Abstract

This mini review explores high school students’ perceptions of ChatGPT as a feedback tool in English as a Foreign Language (EFL) writing instruction. With the increasing integration of artificial intelligence (AI) in education, understanding students’ experiences with AI-driven feedback is crucial for assessing its pedagogical viability. The study employed a qualitative research design, collecting data through semi-structured interviews with 15 EFL students who received ChatGPT-generated feedback over six writing tasks.

Keywords:Artificial Intelligence; EFL Writing; ChatGPT; Personalized Feedback; AI Feedback; Student Perceptions

Abbreviations: AI: Artificial Intelligence; EFL: English as a Foreign Language

Introduction

The integration of artificial intelligence (AI) into language learning has transformed traditional pedagogical methods, particularly in writing instruction. AI-powered chatbots, such as ChatGPT, offer immediate, automated feedback, potentially improving students’ writing skills. However, the effectiveness of AI-driven feedback is not solely determined by its linguistic accuracy but also by students’ perceptions, engagement, and willingness to incorporate such feedback into their learning processes.

This chapter investigates high school students’ opinions on using ChatGPT as a feedback tool in their English as a Foreign Language (EFL) writing task. The study explores their experiences, challenges, and the perceived advantages of AI-generated feedback over traditional teacher feedback. Understanding how students perceive AI-generated feedback is critical in assessing the viability of such tools for future language education, particularly in non-native English-speaking contexts.

The significance of AI in education has been widely debated, with scholars emphasizing its potential to enhance learning experiences. While some researchers argue that AI-generated feedback can bridge gaps in traditional instruction, others caution against over-reliance on automated tools. A study by Marzuki et al., [1], assessed the impact of AI writing tools on student writing, particularly in terms of components and organization, as perceived by EFL teachers. The findings indicated that integrating AI writing tools can improve the quality of EFL student writing. A study by Mohammed et al., [2], examined the use of ChatGPT by Arab postgraduate students in India for their higher studies. The results suggested that participants found ChatGPT beneficial, impacting their academic writing, language competency, and achievement. Many studies have been conducted on user thoughts and possible fields of using ChatGPT. However, more studies are needed on effectively executing activities that make the most of ChatGPT’s potential in the context of teaching and learning. This study contributes to the ongoing discourse by analyzing the experiences of students who used ChatGPT for writing feedback, providing insights into its practical applications and challenges.

ChatGPT and its use in language learning and teaching

In November 2022, OpenAI introduced ChatGPT, an advanced chatbot powered by deep learning and a large language model. ChatGPT is designed to assist users with creative and technical writing tasks, offering capabilities such as creating, editing, and iterating content. To enhance its performance, ChatGPT utilizes techniques like supervised fine-tuning, reinforcement learning, and conversational prompts [3]. Supervised fine-tuning, a method in natural language processing, involves adapting a pre-trained model to specific tasks or domains. For instance, ChatGPT uses this technique to improve its ability to generate responses for tasks like question answering and fluid conversations. Thanks to its extensive training in diverse textual materials, including books, articles, and web content, ChatGPT possesses a deep understanding of natural language. This enables it to provide contextually relevant answers aligned with user cues.

ChatGPT is a valuable resource for language learners, offering the ability to generate different types of texts, correct errors, and provide definitions and translations. It can also help learners understand the meanings of unfamiliar words in context, explain them, and answer follow-up questions. Additionally, learners can request explanations in their native language. It is a powerful tool for language learning with a vast vocabulary and the capability to produce text that closely resembles human conversation on various topics [4]. Recent advancements in AI have led to the emergence of language learning applications that provide instant corrections on grammar, vocabulary, and sentence structure. AI-driven tools, such as Grammarly and ChatGPT, offer realtime feedback, allowing learners to self-correct and improve their writing. Research by Nguyen [5] and Mohammed et al. [2] suggests that AI-assisted feedback enhances writing skills by offering personalized and timely responses. However, questions remain regarding AI’s ability to replace the nuanced feedback provided by human instructors.

Student perceptions of AI feedback

One of the most promising aspects of AI-driven feedback is its ability to offer individualized learning experiences. Unlike traditional feedback methods, which are often delayed due to workload constraints, AI-powered feedback can be accessed at any time, enabling students to learn at their own pace. Several studies have explored AI’s use to improve students’ writing skills, particularly in personalized feedback and automated writing evaluation systems. These systems, which use AI to provide instant feedback on grammar, punctuation, organization, and style, have enhanced students’ writing performance. In a study conducted by Warschauer and Ware in [6], the effectiveness of writing tools like Criterion and Grammarly on students’ writing skills was examined. The researchers discovered that students who utilized these tools showed significant enhancements in grammatical accuracy and sentence structure over time. This was attributed to the automated feedback provided by the tools, which enabled the students to self-correct and improve their writing. In a study by Roscoe and McNamara [7], the use of AI-based writing tools to support high school students’ argumentative writing was examined. The researchers employed Coh-Metrix, an AI tool, to assess text coherence and cohesion. The findings showed that students who received feedback from the AI tool demonstrated improvement in writing mechanics and higher-order skills, including argument development and using evidence. Furthermore, AI systems can track student progress over time, identifying recurring mistakes and suggesting targeted improvements. Despite these advantages, concerns persist regarding the quality and depth of AI-generated feedback. While AI excels at detecting grammatical errors, it may struggle with more complex aspects of writing, such as coherence, organization, and argumentation

Feedback plays a crucial role in writing development, as it helps learners identify errors and refine their linguistic output. According to Hattie and Timperley [8], effective feedback should be timely, specific, and constructive. Traditional teacher feedback provides expert guidance but is often limited by time constraints. There is a pressing need to explore scalable, personalized solutions to enhance writing skills and feedback processes in language education. Traditional feedback from teachers, though valuable, can be time-consuming, inconsistent, and sometimes inadequate to meet the specific needs of individual learners. AI feedback, on the other hand, offers immediacy but may lack the contextual depth of human input (Algaraady & Mahyoob, 2023).

Students’ perceptions of AI feedback vary depending on their learning styles and preferences. Some students appreciate the efficiency and consistency of AI feedback, while others find it impersonal and less adaptable to their specific needs.

Despite the growing availability of AI tools, there is still a limited amount of empirical research exploring their effectiveness in high school-level EFL contexts, particularly with younger learners, as most existing research focuses on higher education or adult learners. This research is justified by the need to fill this gap and provide data-driven insights on the effectiveness of ChatGPT as a feedback tool for high school EFL students. By focusing on Turkish high school learners, this study expands the cultural and contextual scope of AI research, addressing the unique challenges faced by non-native English speakers in improving their writing skills. The study also aims to explore the psychological benefits of AI-driven feedback. This research will contribute to the ongoing discourse on AI integration in education, offering both theoretical and practical insights into how AI tools like ChatGPT can enhance student learning outcomes, promote learner autonomy, and improve teaching efficiency.

The study aims to see if it is an effective method to integrate ChatGPT into the language teaching and learning process by providing quick and constructive feedback on Turkish high school students’ writing tasks. To achieve this goal, the research is structured around the following question;
1. What are the opinions of the students in the experimental group about studying with the ChatGPT bot for feedback on their writing tasks?

Methodology

To investigate psychological effects of working with ChatGPT, qualitative research design was employed, using semi-structured interviews with students who received ChatGPT-generated feedback on their writing tasks. The study was conducted with 15 high school students from an 11th-grade EFL classroom. The students were part of an experimental group that received feedback exclusively from ChatGPT over six sequential writing tasks. Eight students volunteered for in-depth interviews.

Semi-structured interviews were conducted to gather insights into students’ experiences. The interviews were transcribed and analyzed thematically, focusing on key perceptions related to AI feedback, learning engagement, and future applicability. Thematic analysis was chosen as it allows for an in-depth understanding of students’ perspectives, enabling researchers to identify recurring themes and patterns in their responses. This form was developed after acquiring expert opinions and thoroughly reviewing pertinent literature. The template of the semi-structured interview form is provided in Figure 1.

The forms underwent evaluation by two experienced foreign language educators in qualitative research within the realm of foreign language education, were found appropriate for the study, and ensured the congruence of the interview questions with the research’s scope. Four questions were directed to the 8 participants from the experimental group. Regarding the reliability of the qualitative method, the researchers employed the “member checking” procedure, which entails returning data and interpretations to the participants to verify their responses [9].

Findings

According to the interview results, when the comments made by the participants were analyzed, it was seen that all participants expressed that the role of Artificial Intelligence technology in the language learning process is essential. Most participants stated that this technology can be used to make language practices. The comment made by Participant 3, “It can be used as a teacher. We can learn new language with it. We can improve our skills.” reveals that artificial intelligence technology in language learning holds great promise and shows that students’ awareness is in this direction. Participant 5 supported Participant 3 by saying, “We can use it in language learning. We can translate words. We can learn new subjects. We can make exercises.” Another participant, Participant 6, stated that, “We can communicate with people from all around the world with using AI.” This comment draws attention to the importance of the use of artificial intelligence technology in the field of communication. The view was expressed that artificial intelligence facilitates communication between different languages thanks to its ability to perform instant translation. Participant 8 said, “It saves time. It tells us our errors.” This comment emphasizes the effective use of artificial intelligence technologies, which provide economic savings in different aspects, and the importance of its widespread use in this respect.

Students’ emotional responses when receiving feedback on writing tasks provide critical insights into the pedagogical effectiveness of feedback mechanisms. According to the data collected from the interviews, most students reported positive emotions, with many expressing happiness, excitement, and security. These emotional responses can be understood through the lens of contemporary educational psychology, particularly about feedback as a motivational tool and its role in developing learner autonomy. When asked about their feelings while getting feedback from ChatGPT, Participant 2 stated, “I felt good. Technology is in every area of our lives. It gave me feedback like my teacher. I saw my weaknesses. I was happy because it tells me my writing is good.” The predominance of positive emotions, such as happiness and excitement, in response to feedback suggests that the feedback was perceived as constructive and supportive. Students who felt “happy” highlighted the feedback’s corrective nature, allowing them to identify and rectify their mistakes. Participant 7 said, “I was happy to use technology. I saw my problems and fixed them. It gave me new ideas.” Another comment made by Participant 4 “I felt happy while getting feedback from ChatGPT. I used ChatGPT before for translation. I did not know it can check my homework like my teacher. I was surprised. It showed me the word that I write falsely.” this may indicate that students view it as an opportunity for mastery and growth rather than simply as a critique of their work. Interestingly, several students reported feeling safe and not stressed during the feedback process. This sense of psychological safety is critical for effective learning, as it allows students to engage with feedback without fear of judgment or failure. Participant 3 stated, “I felt happy because I could learn from my mistakes and fix them. I felt safe and not stressed.”

The interviews with students reveal a predominantly positive attitude toward using Artificial Intelligence in their future language learning journeys. The students expressed a range of practical applications for AI, from learning new languages to enhancing vocabulary and speaking skills. These findings align with broader trends in educational technology, where AI is increasingly recognized for its potential to transform language learning by offering personalized, flexible, and accessible tools. Participant 6 said, “I will use it. I love Germany and I want to learn German. I can learn German by myself.” Most students indicated that they plan to incorporate AI into their language-learning processes, particularly when learning new languages. Several participants said that they would use AI tools to make speaking exercises. Participant 2 said, “I will use it to improve my speaking. I will talk with ChatGPT.” Similarly, Participant 5 stated, “I will use technology to improve my other language skills. I can use it to correct my grammar mistakes and speak better.”

Discussion and Conclusion

The findings revealed that students had overwhelmingly positive experiences with ChatGPT, praising its accessibility, speed, and non-judgmental nature. Many students reported feeling more confident in their writing abilities due to the immediate feedback they received. This reflects findings from Fryer et al., [10], who noted that chatbots provided a low-stress environment for learners to practice language skills without fear of embarrassment or judgment.

The students appreciated the efficiency and accuracy of ChatGPT’s feedback on fundamental writing issues, echoing the findings of Roscoe and McNamara [7] that AI feedback tools improve learners’ grammatical accuracy and sentence structure. The students’ intention to use AI in the future suggests that they see it not only as a current tool but as an integral part of their ongoing, lifelong language learning. AI’s potential to adapt to a learner’s evolving proficiency level, track progress over time, and continuously introduce new content. The findings from this study offer several pedagogical implications for integrating AI tools like ChatGPT into EFL writing instruction. First, the study shows that ChatGPT can effectively serve as a supplementary feedback tool in classroom settings. Teachers, particularly those working with large groups of students, often face challenges in providing timely, detailed feedback to every student. ChatGPT offers a solution by providing instant, specific feedback on students’ writing, which can be especially useful in early drafting stages where grammatical and structural issues are most common. This reduces the burden on teachers and allows them to focus more on higher-order concerns such as content development and critical thinking. The AI’s feedback can foster student autonomy. By allowing students to receive real-time feedback without relying on their teacher, AI tools allow learners to take greater responsibility for their learning process. Students in the experimental group appreciated the immediate nature of the feedback, suggesting that AI tools like ChatGPT could promote self-regulation in learning by encouraging students to engage in self-correction and reflection. This could lead to more engaged and independent learners, as they are able to track their progress and make adjustments based on feedback they receive in real-time. However, while AI tools offer many advantages, they cannot replace human teachers nuanced, personalized feedback. The study found that while ChatGPT effectively improved writing mechanics, its ability to address more complex writing issues was limited. Therefore, teachers should view AI tools as complementary rather than replacements for traditional feedback. For example, teachers might use AI to handle basic grammatical corrections while reserving time to provide feedback on content, organization, and creativity. Given the positive reception of ChatGPT by students, it is clear that AI can also play a role in reducing feedback-related anxiety. AI feedback’s non-judgmental, consistent nature allows students to make mistakes and learn from them without fear of criticism, creating a more supportive learning environment.

This research offers significant insights into using ChatGPT to provide feedback in EFL writing instruction. However, it is important to recognize certain limitations. The study’s small cohort of limited number of students from a single high school constrains the broader applicability of its conclusions. Conducted among 11th-grade students in a Turkish high school, the study’s context was quite specific, focusing on students learning English as a foreign language. Consequently, its outcomes might not be entirely relevant across different educational environments, age demographics, or cultural backgrounds. To improve the findings’ applicability in future research, an effort should be made to involve a more extensive and varied group of participants.

Future studies should aim to explore the long-term effects of AI feedback on writing performance. While the current study demonstrated short-term gains, whether these improvements are sustainable over time remains unclear. Longitudinal research, spanning several months or even years, could provide insight into how students’ writing skills evolve with sustained exposure to AI feedback. Future research could focus on the differential impact of AI feedback on various aspects of writing. While this study primarily focused on grammar and sentence structure improvements, future studies could examine how AI tools like ChatGPT influence more complex writing skills such as creativity, critical thinking, and argument development. Another area for further research involves hybrid models of feedback that combine AI and human input. Future studies could investigate how integrating AI-generated and teacher-provided feedback could create a more holistic feedback system.

References

1. Marzuki M, Widiati U, Rusdin D, Darwin R, Indrawati I (2023) The impact of AI writing tools on the content and organization of students’ writing: EFL teachers’ perspective. Cogent Education 10(2).
2. Mohammad A, Al-ghazali A, Khalid A (2023) Exploring ChatGPT Uses in Higher Studies: A Case Study of Arab Postgraduates in India. Journal of English Studies in Arabia Felix 2(2): 9-17.
3. Greyling C (2022) The Large Language Model Landscape. Medium.
4. Shahriar S, Hayawi K (2023) Let’s Have a Chat! A Conversation with ChatGPT: Technology, Applications, and Limitations. Artificial Intelligence and Applications 2(1): 11-20.
5. Nguyen T (2023) EFL Teachers’ Perspectives toward the Use of ChatGPT in Writing Classes: A Case Study at Van Lang University. International Journal of Language Instruction 2(3): 1-47.
6. Warschauer M, Ware P (2006) Automated writing evaluation: defining the classroom research agenda. Language Teaching Research 10(2): 157-180.
7. Roscoe R, McNamara D (2013) Writing pal: Feasibility of an intelligent writing strategy tutor in the high school classroom. Journal of Educational Psychology 105(4): 1010-1025.
8. Hattie J, Timperley H (2007) The power of feedback. Review of Educational Research 77(1): 81-112.
9. Creswell JW, Miller D (2000) Determining Validity in Qualitative Inquiry. Theory Into Practice 39(3): 124-130.
10. Fryer L, Ainley M, Thompson A, Gibson A, Sherlock Z (2017) Stimulating and sustaining interest in a language course: An experimental comparison of Chatbot and Human task partners. Computers in Human Behavior 75: 461-468.