OpenAlex · Aktualisierung stündlich · Letzte Aktualisierung: 17.03.2026, 10:24

Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.

Assessment of the Capacity of ChatGPT as a Self-Learning Tool in Medical Pharmacology: A Study Using MCQs

2023·3 Zitationen·Research Square (Research Square)Open Access
Volltext beim Verlag öffnen

3

Zitationen

1

Autoren

2023

Jahr

Abstract

Abstract Background ChatGPT is a large language model (LLM) developed by OpenAI. Drawing upon a substantial textual corpus, ChatGPT exhibits a remarkable ability to simulate human speech. This investigation attempts to evaluate the potential of ChatGPT as a self-directed learning resource, with specific attention to its efficacy in answering multiple-choice questions (MCQs) and furnishing coherent rationale for its responses. Methods The study used 78 MCQs from the Korean Basic Medical Science Comprehensive Assessment (K-BMSCA) for years 2019 to 2021. The test items were translated from Korean to English, and lead-in prompts were engineered to test ChatGPT's ability to generate correct and relevant responses. Each test item produced four questions with different prompts. The questions were submitted to ChatGPT, and the responses were analyzed for correctness, consistency, and relevance. Results The study evaluated the efficacy of ChatGPT in responding to a total of 312 questions with an overall accuracy of 76.0% (237/312). The model demonstrated a commendable performance in addressing questions related to recall and interpretation, whereas its performance in solving problems was comparatively poor. ChatGPT offered correct rationales for 77.8% (182/234) of the responses, with errors primarily arising from two sources: faulty information and flawed reasoning. In terms of references, ChatGPT furnished incorrect citations for 69.7% (191/274) of the responses. While the veracity of reference paragraphs could not be ascertained, 80.4% (41/51) were deemed pertinent and accurate with respect to the answer key. Conclusion While ChatGPT has potential utilities in answering MCQs and generating the correct associated rationale, it shows poor performance in referencing. The limitations of ChatGPT as a self-learning tool should be explained and emphasized to students for a wise usage of AI technology.

Ähnliche Arbeiten

Autoren

Institutionen

Themen

Artificial Intelligence in Healthcare and Education
Volltext beim Verlag öffnen