Treffer: Role-based Prompting Technique in Generative AI-Assisted Learning: A StudentCentered Quasi-Experimental Study.

The education landscape has known remarkable transformations with the emergence of AI which has shown great potential in optimizing different educational processes provided its solutions are utilized effectively. Hence, it has become a necessity to open eyes to practices promoting worldwide efficient use. To this aim, the present paper investigates role-based prompting, its importance in enhancing the output quality of GPT-4 model in AI-assisted learning, and students’ satisfaction with the overall performance with and without using role-based prompts. To achieve the required results, the present study adopts a quasiexperimental pretest-posttest research design with a student-centered approach. The sample of this study includes (N=43) education bachelor’s students of the Higher School of Teachers – Moulay Ismail University, whose ratings were measured before and after the researchers’ intervention. For data analysis, following the ordinal non-normally distributed nature of data, the study adopts Wilcoxon Signed-Rank as a non-parametric test for paired samples conducted using both SPSS 25 and Python codes executed on Google Colab coding space for robust statistical transparency and evidence. Results revealed a strong statistically significant difference between output quality before and after using role-based prompting technique. Additionally, results demonstrated that role-based prompts optimize the output quality in terms of clarity, depth, professionalism, insightfulness, innovativeness, relevance, and generosity. It was also found that students’ satisfaction with the output quality significantly increases with the use of role-based prompts. Furthermore, the paper at hand sheds light on limitations and recommendations to guide future research projects in the field or in fields that relate to it. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Computer Science & Technology Studies is the property of Al-Kindi Center for Research & Development and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)