Hello everyone. This is Tanaka.
I would like to introduce a paper I read at the recent English seminar.

Paper Title: English as a foreign language (EFL) secondary school students’ use of artificial intelligence (AI) tools for developing writing skills: unveiling practices and perceptions
Journal: Cogent Education
Year of Publication: 2025
Author: Maha B. Alghasab

This study aims to investigate the actual usage and learner perceptions of AI tools for improving writing skills among 69 high school students in Kuwait who are learning English as a foreign language (EFL).

1. Introduction and Research Objectives
In recent years, the introduction of AI writing tools in English education has progressed, and their impact on learners’ writing abilities has garnered attention. Writing in EFL, in particular, is a complex skill, and AI tools have the potential to serve as scaffolding within Sociocultural Theory (SCT), supporting various stages of writing such as planning, generating, and revising. While AI tools are expected to contribute to the enhancement of vocabulary and grammar and the promotion of Self-Regulated Learning, concerns have also been pointed out, such as plagiarism, a decline in thinking skills due to excessive reliance on AI, and the inhibition of creativity.
Based on the awareness that it is important to grasp the actual usage of learners and build instructional guidelines to incorporate them into classes educationally and ethically, rather than uniformly regulating AI tools, this study investigated the following questions:
1. What are the AI writing tools commonly used by secondary education learners?
2. How are these tools used to improve learners’ writing skills?
3. How do learners perceive the impact of using AI writing tools on their writing skills?

2. Literature Review
This study captures the educational significance of AI tools from the perspective of SCT. Through the SCT concepts of the Zone of Proximal Development (ZPD) and mediation, it explains that AI tools function as “tools” that support learners’ cognitive activities. AI tools support learners’ ZPD and promote autonomous learning through suggestions for grammar and vocabulary corrections, automated feedback, and the like.
In empirical research, it has been reported that Grammarly, ChatGPT, Google Translate, etc., contribute to linguistic accuracy, vocabulary diversity improvement, and the cultivation of self-correction abilities. In particular, automated writing evaluation and automated writing correction feedback tools enable interactive learning through real-time correction suggestions. However, concerns have also been pointed out regarding the inaccuracy of AI feedback, its temporary effects, the suppression of creativity and critical thinking due to excessive reliance on AI, and academic integrity.

3. Research Methodology
A mixed-methods approach using online questionnaires and semi-structured interviews was conducted with 69 EFL learners at a secondary educational institution in Kuwait. The questionnaire measured the level of understanding of AI tools, actual usage, and impact on writing skills, and interviews were conducted to delve deeper into the students’ specific usage situations and perceptions. Data were analyzed using descriptive statistics and thematic analysis.

4. Results
・Actual Usage of AI Tools
The most widely used tools were Google Translate (81%), followed by Grammarly (65%), and ChatGPT (50%). When selecting these tools, students prioritized familiarity and ease of access (accessibility) from multiple devices.
Students were aware of an average of 3.75 types of AI tools and used them complementarily according to their purpose, such as information gathering, outlining, vocabulary selection, and grammar checking, at each stage of writing (pre-writing, writing, and post-writing).

・Perceptions Regarding Impact on Writing Skills
It was suggested that the use of AI tools has a moderate impact on writing skills, and students’ views included both positive aspects and negative concerns. In particular, many students expressed the following concerns:
1. Long-term reliance: The possibility that relying on AI may hinder the development of independent writing ability.
2. Creativity: The possibility that it inhibits independent idea generation because the information suggested by AI is limited.
3. Academic integrity: Feeling anxious about the possibility that the use of AI will be considered cheating.
In addition, some students also showed a tendency to intentionally avoid using AI tools from the perspective of responding to exam environments and maintaining critical thinking skills.

5. Discussion and Conclusion
The results of this study suggest that AI writing tools act as “scaffolding” in SCT, mediating the learners’ writing process and supporting their development as writers. In particular, Google Translate, Grammarly, and ChatGPT contribute to self-correction and structural improvement within the learners’ ZPD.
On the other hand, while many students positively evaluated the immediate support provided by AI, they were aware of concerns regarding long-term writing skill development and issues related to maintaining academic integrity. This suggests the need for educational design that considers the balance with traditional teaching methods while enjoying the benefits of AI. Teachers are required to appropriately integrate AI utilization and traditional teaching methods to provide balanced support that fosters critical thinking and academic integrity.
As limitations of this study, the small number of participants limits the validity of the findings, and reliance on self-reported data and the lack of verification of long-term effects on writing ability are mentioned.

6. Reflections
I chose this study because it is highly relevant to my own research theme and I thought I could gain useful insights as a literature review. It was also helpful in considering how to collect and analyze interview data that I plan to conduct in the future. Understanding the perceptions of EFL learners toward AI tools is important for educational practitioners when designing classes, and this study was useful for considering how to utilize AI in writing learning.
I feel it is noteworthy that while K–12 learners recognize the convenience of AI tools, they have anxieties and concerns about the long-term growth of their writing ability. Although not mentioned in detail in this study, I believe it is necessary to explore more deeply the background of what specific contexts (e.g., idea generation in pre-writing, grammar correction in the revision stage, etc.) they have concerns about in their use.

Also, the following are mentioned as issues for this study:
1. The evaluation of the effectiveness of AI tool usage relies only on subjective data based on self-reports, and it does not directly verify the causal relationship with the improvement of actual writing skills. It is expected that more reliable findings can be obtained by using objective data that incorporates pre- and post-tests.
2. Although it is stated as “utilization of AI tools for writing,” the content of specific tasks is not clearly specified. Since the selection and utilization methods of AI tools vary depending on the context, I believe that more precise analysis will be possible by collecting data on the types of tasks (e.g., English composition assignments outside of class, academic writing in class, etc.) and the implementation situation.
3. Given that the subjects are Year 10–12 students, there are differences in the nature and difficulty of writing assignments imposed for each grade level, as well as the learners’ experience in using AI tools, so analysis according to the grade level of the survey subjects is also an issue.
I would like to proceed with research that develops the findings obtained in this study in the future.

Report by: Sayo Tanaka