The challenges of patient autonomy in the face of the use of artificial intelligence in health

Abstract

Objective: to conduct a critical analysis of how the integration of artificial intelligence in healthcare can impact patient autonomy, addressing issues such as algorithmic paternalism, ethical data governance, and the need for effective regulation. Methodology: a qualitative, critical-narrative review was carried out, structured in six stages: (1) formulation of the research question; (2) search and selection; (3) data extraction; (4) critical analysis; (5) interpretation/discussion; and (6) integrated presentation of findings. Searches were conducted between March 2024 and October 2025 across a range of online sources relevant to health, bioethics, and AI governance. Descriptors were drawn from Descritores em Ciências da Saúde and Medical Subject Headings in Portuguese and English, and were combined using the Boolean operators “AND” and “OR.” Results: the findings show that artificial intelligence offers significant advancements. However, risks to the principle of patient autonomy were identified, especially in cases of low algorithmic transparency or the absence of human oversight. It was also observed that algorithmic paternalism may limit patients’ active participation in clinical decisions, reinforcing the need for ethical guidelines and effective regulations to ensure the safe and person-centered use of artificial intelligence. Conclusion: it is essential that the application of artificial intelligence preserves patient autonomy. The implementation of ethical guidelines, continuous human oversight, and system explainability are crucial to ensuring that technology strengthens — rather than limits — individual control.

Submitted: 02/26/25| Revision: 10/06/25| Approved: 10/07/25