Abstract
Abstract Background Artificial intelligence (AI) is increasingly integrated into higher education, including medical education, particularly for research training and academic writing. Research methodology and biostatistics courses are often perceived as challenging, with students reporting difficulty structuring research proposals, interpreting statistical concepts, and applying theoretical knowledge to research practice. This study aimed to assess the perceived benefits, challenges, ethical considerations, and implementation needs associated with a customized AI tool for research proposal writing among fourth-year medical students enrolled in a Research Methodology and Biostatistics course. Methods A convergent mixed-methods design was used. A customized ChatGPT-based Research Proposal Smart Tool (RPST-GPT) was developed and provided to students to support proposal development. Data were collected through a self-administered questionnaire assessing perceived benefits, challenges, and ethical awareness related to AI use in research proposal writing, together with open-ended questions exploring students' recommendations for future AI integration. Results A total of 228 students participated. Students reported high perceived benefits of AI use in research proposal writing (mean 46.48 out of 60) and high ethical awareness regarding appropriate AI use (mean 14.32 out of 20), reflecting recognition of responsible practices and avoidance of overreliance. Qualitative responses indicated that students viewed AI as a supportive learning tool requiring foundational AI literacy and ethical guidance. Conclusions A customized AI tool can enhance undergraduate medical students' research skills when embedded within a structured and governed curriculum. Intentional integration, supported by training, clear policies, and active faculty involvement, is needed to maximize educational benefits and mitigate potential risks.