. AI may change how medicine knows; only practical wisdom can determine how medicine ought to act.