Healthcare sciences lecturers' views on the use of artificial intelligence for patient diagnosis in Gauteng province, South Africa: qualitative study

Abstract

Introduction Artificial intelligence (AI) research is critical in healthcare to improve accuracy of diagnosis, optimize operations, and customize treatment approaches, resulting in better patient results. The aim of this study was to explore perceptions and acceptability of AI tools among healthcare sciences lecturers when students use them for patient diagnoses in a selected higher education institution (HEI), Gauteng Province, South Africa. Methodology The study used a qualitative, exploratory-descriptive approach. A purposive sampling technique was used to select healthcare sciences lecturers. Data was collected through semi-structured one-on-one interviews. An interview guide with open-ended and probing questions was used. All interviews were conducted in English, recorded using a digital recorder and later transcribed verbatim. Data were analysed using thematic content analysis and NVivo 14. Results Thirteen themes emerged during the data analysis process. Healthcare sciences lecturers were concerned about accuracy in diagnosing, ethical and human oversight, lack of emotional connection and Ubuntu (humanity towards others) ; some had apprehension and discomfort with AI use; regardless, the majority appreciated AI as a tool for enhancing education and in specific healthcare areas; AI improving efficiency in education and diagnosis; the usefulness of AI tools in general and the importance of AI tools usability in research and education. Lastly, healthcare sciences lecturers stated that the influence of AI use depends on generational influences on AI adoption; the influence of pandemics and networks on AI acceptance; and the need for institutional support and training, as there is limited experience with AI tools in clinical settings. Conclusion AI is a sensitive topic for many professions regarding its use for diagnostic purposes. The majority of healthcare sciences lecturers are interested in using AI for diagnostic reasons, but only if properly educated. However, some believe it can only be used for teaching and learning, not for diagnosis. Inconsistencies exist in instructors' judgments and acceptability of AI tools for students diagnosing patients. Barriers include concerns about humanity, errors, misdiagnosis, suppression of critical thinking, and production of future lazy practitioners. Therefore, this institution's faculty id health sciences should train and support staff in using technological tools without losing the human touch.