https://ijmri.de/index.php/jmsi
volume 4, issue 5, 2025
110
A COMPARATIVE STUDY OF TRADITIONAL AND AI-ASSISTED APPROACHES
TO TEACHING WRITING IN ENGLISH AS A FOREIGN LANGUAGE
Munisa Mansurovna Bahrombekova
Teacher, Tashkent State University of Oriental Studies
munisabahrombekova1990@gmail.com
Nasretdinova Mukhlisa Nizametdinovna
Senior Teacher, Tashkent State University of Oriental Studies
Lobarkhon Sobitovna Artikova
Teacher, Tashkent State University of Oriental Studies
lobarkhon55@gmail.com
Abstract:
This study investigates the comparative effectiveness of traditional and AI-assisted
approaches to teaching writing in English as a Foreign Language (EFL), conducted at Tashkent
State University of Oriental Studies (TSUOS). The research focuses on multilingual
undergraduate students majoring in philology (Chinese/English, Persian/English), history
(Pashto/English), and economics (Arabic/English). At TSUOS, students pursue in-depth study of
both an Eastern and a Western language, making them ideal participants for evaluating diverse
instructional approaches.
The study adopts a quasi-experimental design involving two groups: one receiving conventional
writing instruction, and another using AI tools such as ChatGPT and Grammarly to support
writing development. Over 12 weeks, both groups engaged in parallel writing tasks, assessed
through standardized writing rubrics and student motivation surveys. The research aims to
measure improvement in writing proficiency and student attitudes toward the mode of instruction.
The results indicate that students in the AI-assisted group demonstrated significantly higher gains
in writing fluency, lexical diversity, and grammatical accuracy. Additionally, survey responses
revealed increased motivation and engagement among these learners, who cited benefits such as
instant feedback, greater writing autonomy, and reduced anxiety. However, concerns were raised
about overreliance on AI and lack of cultural nuance in automated feedback.
This research highlights the potential of AI-assisted instruction as a valuable supplement to
traditional teaching. The study recommends a blended approach that leverages the strengths of
both human-guided instruction and AI-powered learning tools. These findings have practical
implications for curriculum development in multilingual and culturally diverse academic settings
like TSUOS.
Keywords:
AI-assisted learning, writing instruction, academic writing, multilingual learners,
digital literacy, language motivation, technology-enhanced learning, student perception,
autonomous learning
Introduction
Writing in a foreign language is a cognitively demanding skill that integrates multiple linguistic
and cultural competencies. For learners of English as a Foreign Language (EFL), developing
proficiency in academic writing poses significant challenges due to constraints in vocabulary,
grammar, organization, and confidence. Traditional methods of teaching EFL writing often rely
on structured instruction, teacher-led feedback, peer editing, and model-based imitation. These
https://ijmri.de/index.php/jmsi
volume 4, issue 5, 2025
111
approaches, while effective in fostering linguistic awareness and accuracy, are often time-
consuming, resource-intensive, and limited in terms of individualized support (Hyland, 2003).
Recent advancements in Artificial Intelligence (AI) have introduced transformative tools for
writing instruction, enabling learners to receive instant, personalized, and continuous feedback
on their work. Tools such as ChatGPT and Grammarly can analyze grammar, suggest lexical
improvements, and even provide content development assistance. These platforms offer the
potential to reduce teacher workload, enhance learner autonomy, and accelerate writing
development (Liu et al., 2023; Kim et al., 2023). However, the pedagogical implications of such
tools — especially in multilingual and culturally diverse environments — remain underexplored.
This study was conducted at Tashkent State University of Oriental Studies (TSUOS), where
students pursue a unique bilingual academic track. In addition to their core specializations —
philology (Chinese/English, Persian/English), history (Pashto/English), or economics
(Arabic/English) — students receive intensive training in both an Eastern and a Western
language. This diverse linguistic foundation positions TSUOS students as valuable participants
for evaluating the applicability of AI tools in complex multilingual learning environments.
Despite growing enthusiasm for integrating AI into the classroom, key concerns remain: Can AI-
assisted learning replace or supplement human instruction? How does it impact writing
proficiency and learner motivation? Are there risks of overreliance or ethical misuse? This study
seeks to address these questions by comparing the effectiveness of traditional and AI-assisted
writing instruction for EFL learners in the TSUOS context.
The research aims to assess measurable improvements in students' writing abilities and analyze
their motivational responses and attitudes toward each instructional method. The findings are
intended to support evidence-based curriculum decisions and promote a balanced, context-
sensitive integration of technology in language education.
Methods
Participants
The study was conducted at Tashkent State University of Oriental Studies and involved a total of
60 second-year undergraduate students. The students were majoring in philology, history, and
economics, with language specializations as follows: Chinese/English, Persian/English,
Pashto/English, and Arabic/English. At this university, students study both an Eastern language
and a Western language (primarily English), making them well-suited for examining how
multilingual learners respond to different methods of English writing instruction.
The participants were randomly divided into two equal groups of 30 students:
The control group received traditional classroom-based writing instruction.
The experimental group engaged with AI-supported writing tools, including ChatGPT and
Grammarly.
Instructional Design
The research followed a pre-test and post-test comparative design over the course of one
academic semester (12 weeks). Both groups followed the same writing syllabus, which
emphasized academic essay writing, paragraph structure, coherence, grammar usage, and
argumentative development.
The control group was taught through teacher-led lectures, written feedback, in-class writing
exercises, and peer-review activities. Written feedback was typically given within two to three
days.
The experimental group used AI tools throughout their writing process. During the first two
weeks, students were trained on how to use AI ethically and effectively, including how to write
clear prompts and evaluate AI-generated suggestions. ChatGPT was used for brainstorming and
organizing content, while Grammarly supported revision at the grammar and vocabulary level.
Data Collection Instruments
Two main instruments were used to collect data:
1. Writing assessments: Students wrote a 250–300 word essay before and after the instructional
period. These essays were evaluated by two experienced instructors using a standardized rubric
https://ijmri.de/index.php/jmsi
volume 4, issue 5, 2025
112
based on IELTS writing criteria, including task achievement, coherence and cohesion, lexical
resource, and grammatical range and accuracy.
2. Motivation and perception surveys: After instruction, students completed a short survey in
which they rated their level of motivation, engagement, and the perceived usefulness of the
teaching method. The survey used a simple 1-to-5 scale, where 1 meant “strongly disagree” and
5 meant “strongly agree.” In addition, six students from each group participated in interviews to
provide more detailed insights into their experiences and attitudes.
Data Analysis
To evaluate the effectiveness of the two teaching approaches, the study analyzed both the writing
test scores and student feedback. The essays written before and after the 12-week course were
compared to determine each student’s progress. The results from the control and experimental
groups were also compared to see which method led to greater improvement.
To support these findings, interview responses were reviewed to understand students’
experiences with each approach. Themes such as motivation, independence, clarity of feedback,
and ease of use were identified. This combination of writing scores and learner perspectives gave
a well-rounded view of how traditional and AI-assisted methods influenced both the writing
outcomes and the attitudes of EFL students at Tashkent State University of Oriental Studies.
Results
Writing Proficiency Outcomes
The analysis of pre- and post-test writing scores revealed that both the control and experimental
groups made measurable progress over the 12-week instructional period. However, the degree of
improvement differed significantly.
Control Group (Traditional Instruction):
The group’s average writing score improved from 5.5 to 6.0 (on a 9-point IELTS-based rubric),
reflecting modest gains in sentence structure, coherence, and vocabulary control. The largest
improvements were observed in grammatical accuracy and paragraph organization.
Experimental Group (AI-Assisted Instruction):
This group showed a more substantial increase, with average scores rising from 5.5 to 6.5
.
In
addition to improved grammar and cohesion, students in this group demonstrated increased
lexical variety, clearer argumentative structure, and stronger topic development. The independent
samples t-test indicated a statistically significant difference in score gains between the two
groups.
These results suggest that the integration of AI tools, particularly ChatGPT for brainstorming and
Grammarly for linguistic refinement, positively impacted students’ writing performance.
Motivation and Engagement
Survey data revealed that students in the experimental group reported higher motivation levels
throughout the course. Key responses included:
87% agreed or strongly agreed that AI tools increased their confidence in writing.
81% felt more independent and autonomous in handling assignments.
76% reported that immediate feedback helped them learn from mistakes more effectively than
delayed teacher feedback.
In contrast, the control group showed moderate motivation levels. While students appreciated
personalized teacher feedback, they expressed a desire for faster error correction and more
interactive feedback mechanisms.
Qualitative Insights
Interview responses from experimental group participants emphasized that AI tools allowed
them to experiment with language more freely, improve drafts efficiently, and reduce anxiety
about correctness. However, some expressed uncertainty about trusting AI-generated suggestions
and acknowledged that human feedback remained essential for more nuanced issues like tone
and argument logic.
https://ijmri.de/index.php/jmsi
volume 4, issue 5, 2025
113
Control group interviewees appreciated teacher guidance but noted the slow feedback loop as a
limitation. Several students expressed curiosity about AI tools and a willingness to try them in
future writing tasks.
Discussion
The findings of this study suggest that AI-assisted instruction can significantly enhance EFL
writing proficiency and learner motivation when compared to traditional teaching methods. The
students at TSUOS — who are trained in both Eastern and Western languages — benefited
notably from AI integration, especially in terms of writing fluency, lexical richness, and self-
editing skills. These results support prior research (e.g., Liu et al., 2023; Kim et al., 2023)
indicating that AI tools provide timely, context-sensitive feedback that fosters writing
development and learner autonomy.
One of the most important contributions of AI tools in this study was the acceleration of the
revision process. Students using Grammarly and ChatGPT could instantly identify grammatical
errors, receive alternative vocabulary suggestions, and restructure sentences in real time. This
kind of immediate interaction is particularly beneficial in multilingual environments like TSUOS,
where learners may transfer syntactic structures from their Eastern language backgrounds into
English writing. The AI tools thus acted as a form of individualized scaffolding that traditional
instruction, with its longer feedback loops, could not provide as efficiently.
Additionally, the motivational impact of AI-assisted instruction was substantial. Students
reported feeling more confident, in control of their learning, and willing to take linguistic risks
— key indicators of increased engagement. These affective benefits are vital in EFL writing
contexts, where fear of making mistakes often inhibits student performance. However, this boost
in motivation must be balanced with the development of critical digital literacy skills.
Despite its advantages, the AI-assisted approach is not without limitations. Some students
showed signs of overreliance on automated suggestions, raising concerns about surface-level
editing and diminished critical thinking. Furthermore, while AI can correct mechanics and
structure, it lacks the ability to interpret context, tone, or cultural appropriateness in complex
academic writing. In this regard, teacher feedback remains indispensable, especially for higher-
order thinking, content accuracy, and cultural nuance.
Thus, the evidence supports the adoption of a blended approach — one that integrates AI tools
into traditional instruction, rather than replacing it. Such integration allows for technological
affordances to complement pedagogical expertise, offering a more holistic and effective writing
instruction model for EFL learners at institutions like TSUOS.
Conclusion
This comparative study, conducted at Tashkent State University of Oriental Studies (TSUOS),
offers clear evidence that AI-assisted instruction can significantly enhance the writing
proficiency and motivation of EFL learners across multiple disciplines. Students majoring in
philology (Chinese/English, Persian/English), history (Pashto/English), and economics
(Arabic/English) demonstrated measurable gains in their writing skills when supported by tools
such as ChatGPT and Grammarly. The most substantial improvements were seen in grammatical
accuracy, lexical diversity, coherence, and task achievement — key components of effective
academic writing.
In addition to cognitive and linguistic benefits, the use of AI tools had a profound impact on
students’ affective engagement. Learners in the experimental group reported increased
confidence, reduced writing anxiety, and greater autonomy in drafting and revising their work.
These findings are particularly relevant for multilingual students at TSUOS, who face the unique
challenge of navigating writing across multiple linguistic and cultural systems. The immediacy
and personalization provided by AI tools created a more dynamic, responsive learning
environment, which in turn supported higher levels of learner engagement.
However, the study also reveals important limitations. AI tools, while powerful, cannot replace
the nuanced and context-sensitive feedback provided by human instructors. Several students
expressed confusion or skepticism about AI-generated content, particularly when it came to
https://ijmri.de/index.php/jmsi
volume 4, issue 5, 2025
114
rhetorical structure, argument development, or cross-cultural communication. Furthermore, there
is a risk of students becoming overly reliant on automated corrections without fully
understanding their underlying language issues.
Therefore, the study recommends a blended instructional model that integrates AI tools into a
teacher-guided curriculum. This approach maximizes the advantages of both methods: the speed
and accessibility of AI, and the depth and human insight of traditional instruction. For
institutions like TSUOS, where students navigate complex multilingual and multicultural
learning environments, such a model offers a promising path forward. Future research should
explore long-term impacts, potential ethical considerations, and strategies for training both
learners and educators in responsible AI use.
References
1. Liu H., Zhang Y., Wang L. Enhancing academic writing skills and motivation through AI-
assisted language learning // Frontiers in Psychology. – 2023. – Vol. 14. – Article 1260843. –
URL:
https://doi.org/10.3389/fpsyg.2023.1260843
2. Kim Y., Lee H., Park J. Investigating the effectiveness of AI writing assistants for EFL
learners: A classroom-based study
//
Computer Assisted Language Learning
.
– 2023. – Vol. 36,
No. 4. – P. 398–417. – DOI:
https://doi.org/10.1080/09588221.2022.2100561
3. Hyland K. Second Language Writing
.
– Cambridge: Cambridge University Press, 2003. – 256
p.
4. Yan Z. Exploring the impact of ChatGPT on EFL learners' writing performance // Frontiers in
Psychology.
–
2023.
–
Vol.
14.
–
Article
1260843.
–
URL:
https://doi.org/10.3389/fpsyg.2023.1260843
5. Harunasari S. Y. Examining the effectiveness of AI-integrated approach in EFL writing: A
case of ChatGPT // International Journal of Progressive Sciences and Technologies. – 2023. –
Vol. 39, No. 2. – P. 357–368. – URL:
https://www.researchgate.net/publication/372961015
6. Nazarova Y. The use of traditional and non-traditional methods in EFL classroom //
Innovative Technological Methodical Research Journal. – 2024. – Vol. 3, No. 4. – P. 6–12. –
URL:
https://www.researchgate.net/publication/387652186
7. Fitria T. N. The effectiveness of Grammarly as an AI-assisted language learning tool in
enhancing EFL learners' writing performance // Journal of English Teaching. – 2021. – Vol. 7,
No. 2. – P. 123–134.
8. Su Y., Wang X. Integrating ChatGPT in EFL writing courses: Benefits and challenges //
Journal of Language Teaching and Research. – 2023. – Vol. 14, No. 3. – P. 456–467.
9. Munisa Mansurovna Bahrombekova. Linguocultural Insights Into the Ethical Paradigms
Embodied in English Proverbs: A Structural And Pragmatic Analysis. (2024). American Journal
of Language, Literacy and Learning in STEM Education (2993-2769), 2(5), 401-404.
https://www.grnjournal.us/index.php/STEM/article/view/4918
10.
Nasretdinova, M. N., Saydikramova, U. X., Saydikramova, Fuzaylova, N. A., Fayziyeva,
A. K., & Jabborova, Z. T. (2024). Analysis of Stylistic Aspects of the Appearance of Ellipsis in
Speech
.
South
Eastern
European
Journal
of
Public
Health
,
1676–1680.
https://doi.org/10.70135/seejph.vi.2184
11.
Nasretdinova, Mukhlisa Nizometdinovna, and Xamidova Ruxshona. "Methods of
Analysizing New Words by Reading Texts."
Journal of Higher Education and Academic
Advancement
1 (2024): 229-232.
