Pakistani Undergraduates’ Use of ChatGPT for Writing Assistance: A Pilot Study

Abstract

This pilot study examined the extent to which Pakistani English as a Second Language (ESL) undergraduates use artificial intelligence (AI) tool, ChatGPT, for writing assistance and tested the reliability of a questionnaire developed for this purpose. The study is guided by the Technology-to-Performance Chain (TPC) theory. The pilot was conducted with 25 ESL undergraduates. The questionnaire had 27 items: 22 closed-ended items grouped into four constructs, which measured the extent and frequency of ChatGPT use for writing on a 5-point scale (1 = Never to 5 = Always), and 5 open-ended questions asking how ChatGPT supports students’ vocabulary, sentence structure, paragraph writing, essay writing, and overall academic writing. The closed-ended data were analyzed in SPSS using Cronbach’s alpha, means, standard deviations, medians, interquartile ranges, and percentages. The open-ended responses were read to identify the main patterns. The questionnaire was found to be reliable, with alpha values ranging from .718 to .809 for the four constructs and .929 for the 22 closed-ended items. Mean scores ranged from 2.93 to 3.14, indicating that most participants were at a moderate level of ChatGPT use across all constructs. The open-ended responses showed that participants found AI supportive for specific writing tasks, such as replacing basic words with academic ones, correcting grammar, organizing paragraphs, and making essay outlines. Some also reported greater confidence in academic writing. Since use was moderate, participants may benefit from training on using AI tools more effectively. The findings are preliminary due to the small sample size, and a full analysis will be conducted in the main phase of the study with a larger sample.