https://jurnal.
id/index.
php/jall/index JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
Received: December 06th, 2025.
Accepted: January 12th, 2026.
Published February 27th, 2026
STUDENTSAo PERCEPTIONS OF THE USE OF CHATGPT IN IMPROVING
ENGLISH WRITING SKILLS
Vanes Kuera*.
Maya Pinkan Warouw.
Ari Anggraini Sebayang Sam Ratulangi University.
Indonesia vanneskuera092@student.
ABSTRACT
The use of artificial intelligence in education has become increasingly important, particularly in supporting English as a Foreign Language (EFL) writing.
Although many previous studies have examined the effectiveness of ChatGPT on writing performance, most have used quantitative approaches, and limited qualitative research has explored studentsAo independent use of ChatGPT in higher education EFL contexts.
This study aims to investigate studentsAo perceptions of using ChatGPT to improve English writing skills, focusing on its benefits, challenges, and user strategies.
This study employed a qualitative descriptive design.
Data were collected through semi-structured interviews with six undergraduate students from the English Department of Sam Ratulangi University who had experience using ChatGPT in academic writing.
The data were analyzed thematically following CreswellAos qualitative data analysis procedures.
The findings show that students perceive ChatGPT as useful for improving grammar and tense accuracy, constructing academic sentences, generating ideas, and supporting self-directed learning.
However, students also reported challenges, including rigid and generic output, increased revision demands, and the risk of psychological dependence that may reduce independent writing ability.
To overcome these challenges, students used strategies such as paraphrasing AI-generated text, refining prompts, selecting essential ideas, and maintaining control over drafting and revision.
In conclusion.
ChatGPT can effectively support EFL writing development when used critically and strategically.
This study recommends that educators guide students in using ChatGPT as a supportive tool rather than a replacement for independent writing to preserve learner autonomy and writing competence.
Keywords: Academic writing.
Artificial Intelligence.
ChatGPT.
EFL writing, studentsAo
INTRODUCTION
English language proficiency plays a crucial role in academic and professional success in todayAos globalized world.
As an international language widely used across fields such as education, science, and technology.
English requires learners to master the four fundamental language skills from an early stage in order to support future competence (Warouw, 2.
Since a large proportion of global knowledge is published in English, students are expected to develop sufficient English proficiency, particularly in writing, to engage effectively with academic and professional information.
Among the four language skills, writing is considered one of the most complex yet essential JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
Writing not only functions as a means of communication but also serves as a medium for expressing critical and analytical thinking, fostering creativity, and supporting academic and professional development (Suastra & Menggo et al.
, 2.
educational contexts, strong writing skills significantly influence studentsAo academic achievement and readiness to compete in the global arena.
Similar to writing, literacy studies also emphasize that reading comprehension is a core academic skill that supports studentsAo ability to process knowledge and develop critical thinking (Sugiarto, 2024.
Ahmad, 2025.
Hidayati, 2.
Despite its importance, mastering English writing skills remains a major challenge for students learning English as a Foreign Language (EFL).
Many EFL learners encounter difficulties related to grammatical accuracy, vocabulary use, coherence, and the organization of ideas.
In addition, some students still rely heavily on rote memorization rather than developing critical and creative thinking skills in the learning process (Feng et , 2.
Research on reading comprehension shows similar problems, where EFL learners struggle to understand academic texts and apply effective strategies (Rinantanti.
These challenges indicate the need for learning tools and instructional strategies that can better support studentsAo writing and literacy development.
Recent developments in information technology have introduced tools designed to support language learning, including artificial intelligence (AI).
One notable AI-based tool is ChatGPT (Generative Pre-trained Transforme.
, developed by OpenAI, which provides real-time feedback, assists in organizing ideas, and supports learners throughout the writing process, from brainstorming to revision.
In EFL contexts.
ChatGPT offers opportunities to address common writing difficulties by providing immediate assistance during academic writing tasks.
As a learning tool.
ChatGPT helps students by correcting grammatical errors, suggesting vocabulary, and providing models of academic sentences and text Through flexible interaction.
ChatGPT supports self-directed learning and encourages students to engage more actively in writing (Maharani et al.
, 2.
Previous studies report that AI technologies, including ChatGPT, can increase engagement in grammar and writing learning .
nzyelik, 2.
and support idea generation and revision JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
(Alamri et al.
, 2.
However, the effectiveness of ChatGPT is strongly influenced by studentsAo StudentsAo engagement is shaped by how useful they believe the tool is and how they experience its role in learning (Polakova, 2.
Some students see ChatGPT as helpful for improving writing, while others see it as a shortcut that may reduce learning effort and weaken independent writing ability (Dewi, 2.
This shows that the value of ChatGPT cannot be understood only through test results.
Although many studies have examined ChatGPT and writing improvement, several gaps remain.
Most previous studies focus on quantitative outcomes, with limited attention to studentsAo real experiences of using ChatGPT independently, especially in higher education EFL contexts.
Few studies also explore how students manage issues such as autonomy, dependence, strategy use, and critical control of AI output.
This study offers a novel contribution by providing qualitative insights into how university EFL students independently use ChatGPT in academic writing, with specific attention to autonomy, dependence, and strategy use.
Unlike previous perception studies that only describe attitudes, this study examines how students negotiate the benefits and risks of AI support in their writing process, showing how technology shapes learning behavior rather than simply improving outcomes.
Understanding studentsAo perceptions is important for effective integration of ChatGPT in EFL education.
Examining how students interact with AI tools reveals their learning strategies and difficulties (Tarchi et al.
, 2.
This study focuses on students from the English Department at Sam Ratulangi University, who have different levels of writing ability.
As future professionals in academic and educational fields, they are required to develop strong writing skills.
Therefore, this study aims to investigate studentsAo perceptions of using ChatGPT to enhance English writing skills, focusing on its benefits, challenges, and the strategies students use to maintain writing autonomy.
The findings are expected to be useful for EFL educators, curriculum designers, and applied linguistics researchers who seek to integrate AI tools responsibly in literacy and writing instruction.
JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
REVIEW OF THE LITERATURE
Recent developments in educational technology have significantly influenced language learning, particularly in the area of English as a Foreign Language (EFL) Writing is a complex skill involving planning, drafting, revising, and editing (Harmer, 2.
For EFL learners, difficulties in grammar, vocabulary, coherence, and idea organization often hinder effective written communication (Feng et al.
, 2.
These challenges have encouraged educators and learners to adopt technological tools to support writing development.
However, writing is also closely connected to broader literacy skills, especially reading comprehension.
Studies in EFL contexts show that studentsAo ability to understand academic texts influences how they organize ideas and develop arguments in writing (Ahmad, 2025.
Hidayati, 2.
In digital environments, learners rely on online reading strategies before transforming information into written form (Rinantanti, 2.
This indicates that challenges in writing cannot be fully separated from challenges in reading, even though the present study focuses mainly on writing.
Artificial intelligence (AI) has recently emerged as one of the most influential technologies in education.
Among AI-based tools.
ChatGPT has gained widespread attention due to its ability to generate text, provide feedback, and assist learners throughout the writing process.
Fitria .
argues that ChatGPT functions as a fast and practical source of written information, making it useful for academic writing tasks.
Similarly.
Song and Song .
found that students who used ChatGPT showed improvement in writing skills and learning motivation compared to those using traditional However, their study was conducted in a structured classroom setting where teachers guided the use of ChatGPT, which may differ from situations where students use the tool independently.
Several studies specifically highlight the positive impact of ChatGPT on writing Anggeraja et al.
reported that students perceived ChatGPT as helpful in improving vocabulary, grammar, and sentence construction.
Roisah et al.
quantitatively demonstrated that students using ChatGPT achieved higher writing scores than those taught using conventional methods.
Many of these studies were conducted in secondary school contexts, where learners usually depend more on teacher guidance.
JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
contrast, university students are expected to be more autonomous, suggesting that findings from secondary-level studies cannot be fully generalized to higher education Moreover, these studies mainly focus on learning outcomes and do not explain how students experience and manage ChatGPT in real writing situations.
Beyond linguistic accuracy, some researchers emphasize ChatGPTAos role in idea generation and writing process support.
Soelistiyowati et al.
found that ChatGPT helped college students generate ideas and improve sentence structure, although concerns about originality and ethics were also noted.
Dewi .
revealed that students preferred ChatGPT because it provided clear and immediate feedback, helping them overcome writing anxiety and build confidence.
These findings relate to themes of idea generation and academic sentence construction that are later discussed in this study.
Nevertheless, they provide limited insight into how students regulate their dependence on AI assistance and how they maintain authorship.
Despite these benefits, other studies raise concerns about the negative effects of technological dependence.
Alhusban .
, although not focused on AI, found that excessive use of technological tools could reduce formal writing quality.
In the context of AI.
Yazid and Dzulfikri .
reported that heavy reliance on ChatGPT led to dependency, reduced comprehension, and uniform writing style.
These concerns relate directly to the themes of dependence and writing autonomy explored in this study.
However, most of these studies describe risks in general terms and do not examine how learners themselves attempt to control or limit their reliance on technology.
From a pedagogical perspective, several studies discuss how ChatGPT should be integrated into instruction.
Daza et al.
demonstrated that AI-based didactic strategies increased student engagement and helped overcome writing difficulties through interactive learning.
Polakova and Ivenz .
showed that ChatGPT feedback improved studentsAo writing quality, but only when students actively revised and reflected on the feedback.
These findings emphasize that strategies and learner control are essential, which directly connects to the theme of student strategies discussed in the findings of this study.
However, these studies mainly focus on teacher-guided use in classrooms, not on independent use outside formal instruction.
JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
Although previous studies confirm that ChatGPT can improve grammar, vocabulary, idea clarity, motivation, and writing confidence (Anggeraja et al.
, 2024.
Song & Song.
Soelistiyowati et al.
, 2.
, several gaps remain.
First, most studies use quantitative or experimental methods that focus on test scores rather than studentsAo lived Second, many studies are conducted in secondary school or teachercontrolled contexts, leaving limited evidence from independent use in higher education EFL settings.
Third, few studies explore how students balance the benefits of ChatGPT with risks such as dependence, increased revision demands, and loss of autonomy.
Therefore, there is a need for qualitative research that explores studentsAo real experiences in using ChatGPT independently for academic writing in higher education EFL contexts.
This study addresses this gap by focusing on studentsAo perceptions, strategies, and reflections, particularly regarding grammar support, idea generation, revision workload, dependence, and writing autonomy.
By explicitly linking reviewed studies to these themes, this research contributes a deeper understanding of how ChatGPT functions not only as a writing assistant, but also as a factor shaping learnersAo writing behavior and independence.
METHODOLOGY
Research Design This study employed a qualitative research approach with a descriptive design to explore studentsAo perceptions of using ChatGPT to improve English writing skills.
qualitative approach was selected because it allows in-depth exploration of participantsAo experiences, meanings, and interactions with technology in real academic contexts (Creswell, 2.
Rather than measuring learning outcomes numerically, this design focuses on understanding how students experience, interpret, and manage the use of ChatGPT in their writing process.
Participants and Data Sources The participants of this study were selected using purposive sampling.
This technique was chosen to ensure that participants met specific criteria relevant to the research focus and were able to provide rich and meaningful information (Nunan, 1.
JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
The participants consisted of six undergraduate students from the English Department of Sam Ratulangi University.
Manado, batch 2022.
To be included in the study, participants were required to:
be active students, .
have completed the Research Method course, and .
have experience using ChatGPT in academic or English writing activities.
The six participants represented three fields of interest: Applied Linguistics.
Linguistics, and Literature.
Although the number of participants was relatively small, the sample size was considered sufficient because data saturation was achieved, meaning that no new themes emerged in the later interviews.
Data Collection Data were collected through semi-structured face-to-face interviews.
This method was chosen because it allows flexibility in exploring participantsAo responses while maintaining focus on the research objectives (Creswell, 2.
The interview guide consisted of open-ended questions covering the following areas:
experiences using ChatGPT in writing, .
perceived benefits, .
perceived challenges, .
strategies used to manage ChatGPT output, and .
views on writing independence.
Before the interviews, participants were informed about the purpose of the study and gave their consent.
Interviews were conducted individually at the Faculty of Humanities.
Sam Ratulangi University.
Each interview lasted approximately 25Ae40 minutes and was audio-recorded to ensure accurate data capture.
Data Analysis The interview data were analyzed using thematic analysis following CreswellAos .
qualitative data analysis procedures.
First, all recordings were transcribed The transcripts were read repeatedly to gain a general understanding of the Next, coding was conducted by identifying meaningful segments of text and labeling them with descriptive codes.
JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
These codes were then grouped into categories based on similarities and patterns.
The categories were further synthesized into broader themes related to the research questions, such as perceived benefits, challenges, and user strategies.
Finally, the themes were interpreted and connected to relevant theories and previous studies to explain their meanings in the context of EFL writing and AI-assisted learning.
To enhance trustworthiness, the researcher carefully rechecked transcripts, coding, and theme formation.
ParticipantsAo direct quotations were used in the findings section to represent their authentic voices and strengthen the credibility of the analysis.
FINDINGS AND DISCUSSION
This section presents the findings of the thematic analysis and discusses their meanings in relation to relevant theories and previous studies.
The findings answer the research questions by describing studentsAo perceptions of using ChatGPT in improving English writing skills, including its benefits, challenges, and the strategies students use to manage AI assistance.
Integrative Overview of Findings Overall, the findings indicate that students experience ChatGPT as both supportive and challenging.
ChatGPT helps with language accuracy, academic expression, and idea generation, but it also creates new demands for revision and raises concerns about These themes are interconnected: support from ChatGPT often increases the need for critical editing, which then encourages students to develop strategies to maintain control over their writing.
This shows that ChatGPT does not automatically improve its impact depends on how students use and manage it.
Linguistic Support and Writing Accuracy Students reported that ChatGPT helped them check grammar, tense, and sentence They commonly compared their original sentences with the revised versions to understand their mistakes.
This finding supports HarmerAos .
view that feedback and revision are essential parts of writing development.
From a sociocultural perspective.
ChatGPT can be seen as temporary support that helps learners in areas they still find difficult (Vygotsky, 1.
Similar results were found by Anggeraja et al.
and JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
Roisah et al.
, who reported improvements in grammatical accuracy when students used ChatGPT.
However, in this study, improvement is based on studentsAo perceptions rather than measured writing scores.
Academic Expression and Idea Generation Students also perceived ChatGPT as useful for constructing academic sentences and generating ideas, especially when they felt stuck.
This supports HarmerAos .
view of writing as a process involving planning and revising.
In line with Dewi .
ChatGPT helped reduce hesitation at the beginning of writing tasks.
However, many students felt that ChatGPTAos language was often too general or rigid, so they needed to revise it carefully.
This suggests that ChatGPT supports idea generation but does not replace studentsAo responsibility to shape their own writing.
Increased Complexity of the Writing Process Although ChatGPT was expected to make writing easier, many students felt that it sometimes made the process longer because the output required heavy editing.
This shows that technology does not always reduce workload.
According to DavisAos .
Technology Acceptance Model, a tool may be seen as less useful if it increases effort.
This finding differs from studies such as Song and Song .
, which highlight efficiency gains, suggesting that efficiency depends on how critically students use AI Dependence and Writing Autonomy Some students worried about becoming too dependent on ChatGPT, especially under time pressure.
This supports Yazid and DzulfikriAos .
finding that heavy reliance on ChatGPT can lead to dependency and uniform writing styles.
From VygotskyAos .
perspective, support should be temporary.
if it becomes constant, it may reduce independent skill development.
However, this study does not claim that all students become dependent.
it only shows that the risk is perceived by some students.
Strategies to Maintain Control To avoid overdependence, students used strategies such as paraphrasing, refining prompts, selecting only important points, and drafting ideas before consulting ChatGPT.
These strategies show that students actively manage AI support rather than using it JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
This aligns with HarmerAos .
emphasis on learner involvement in drafting and revising and with constructivist ideas that learning happens through active engagement (Vygotsky, 1.
Counter-Interpretation: Productivity Tool or Learning Tool ChatGPT can be viewed as a productivity tool that speeds up writing.
However, the findings suggest a more cautious interpretation.
While ChatGPT can help students start writing more quickly, it often requires careful editing to fit academic style and personal Therefore, increased speed is not guaranteed.
This challenges the idea that ChatGPT always makes writing easier and suggests that its value depends on how critically students use it.
Synthesis of Findings In summary.
ChatGPT supports students in grammar, academic expression, and idea generation, but it also introduces challenges related to revision workload and Students respond to these challenges by developing strategies to maintain control over their writing.
These findings suggest that ChatGPT is most helpful when used as a supportive tool rather than as a replacement for studentsAo own writing efforts.
The role of ChatGPT in writing is therefore shaped more by studentsAo strategies and awareness than by the technology itself.
CONCLUSION
This study explored university studentsAo perceptions of using ChatGPT to support English writing in an EFL context through a qualitative descriptive approach.
examining studentsAo experiences, this study aimed to understand the perceived benefits, challenges, and strategies related to the use of ChatGPT in academic writing.
The findings indicate that students generally perceive ChatGPT as helpful for improving grammar and tense accuracy, constructing academic sentences, and generating At the same time, students also experience challenges, including rigid or generic AI output, increased revision demands, and concerns about dependence that may affect writing autonomy.
To manage these challenges, students use strategies such as paraphrasing, refining prompts, selecting essential ideas, and limiting ChatGPTAos role JALL (Journal of Applied Linguistics and Literac.
ISSN 2598-8530.
February.
Vol.
10 No.
1, 2026
during drafting.
These findings suggest that ChatGPT is most useful when used critically and strategically as a supportive tool rather than as a replacement for independent This study has several limitations that should be acknowledged.
First, the study involved a small number of participants from a single university, which limits the generalizability of the findings.
Second, the qualitative design relies on self-reported perceptions rather than direct measurement of writing improvement.
Third, the study focused only on studentsAo perspectives and did not include teachersAo views or classroom Future research is recommended to involve larger and more diverse participant groups, use mixed-method or longitudinal designs, and examine the long-term impact of AI-assisted writing on learner autonomy, creativity, and critical thinking.
Further studies may also explore teachersAo perspectives and instructional strategies for integrating AI tools such as ChatGPT in EFL writing instruction.
REFERENCES