Fostering independent expository essay writing development using Song-Assisted Learning (SAL) to avert AI over-reliance

AI over-reliance on expository essay writing aspects lead to decreased cognitive abilities and critical thinking skills, diminished confidence, and waning core skills that are crucial to selfdevelop an expository essay. Its over-reliance leads to like-minded idea production in the students’ work, as reflected in their main ideas, supporting details, and examples. Thus, this study aims to analyse the effects of using a Song-Assisted Learning (SAL) approach to reduce over-reliance on AI for expository essay writing among students in a public university in Peninsular Malaysia. The research design was quantitative in nature, focusing on a quasiexperimental design involving 30 students from a public university located in Perak, Malaysia. Two distinct approaches were used in this study: (i) AI-permissible condition, and (ii) SAL approach. The essays were marked by two independent raters using a reliable rubric adapted from an established university. The data were analysed using SPSS v.29 involving independent samples T-test. For the AI condition test, the findings revealed poorer scores (M = 6.967) resulted from a high plagiarism rate, and points that drifted away from the question’s main requirements. On the other hand, under the SAL condition, the submitted essays contained more human touch – more authentic, personal, creative, and emotion-driven, resulting in better scores (M = 14.233). The difference was significant (p = 0.006), highlighting the negative standing of AI’s over-reliance on students’ expository essay writing capabilities, and the need to use a different approach to foster students’ cognitive and problem-solving skills.