AIQUE framework for AI-generated single best answer questions

Nanyan, Suhaila and Ramly, Nur Fariza (2025) AIQUE framework for AI-generated single best answer questions. In: 2nd International Conference on Artificial Intelligence in Medical Education, 12 - 14 Dec 2025, UKM Cheras.

PDF - Published Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Introduction: The Single Best Answer (SBA) question is one of the widely used Multiple Choice Question (MCQ) formats in medical education. Creating SBA questions is challenging and labour-intensive. Large language models (LLMs) such as ChatGPT have demonstrated the ability to produce a structured question format. Currently, there is no established framework for educators to guide the integration of artificial intelligence (AI) into valid and reliable SBA question design. Despite increasing AI adoption, the absence of guidelines may lead to variable standards. Objective: To propose the AIQUE Framework, a theoretically grounded model for quality-assured AI-generated SBA question development in medical education. Methodology: This conceptual paper is developed based on an integrative conceptual synthesis of four evidence streams: Classical validity theory, national AI governance standards, guidelines on SBA construction and literature on AI-generated examination questions. Concepts were extracted, clustered, aligned, and reorganized into a framework that guides ethical use of AI and production of high-quality AI-assisted SBA questions. Results and discussion: AIQUE Framework comprises five interdependent domains: Alignment, Integrity, Quality, User-moderated Iteration and Ethical Governance. The framework integrates key principles of assessment validity and reliability. Alignment ensures each question matches the blueprint, learning outcomes and required cognitive level. Integrity focuses on factual accuracy and avoidance of AI hallucinations. Quality emphasizes proper SBA structure, functional distractors, clarity and sound psychometric characteristics. User-moderated iteration involves expert review, refinement and content validation process that improves overall validity and reliability. Ethical governance ensures transparency, accountability and responsible AI usage supporting fair and defensible assessment items. Conclusion: AIQUE Framework offers a systematic approach for integrating AI into SBA question development. It outlines best practices, mitigates risks associated with AI-generated questions and provides a theoretical foundation for future research and institutional policy development.

Item Type:	Proceeding Paper (Poster)
Uncontrolled Keywords:	Artificial Intelligence; medical education; question design; single best answer
Subjects:	L Education > LB Theory and practice of education > LB2300 Higher Education T Technology > T Technology (General)
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Medicine Kulliyyah of Medicine > Department of Anaesthesiology & Intensive Care
Depositing User:	Dr Nur Fariza Ramly
Date Deposited:	19 Jan 2026 09:58
Last Modified:	19 Jan 2026 09:58
Queue Number:	2026-01-Q1631
URI:	http://irep.iium.edu.my/id/eprint/126339

Actions (login required)

View Item