Behind Nurses' AI Ethics Awareness: Moral Sensitivity, Reflection and Digital Stress
The study examines whether recognizing ethical concerns, reflecting on personal judgments and experiencing digital stress are connected with awareness of AI ethics.
An AI recommendation can give a nurse more information, but deciding whether that information deserves trust brings questions about fairness, privacy and responsibility into everyday care. A prediction may look convincing without explaining how it was reached or whether it reflects the needs of a particular patient. Recognizing those concerns requires attention to the people whose care could be affected.
Research by Gyuli Baek and Eunju Lee, published in Healthcare, explores how nurses' ethical awareness relates to their thinking habits and the pressures of working with technology. The original study is titled "Associations of Moral Sensitivity and Metacognition with Artificial Intelligence Ethics Awareness Among Clinical Nurses: The Potential Mediating Role of Technostress." It examines whether recognizing ethical concerns, reflecting on personal judgments and experiencing digital stress are connected with awareness of AI ethics.
The survey of 120 nurses in South Korea found positive relationships between AI ethics awareness and all three characteristics. Its most delicate finding involved a possible indirect connection between moral sensitivity and ethical awareness through technology-related stress. The results describe patterns in questionnaire responses, without proving that stress improves ethical judgment or that greater awareness produces safer patient care.
Inside the Survey of 120 Nurses
The researchers collected online responses from nurses at four tertiary hospitals in Daegu between 10 and 31 October 2025. Participants needed at least six months of nursing experience and had to identify themselves as familiar with AI-related technologies relevant to nursing. Of 124 questionnaires collected, four contained missing responses and were excluded.
Women accounted for 95% of participants, nearly half were aged 30–39, and average nursing experience was about 12 years. Their roles covered general wards, specialist units, outpatient services and other nursing departments. The survey did not establish which AI systems they had used, how frequently they used them or whether their familiarity included direct clinical experience.
Questionnaires measured moral sensitivity, meaning the ability to recognize ethical concerns, and metacognition, which involves checking and adjusting one's thinking. A separate measure assessed technostress, including digital overload, complexity, uncertainty, anxiety and technology intruding into personal life. The AI ethics questionnaire covered accountability, safety, fairness, transparency, human-centered services, employment, permission and limitations, and robot rights.
Ethical Awareness Was Linked to Reflection and Sensitivity
Average AI ethics awareness was 3.46 out of five, with metacognition averaging 3.94 and technostress 3.30 on their respective five-point scales. Moral sensitivity averaged 146.27 out of a possible 180. These scores describe participants' self-reports and cannot establish how well they would handle an ethical problem involving AI during patient care.
Higher AI ethics awareness was associated with greater moral sensitivity, more reflective thinking and higher technostress, with correlation values of 0.36, 0.33 and 0.22, respectively. Moral sensitivity had the strongest of these three relationships and was itself connected with metacognition and technostress. Metacognition showed no clear simple relationship with technostress, suggesting that the characteristics did not all move together.
Nurses satisfied with their work reported greater moral sensitivity than those who felt neutral or dissatisfied, and stronger metacognition than those who felt neutral. Technostress was higher among nurses aged 40–49 than those aged 20–29, and among nurses with at least 20 years of experience compared with those with under five or 10–14 years. Women reported more technostress than men, but only six men participated, and the comparisons found no clear differences in AI ethics awareness across demographic or employment groups.
The Stress Finding Needs Careful Reading
Using models that repeatedly resampled the survey data 10,000 times, the researchers explored whether technostress formed an indirect connection between personal characteristics and AI ethics awareness. The moral-sensitivity model produced a small positive indirect estimate of 0.058, with an uncertainty range that excluded zero. Accounting for age, experience and sex reduced the estimate to 0.043, bringing the lower end of that range close to zero.
Removing one participant at a time exposed the finding's fragility, with 31 of the 120 repeated analyses producing uncertainty ranges that included zero. The estimates stayed positive and similar in size, but confidence in the finding depended on which individual responses were included. The study offers preliminary evidence of a relationship rather than a dependable explanation of how ethical awareness develops.
Metacognition showed a direct positive relationship with AI ethics awareness after technostress was included in the model, despite an unclear total association in the model accounting for moral sensitivity. Its proposed indirect connection through technostress was not supported, and moral sensitivity's direct association became less clear after accounting for stress. The combined model explained 18.1% of differences in AI ethics awareness, leaving much of the variation unexplained.
What Nursing Teams Can Take From the Research
The stress questionnaire measured demands from general information and communication technologies, rather than pressures specifically caused by AI. It could not separate manageable challenges from harmful distress or show whether awareness preceded stress, stress preceded awareness, or other factors influenced both. These findings provide no basis for increasing nurses' stress as a way to strengthen ethical awareness.
A single survey from four hospitals in one region offers a limited picture, especially with voluntary recruitment and small participant subgroups. Self-reports can reflect socially desirable answers, and overlap between the ethical questionnaires may partly explain their relationships. Missing hospital identifiers prevented examination of hospital-level differences, and the sample-size calculation did not establish whether the study could reliably detect small indirect relationships.
The researchers identify case-based ethics teaching, reflection on reasoning and support for managing technological demands as approaches worth testing. Larger studies following nurses over time would need detailed measures of AI use, AI-specific stress and observed decisions to establish whether these approaches help. The practical question is whether greater ethical awareness translates into more thoughtful use of AI and better protection for patients, outcomes this study did not measure.
- FIRST PUBLISHED IN:
- Devdiscourse
ALSO READ
-
Paris Event Puts Human Rights, Inclusion and Information Integrity at Centre of AI
-
Viet Nam Puts Workers and Business Productivity at the Heart of AI Transformation
-
Indian steelmakers rank among top 10 most competitive globally: Ashok Sahay
-
SPZAORA: Transforming Interior Design for Architects Through the Power of Artificial Intelligence
-
EU Job Quality Improves as ILO Calls for Fair Wages and Responsible AI Use
Google News