Using locally-hosted Small Language Models (SLMs) to protect student, patient and research subject data in Health Professions Education

Abstract

Cloud-based Large Language Models (LLMs) are being increasingly used for Health Professions Education (HPE) teaching and research. A major concern is data privacy, resulting in a potential exposure of student, patient, and research participant data. Using locally-hosted <i>Small</i> Language Models to assist the teacher and researcher in completing tasks while minimising the risk of data exposure. The main ethical task is achieved, but there may be limitations. In addition, technical details are given to assist in the effective implementation of the solution. Further detailed research in a range of environments will demonstrate their practicability, especially as the technology improves. The implications are far broader than the focus on research and teaching covered in this article.

Ähnliche Arbeiten