University students' cyclical self-assessment process mediated by artificial intelligence in academic writing

Abstract

This study was designed to investigate the cyclical self-assessment process in the Academic Writing class mediated by Artificial Intelligence (AI). This narrative inquiry elicited data from three university students having different writing proficiency levels: high, middle, and low levels at one university in Indonesia. The data was collected through reflective notes and interviews. The data was then analysed using thematic analysis. The study revealed two significant findings. First, the three university students with different writing proficiency levels engaged in different stages of cyclical self-assessment caused by two main factors, namely learning motivation and level of trust in AI. The more motivated university student was to learn, the more likely they were to seek external feedback actively. Additionally, their level of trust in Automated Written Corrective Feedback (AWCF) and Automated Writing Evaluation (AWE) engaged them in an evaluation and revision process that engaged them in a cyclical self-assessment process that would improve their final results. Second, Artificial Intelligence (AI) could facilitate an effective cyclical self-assessment process with various features. The implication of the study is discussed.