Analysis · Technology

When Nurses Describe Patients’ AI Fears, Whose Views Are Measured?

By AWEI · AI-compiled · Published 2026-10-26 · Analysis prepared 2026-10-26 · 1 source · doi.org

Fear of losing human connection can sound like a finding about patients’ acceptance of AI. In a 2024 study by Moustaq Karim Khan Rony and colleagues, however, that concern came from nurses describing patients. The distinction matters for deployment: staff interviews can identify questions worth investigating without establishing what patients actually think.

Published in Heliyon, the study reported interviews with 20 nurses from four tertiary hospitals in Dhaka, Bangladesh. Participants were deliberately selected and all held at least a master’s degree. Interviews took place in Bengali between December 2023 and January 2024, lasting 50–60 minutes, with translation into English supported by bilingual experts.

The researchers used an interpretive qualitative approach to explore participants’ perspectives. Three researchers independently examined transcripts, then developed shared themes. They checked findings with participants and maintained a record of analytical decisions. Six themes included privacy, professional responsibility, human connection and preparation for ethical AI use.

One revealing distinction emerged within these themes: nurses discussed preserving empathetic communication alongside protecting confidential information. Their concerns extended beyond whether a system worked technically to how care would feel and how responsibilities would be supported. These were expressed priorities, not measured changes in patient experience or evidence that AI caused emotional distance.

The interview questions also repeatedly asked nurses to anticipate future developments. That makes the findings partly a map of expectations and proposed safeguards. Checking interpretations with nurses strengthens confidence that their perspectives were represented; it cannot verify their assumptions about patients. Patients were not interviewed, and the study measured neither clinical outcomes nor privacy breaches. Selection and translation further limit interpretation.

For a hospital evaluating AI, these interviews offer a useful starting point: investigate confidentiality and communication separately, and ask patients directly about their experience. Nurses’ concerns can shape an evaluation, but they cannot substitute for the patient evidence that evaluation needs.

All newsNews archive