Document Type : Short Communication
Authors
1
PhD Student, Faculty of Nursing and Midwifery, Mashhad University of Medical Sciences, Mashhad, Iran
2
Department of Nursing, Torbat Heydariyeh University of Medical Sciences, Torbat Heydariyeh, Iran
10.22038/fmej.2025.90396.1690
Abstract
Background: The rise of generative artificial intelligence (AI) has created unprecedented challenges for maintaining academic integrity in online medical examinations. AI tools such as large language models can produce human-like, contextually relevant responses that evade traditional plagiarism detection, undermining the validity of assessment results. Traditional assessments are increasingly vulnerable to AI-assisted cheating, threatening the reliability of evaluation outcomes.
Method: This short communication reviews recent literature on assessment redesign strategies and synthesizes key principles, including authentic clinical scenarios, localized data interpretation tasks, process-tracing methods, multimodal assessment, and open-book application-focused formats, to mitigate AI-related risks.
Results: Redesigned assessments emphasizing higher-order cognitive skills, context-bound reasoning, and real-time performance can limit AI's effectiveness in generating undetectable responses. Practical implementation guidelines are presented for both high-resource and resource-limited settings, with particular relevance to countries facing infrastructure limitations, such as Iran.
Conclusion: By re-engineering assessment tasks rather than relying solely on surveillance technologies, medical schools can safeguard examination validity, preserve public trust in medical qualifications, and ensure graduates possess the competencies essential for safe clinical practice. Future research should evaluate the reliability, feasibility, and scalability of redesigned assessment models across diverse medical education settings.
Keywords
Main Subjects