The purpose of this study is to determine how students might enhance their essay-writing abilities and how they respond to the essay writing GPT. Explanatory sequential research was used in this mixed method study to analyze both the qualitative and quantitative data in distinct steps. In order to conduct the study quantitatively, the researcher used 60 students who were enrolled in an undergraduate English study program at IAIN Curup, Indonesia, that included an essay writing course. With 30 students of experimental and control group. The test is validated by three experts from UIN Raden Fatah Palembang, University of Bengkulu, and UIN Fatmawati Sukarno Bengkulu. The results of this research are that GPT has no significant impact on students' essay writing skill. Experimental and control groups have almost the same score in essay writing tests. GPT has made students simpler to build the idea because they do not need to think harder like traditionally writing. Instead, the students who are in control group easier to arrange the structure of the essay as what they wanted.
TopIntroduction
Writing holds great significance for university students, particularly those who are EFL (English as a Foreign Language) learners, since it is considered one of the essential skills for effective communication (Jabali, 2018; Toba et al., 2019). Al Khazraji (2019) asserts that in an academic atmosphere, the main objective is to develop proficient writing skills. Lecturers should foster the development of students' writing skills by promoting the production of well-structured papers (Ceylan, 2019). Moreover, the development of organisational skills, together with proficiency in behaviour management, composition revision, and reader awareness, has become essential (Bakry & Alsamadani, 2015).
Although writing programmes are highly beneficial for college students, mastering the art of writing has become challenging for them. This phenomenon has occurred in many situations worldwide, where pupils have made a range of writing errors in the subsequent areas: The citation “Toba et al., 2019” refers to a publication by Toba and colleagues in the year 2019. The four categories include: (1) structural, (2) grammatical, (3) mechanical, and (4) vocabulary. Previous research has indicated a correlation between learning styles and writing difficulties (Bakry & Alsamadani, 2015). Consequently, they were incapable of articulating their perspectives in organised and cohesive paragraphs or essays.
Okpe & Onjewu (2017) assert that acquiring the skill of essay writing can enhance everyday communication, facilitate academic achievement, and foster professional growth. Consequently, the essay writing course becomes increasingly important among university students. The Indonesian higher education (HE) curriculum emphasises the importance of cultivating proficient essay writing skills among university students, particularly those enrolled in the English Education Department.
The subject of AI language processing has undergone a revolution due to the introduction of large, pre-trained language models that greatly enhance the capabilities of AI systems. The following systems, released from 2018 to 2022, are included: the BigScience Large Open-Science Open-Access Multilingual Language Model (BLOOM), Google's BERT, OpenAI's GPT-3, Pathways Language Model (PaLM), and Meta's OPT-175B. This paper incorporates examples that employ the extensively utilised GPT-3.GPT-3 was directed to compose an op-ed, around 500 words in length, elucidating why humans should not harbour any apprehensions towards AI. This instruction was sent along with an initial paragraph to set the overall trajectory. Each of the eight essays generated by GPT-3 was carefully evaluated by a human. The human then chose and arranged the content for the final article and made edits. It was observed that the editing process took less time compared to editing several op-eds authored by humans. The essay titled “We Asked GPT-3 to Write an Academic Paper about Itself—Then We Tried to Get It Published” was published in Scientific American on June 30, 2022 (Thunström, 2022). For this task, GPT-3 was given the specific directive to compose a scholarly thesis on the subject of GPT-3, containing precisely 500 words and incorporating scientific references and citations within the text. Subsequently, it was provided with explicit guidelines for each of the standard sections found in an academic work, namely the introduction, methodology, findings, and discussion. Each component potentially generated a maximum of three distinct versions, and the human co-authors selected which ones to employ. The preprint of the publication, now under evaluation by a journal, credits the two researchers who created the prompts as co-authors (GPT-3, Thunström & Steingrimsson in Kleiman, 2022). The AI-human writing process has several steps, namely parameter setting, AI instruction, evaluation, curation, and outcome editing.