Impact of AI Assisted Feedback on Academic Performance of Undergraduate Engineering Students in Karachi

Abstract

In Pakistani context, excelling writing skills is considered a strenuous task for second language learners. At undergraduate level, students often struggle in their academic writing classes. It has been a challenge for students to outshine the skills of drafting, organizing, and outlining error-free essays. Since the emergence of AI tools like ChatGPT, it has been known to offer extensive support to ESL learners for writing tasks in particular. With the ease of access for AI tools, there is a need to explore its contribution as a feedback tool in formal academic settings apart from content generation. Previous studies have explored the impact of ChatGPT as a feedback tool on students’ learning and its effectiveness in improving writing skills. However, the impact of AI assisted feedback on the performance of students in their final examination remains underexplored. Using a quasi-experimental design, the present study examined the impact of ChatGPT as a feedback tool on academic writing tasks and formal examination performance. Data was collected from 103 undergraduate students studying at a Public Sector Engineering University in Karachi. Pre-and post-tests were used to measure writing improvements with post-test conducted as part of their final semester exam. A questionnaire was used to examine the perceptions of students about ChatGPT’s usability and effectiveness as a feedback tool. Results from the paired sample t-test revealed a statistically significant improvement in post-tests scores. Descriptive analysis of the questionnaire indicated positive perceptions of students regarding usability, feedback clarity, writing confidence and preference for collective feedback (AI + instructor). These findings suggest that ChatGPT, when integrated with instructional guidance, can enhance writing performance even among students with limited prior AI exposure.