Exploring the Potential of ChatGPT in EFL Language Learning: Learners’ Reflections and Practices

1. Introduction

The integration of technology in education has had a profound impact on the landscape of language learning, with a growing emphasis on tools such as ChatGPT. Following its release by OpenAI in November 2022, ChatGPT has garnered significant attention for its potential to enhance language learning. Technology-enhanced learning tools offer learners unique opportunities to engage with the target language in authentic and interactive ways, promoting the development of language skills in a more immersive and meaningful manner (Hassan Taj et al., 2017; Loncar et al., 2021). Additionally, technology has been demonstrated to enhance motivation and autonomy among learners, empowering them to take control of their learning process (Gikas & Grant, 2013; Peterson, 2017).

A significant advancement in this domain is the development of AI-powered chatbots, which utilize natural language processing and machine learning techniques to facilitate personalized, dialogue-based interactions. These chatbots can adapt to learners' proficiency levels, offer real-time feedback, and foster learner autonomy and self-correction (Guo et al., 2022; Chen et al., 2021).ChatGPT, an AI-powered chatbot developed by OpenAI, is among the most notable of these tools. By employing large-scale language models, ChatGPT generates text that is comparable in quality to that of human writers (Bender et al., 2021).

The potential benefits of ChatGPT for language learning are vast, including supporting skill development, providing personalized practice materials, and assisting with writing, research, and problem-solving tasks (Kasneci et al., 2023). Furthermore, ChatGPT provides a platform for learners to hone their communication skills by simulating real-life scenarios, thereby fostering their confidence in utilizing the target language in diverse contexts (Kasneci et al., 2023).

Notwithstanding the aforementioned advantages, concerns have been raised regarding the potential risks of ChatGPT's improper usage, including issues of fairness, copyright infringement, and breaches of academic integrity (Kasneci et al., 2023). Furthermore, there are inquiries regarding the precision and reliability of the information generated by ChatGPT. While ChatGPT can be a valuable tool, learners must develop the capacity to critically evaluate its outputs to avoid misinformation and ensure a high-quality learning experience.

The integration of ChatGPT into blended and online learning environments presents promising opportunities for collaborative learning. By incorporating ChatGPT into group tasks, learners can co-construct knowledge and deepen their understanding of language concepts through guided AI interactions. This approach is consistent with the findings of Hassan Taj et al. (2017) and Loncar et al. (2021), who underscored the significance of interactive tools in enhancing engagement and comprehension. Educators can further capitalize on ChatGPT to develop customized instructional materials that are aligned with their students' needs. For instance, ChatGPT can generate context-specific exercises or simulate conversational partners, catering to varying proficiency levels and learning objectives (Chen et al., 2021). Moreover, by providing immediate feedback on assignments, ChatGPT can enhance the formative assessment process, encouraging students to engage in iterative learning cycles.

As technology continues to evolve, the role of ChatGPT and similar artificial intelligence (AI) tools in promoting lifelong learning becomes increasingly important. These tools empower learners to explore language learning independently, fostering curiosity and adaptability—key traits necessary for success in an increasingly digital world (Gikas & Grant, 2013; Peterson, 2017). However, as educators and learners increasingly adopt these tools, it is imperative to address the challenges associated with their integration and to continuously evaluate their impact on educational outcomes.

Despite the growing significance of AI-powered tools like ChatGPT in language learning, there is still a considerable gap in understanding the effectiveness and impact of these technologies from the learners' perspective. While prior studies have underscored the potential advantages of AI-driven chatbots, further investigation is necessary to ascertain learners' perceptions of and engagement with ChatGPT, as well as to ascertain their distinct requirements and obstacles (Jeon, 2021). Addressing this research gap is imperative to ensure that such tools are optimized to meet the diverse needs of language learners effectively.

The objective of this study is to address this lacuna by examining learners' experiences and attitudes toward using ChatGPT to complete language-learning tasks. Through a meticulous examination of learners' perceptions and interactions with ChatGPT, the research endeavors to furnish invaluable insights into the optimization of this technology's implementation in language education. The study will concentrate on the specific challenges learners encounter and their expectations from such tools, and the findings will contribute to the development of more adaptive, learner-centric language learning solutions.

The integration of technology into language learning has become an essential element in modern educational practices. Technological tools offer learners opportunities to engage with the target language in ways that were previously inaccessible (Hassan Taj et al., 2017). A notable benefit of incorporating technology is its capacity to provide learners with access to authentic and interactive materials, thereby enhancing the learning process (Loncar et al., 2021). For instance, learners can benefit from engaging with authentic resources such as videos, podcasts, and online articles that accurately represent the language as it is used in everyday contexts. This exposure fosters more immersive learning experiences and contributes meaningfully to the development of learners' language skills (Loncar et al., 2021).

Furthermore, the use of technology has been demonstrated to positively influence learners' motivation (Gikas & Grant, 2013; Smith, 2018), as technology offers greater autonomy and flexibility in how learners engage with the material. Peterson (2017) further posits that technology encourages learners to assume greater responsibility for their learning process, thereby fostering a sense of agency that empowers them to direct their own educational journey.

Recent advancements in the fields of artificial intelligence (AI) and natural language processing (NLP) have given rise to intelligent tutoring systems and adaptive platforms for language learning (Heift & Chapelle, 2012). Among these innovations, ChatGPT, created by OpenAI, has emerged as a prominent tool in this field. Utilizing large-scale language models, ChatGPT generates coherent, human-like responses based on user inputs (Bender et al., 2021). Its capabilities extend beyond mere text generation, encompassing the ability to respond to inquiries, assist in writing tasks, and provide customized feedback for learners at diverse proficiency levels (Kasneci et al., 2023). For instance, ChatGPT can assist students in enhancing their writing abilities, creating customized practice exercises, and providing support for tasks related to research and problem-solving (Kasneci et al., 2023).

Notwithstanding its numerous advantages, the utilization of ChatGPT in language learning is not without concerns. A significant concern is its potential to encourage academic dishonesty, including cheating, which could compromise the integrity of assessments (Kohnke et al., 2023). Additionally, the absence of proper citations when ChatGPT paraphrases or draws from external sources could result in unintentional plagiarism. Another challenge is the variability in responses generated by the system. Kohnke et al. (2023) observed that identical prompts often yield different outputs, suggesting that the precision of the prompts plays a crucial role in obtaining consistent results.

While extant literature on ChatGPT and analogous tools proffers valuable theoretical insights into their potential applications in language learning, there is a conspicuous absence of empirical research regarding the experiences and perspectives of learners. To this end, it is imperative to investigate learners' interactions with these systems and understand how they perceive their effectiveness in supporting language learning tasks. Addressing this knowledge gap is paramount to ensure the optimal utilization of these technologies in the future.

2.1. The Potential of AI-Powered Chatbots

In the realm of artificial intelligence (AI), the potential of AI-powered chatbots has garnered significant attention due to their efficacy in facilitating language learning. These AI-driven conversational agents engage users in dynamic and interactive dialogues, offering real-time, turn-by-turn exchanges (Guo et al., 2022). These interactions emulate natural communication environments, thereby providing learners with opportunities to practice and improve their language skills while receiving immediate feedback (Guo et al., 2022).

Research has consistently underscored the potential of chatbots to transform language education. Leveraging natural language processing (NLP) and machine learning algorithms, these chatbots can adapt their interactions to align with the proficiency levels and unique requirements of individual learners. For instance, Shin et al. (2021) developed an adaptive chatbot capable of adjusting conversation complexity based on the learner's skill level. This approach has been shown to enhance user engagement and lead to demonstrable improvements in learning outcomes. Similarly, Yang et al. (2022) demonstrated that a chatbot emphasizing phonetic accuracy can lead to superior outcomes in pronunciation improvement when compared to traditional learning methods. Furthermore, Chen et al. (2022) have extended the application of chatbots by creating a system that provides feedback on writing tasks, including grammar, vocabulary, and sentence structure, which has led to significant improvements in writing proficiency. Kim's (2018) study introduced a vocabulary-focused chatbot, designed to present new words within meaningful contexts and deliver immediate corrections, thereby fostering better retention rates.

ChatGPT signifies a marked progression from traditional chatbots, chiefly due to its capacity to engage users in sophisticated and reciprocal dialogues. Beyond its conversational functionality, ChatGPT offers capabilities such as summarizing complex texts, drafting presentations, and generating creative ideas. The most recent iteration of the platform, ChatGPT-4, has demonstrated an advanced capability to generate academic-grade responses on specialized subjects. However, it is important to acknowledge that the knowledge of ChatGPT is constrained to information available up to 2021. This limitation necessitates that users critically evaluate the outputs of ChatGPT and supplement them with updated resources (Pavlik, 2023). While ChatGPT demonstrates proficiency in the synthesis and reorganization of existing information, attaining specific objectives frequently necessitates meticulous and iterative prompt design by the user.

In sum, AI-powered chatbots, including ChatGPT, offer considerable advantages in language learning by personalizing educational experiences, adapting to individual proficiency levels, and fostering learner autonomy through real-time feedback. However, it is crucial to acknowledge the limitations inherent to these systems, which include the potential for occasional inaccuracies and reliance on outdated data. This underscores the importance of critical engagement with these tools, as well as the necessity for ongoing refinement and improvement of their capabilities.

In light of the dearth of empirical research concerning how learners perceive and utilize ChatGPT in educational settings, this study endeavors to explore this perspective by addressing the following key question:

• What is the perception of ChatGPT as a tool for enhancing English language learning experiences among learners?

2.2. Rationale for the Study

Investigating learners' perspectives is critical for several reasons:

• Firsthand Insights: Understanding students' firsthand experiences provides essential information about how ChatGPT addresses or fails to meet their needs, preferences, and challenges. Such insights are crucial for optimizing ChatGPT's utility in educational contexts.
• Technology Acceptance: Learners' perceptions significantly influence the acceptance and integration of ChatGPT into educational practices. Positive perceptions may facilitate the adoption of this technology, while negative perceptions could hinder its implementation.
• Ethical Considerations: Learners’ views also contribute to ongoing ethical debates surrounding the use of AI in education. Empirical evidence from students' perspectives can enrich these discussions and inform policy-making.

The objective of this study is to explore learners' perceptions of ChatGPT in the context of English language learning. By gaining insights into these perceptions, this research seeks to develop effective applications of ChatGPT that align with learners' values, goals, and specific needs.

3. Methodology

3.1. Design and Context of the Study

This study adopted a small-scale exploratory approach to examine students' perceptions and experiences with ChatGPT in an Iranian educational context.The investigation was conducted at the Varamin-Pishva Branch of Islamic Azad University, a prominent institution in Iran. At the time of data collection, the university had not officially endorsed the integration of ChatGPT in teaching and learning; instead, its use in graded assignments was explicitly prohibited. Notwithstanding these restrictions, the university’s academic environment encouraged the exploration and experimentation with various technological tools, including ChatGPT.

Initial observations by the first author indicated that most students had prior experience with ChatGPT. The present study focused on ChatGPT 3.5, the free version that was available to participants at the time. All subsequent references to ChatGPT henceforth pertain to this version.

3.2. Participants

The present study involved a sample of ten undergraduate students from various academic disciplines at the Azad University of Varamin-Pishva, spanning years 2–4. Participants were selected through a criterion-based, non-random sampling method to ensure a range of perspectives and experiences. The selection criteria included the following:

• Ali: Frequent use of ChatGPT for discussing language learning strategies.
• Zahra: Use of ChatGPT to enhance English language skills for IELTS preparation.
• Mohammad: Observed use of ChatGPT for completing English literature assignments.
• Niloofar: Experience using ChatGPT for research in English translation.
• Farhad: Utilized ChatGPT for preparing legal documents and legal language learning in law courses.
• Mina: Employed ChatGPT for practicing English pronunciation and speaking skills.
• Reza: Used ChatGPT in the context of nursing-related vocabulary and medical terminology.
• Fatemeh: Employed ChatGPT to review and correct grammar in her English writing tasks.
• Amir: Utilized ChatGPT for enhancing understanding of English syntax in linguistics courses.
• Leila: Investigated the feedback mechanisms of ChatGPT to improve her academic writing.

The presence of a diverse array of participants was instrumental in facilitating an in-depth exploration of ChatGPT's applications and the perceived impact on language learning from the perspective of learners at Azad University of Varamin-Pishva.

3.3. Instrument(s)

To investigate students’ perceptions and experiences with ChatGPT in the context of English language learning, this study utilized the following instruments:

3.3.1. Semi-Structured Interview Protocol

The primary instrument for data collection was a semi-structured interview protocol specifically designed for this study. The protocol was developed based on a theoretical framework informed by prior research on the integration of AI tools in education. It consisted of three main sections:

◦

Understanding ChatGPT: Questions aimed to gauge students’ familiarity with ChatGPT and their perception of its potential for language learning.

◦

Applications in Learning Tasks: Items focused on how students used ChatGPT for specific language learning tasks, such as vocabulary building, grammar correction, and speaking practice.

◦

Challenges and Solutions: Questions explored difficulties faced by students in using ChatGPT, along with strategies they adopted to address these challenges.

3.3.2. Audio-Recording Device

With participants’ consent, interviews were audio-recorded to ensure accuracy and facilitate detailed transcription. This allowed the researchers to preserve participants’ verbatim responses for subsequent thematic analysis.

3.3.3. NVivo 11 Software

The collected data was analyzed using NVivo 11, a qualitative analysis tool. This software enabled efficient coding and categorization of themes, ensuring a systematic and detailed analysis of the student’s responses.

3.3.4. Observation Notes

Complementary to the interviews, observational notes were taken during the interactions. These notes documented non-verbal cues, contextual details, and any additional insights that arose during the interviews

3.3.5. Ethical Compliance Documentation

Ethical approval documents and consent forms ensured adherence to ethical standards. Participants were fully informed about the purpose of the study and their rights, and their participation was voluntary.

The combination of these instruments ensured the reliability and depth of the collected data, allowing for a comprehensive exploration of the role of ChatGPT in language learning among students at the Varamin-Pishva Branch of Islamic Azad University (Tajik et al., 2023).

3.4. Data Collection Procedure

This study employed semi-structured interviews to collect data on students' perceptions and experiences with ChatGPT in the context of language learning. The interview protocol, delineated in Appendix A, was developed based on a theoretical framework tailored to the study's context, informed by previous research on AI in education (Kohnke et al., 2023). The interviews were structured around three primary domains:

• Understanding of ChatGPT and Its Potential in Language Learning: This section explored students' familiarity with the tool, their understanding of its capabilities, and how they perceive its role in enhancing language learning.
• Application of ChatGPT in Language Learning Tasks: This section examined how students have incorporated ChatGPT into their study routines, including the specific tasks it helps them complete, such as vocabulary building, writing improvement, and speaking practice.
• Challenges and Solutions in Using ChatGPT for Language Learning: This section identified the challenges students face while using ChatGPT, including limitations in the tool's responses or technical issues, and how they manage or overcome these challenges.

The interviews were conducted in Persian, the native language of the participants, to ensure clarity and depth in responses. With the participant's consent, the interviews were audio-recorded, transcribed verbatim by trained research assistants, and analyzed thematically. Ethical approval for the study was obtained from the university's ethical review board to ensure compliance with ethical standards in research.

3.5. Data Analysis Procedure

The transcribed interviews were analyzed using NVivo 11, following the thematic framework developed by Yan (2023). This framework categorizes insights into four overarching themes:

• The Power of ChatGPT
  Key codes under this theme included:

  ◦

  Generating Ideas: Students described using ChatGPT to brainstorm ideas for essays and presentations.

  ◦

  Providing Individualized Assistance: Participants reported ChatGPT's ability to tailor suggestions based on specific prompts.

  ◦

  Offering Immediate Feedback: Learners highlighted the tool's real-time grammar and vocabulary corrections.

• Potential Challenges

  ◦

  Accuracy Limitations: Students frequently noted occasional factual errors in ChatGPT's responses, acknowledging the need to verify information from other sources.

  ◦

  Plagiarism Concerns: There were concerns regarding the uncredited use of paraphrased content, particularly when students used ChatGPT to generate text for assignments.

• Proper Use of ChatGPT

  ◦

  Effective Usage Strategies: Students emphasized the importance of crafting precise prompts and cross-verifying information with other sources to ensure the accuracy and quality of the output.

• Critical Reflections on ChatGPT (Emerging Inductively)

  ◦

  Overreliance on AI: Some students expressed concerns about overreliance on AI, stressing the importance of balancing AI assistance with independent learning to avoid diminishing their own cognitive skills and creativity.

These themes provide a nuanced understanding of students' experiences with ChatGPT and their reflections on its use in educational contexts.

3.6. Coding and Reliability

The initial coding was conducted by the first author, who meticulously categorized responses into relevant themes and sub-themes. Subsequently, the second author conducted a thorough review of the coded data to ensure consistency and reliability. Any discrepancies that emerged were addressed through collaborative discussion, ensuring a robust and transparent analytical process.

This systematic approach was adopted to elucidate the intricate insights into learners' experiences with ChatGPT, thereby contributing to the broader discourse on AI-assisted language learning.

4. Findings

4.1. ChatGPT as a Peer Tutor for Providing Individualized Assistance

A significant proportion of the student body at Azad University of Varamin-Pishva regarded ChatGPT as a valuable peer tutor, with its accessibility surpassing that of conventional educators. Several participants described how they leveraged ChatGPT for personalized learning across different fields of study. For instance, Zahra, who employed ChatGPT to augment her English language proficiency in preparation for the IELTS exam, shared her experience:

“I inquired about ChatGPT, a large language model, and what advice it could offer regarding the enhancement of the logical flow of my essay. The generated text highlighted instances of sentences that were not well-connected, and it offered suggestions for enhancing the organization of the information. Additionally, I inquired about the inclusion of additional examples to ensure the coherence of my ideas."

This statement illustrates how ChatGPT played a role similar to that of a language tutor, offering direct, useful feedback that helped Zahra improve her essay. Zahra noted that subsequent to utilizing ChatGPT, she found herself visiting the language center with reduced frequency, as she could obtain immediate feedback on her written work, thereby conserving both time and effort. Furthermore, ChatGPT provided effective strategies to enhance her essay, aligning it with the standards for higher IELTS band scores. This case exemplifies the convenience and efficacy of instant feedback in language learning and improvement.

Similarly, Ali mentioned that he frequently used ChatGPT to discuss strategies for language learning. He noted that ChatGPT provided him with customized advice for structuring essays and preparing academic materials, making it a valuable tool in his study routine.

Furthermore, Mohammad underscored the efficacy of ChatGPT in assisting with English literature assignments. He found that the tool offered useful suggestions for improving his arguments and provided alternative ways to express complex ideas in simpler language, thus enhancing his writing.

Farhad, a law student, employed ChatGPT to assist with legal document preparation and legal terminology acquisition. He found the tool particularly beneficial for understanding the precise wording needed in legal contexts, which made his legal studies more manageable. Additionally, Farhad observed that ChatGPT offered a more accessible alternative to consulting with professors, enabling him to continue his work with greater efficiency.

Mina employed ChatGPT for the specific purpose of refining her English pronunciation and enhancing her speaking skills. She found that the AI helped her articulate her thoughts better by providing immediate feedback on her pronunciation and sentence structure, which boosted her confidence in speaking.

Reza, a nursing student, integrated ChatGPT into his study routine to enhance his comprehension of medical terminology and vocabulary pertinent to healthcare. By engaging in discourse with the AI, he was able to practice technical language relevant to his field, thereby reinforcing his knowledge.

Fatemeh utilized ChatGPT to evaluate and enhance the grammar of her English writing tasks. She found the AI's capacity to swiftly identify and suggest improvements to the clarity and structure of her sentences to be particularly beneficial.

Amir, a linguistics student, leveraged ChatGPT to enhance his comprehension of English syntax. Engaging with ChatGPT enabled him to comprehend intricate syntactical structures and apply them to his academic assignments, contributing to his academic success.

Leila, on the other hand, focused on academic writing and explored ChatGPT's feedback mechanisms to enhance the quality of her writing. She examined how the AI's feedback could be customized to enhance the quality of her academic essays, and she found that this approach contributed to achieving more polished and professional academic work.

The findings indicate that ChatGPT possesses the capacity to provide customized, adaptive feedback, which students have found beneficial for various components of their academic pursuits. While it has been demonstrated to support the development of specific language skills, the effectiveness of ChatGPT appears to be contingent upon its utilization. Specifically, the efficacy of this technology hinges upon its application as a complementary tool or as a substitute for active engagement in learning.

Zahra and Ali reported significant improvements in grammar, vocabulary, and essay coherence, particularly in preparation for international exams, with ChatGPT serving as an auxiliary aid. However, students like Mina have observed that while ChatGPT may superficially enhance aspects of language use such as appearance or form, true language proficiency is derived from sustained practice including reading and listening, rather than solely relying on AI. Consequently, while ChatGPT can serve as a beneficial tool, its role in fostering authentic language mastery is contingent upon its integration into the learning process.

In conclusion, the students' feedback revealed that ChatGPT can help improve fundamental language abilities, like vocabulary and grammar, but is less effective for mastering advanced skills, such as writing coherent, logical essays or mastering complex linguistic concepts. The optimal utilization of ChatGPT, therefore, lies in its judicious integration with more conventional language learning methodologies, thereby complementing and enhancing the overall learning experience.

4.2. ChatGPT as a Source for Generating New Ideas

The employment of ChatGPT as a source for the generation of novel concepts was a matter of universal consensus among the ten participants during the planning and composition of an English essay. The following examples illustrate this phenomenon:

Zahra underscored the significance of ChatGPT as a catalyst for ideation and conceptualization, particularly in moments of uncertainty regarding the initiation of an essay. She further elaborated on this point by stating:

“If I were to compose an essay on a particular subject, I would ask ChatGPT for potential approaches or ways to structure my thoughts. It always provides me with different angles to explore."

Ali also mentioned using ChatGPT for the initial stages of planning. He added:

“When I’m unsure where to start with an essay, I request ChatGPT’s suggestions on potential topics or themes. It helps me brainstorm and gives me ideas to consider."

These responses suggest that ChatGPT is particularly beneficial during the planning phase of language writing tasks. However, Mina noted that while the ideas generated by ChatGPT were not always original, they often reminded her of concepts she had not previously considered. She elaborated on this observation, stating:

"The ideas ChatGPT generates are often broad and not fully developed, so I use them as a starting point. From there, I conduct my research to refine and personalize the ideas."

Reza echoed a similar sentiment, observing that while ChatGPT provides useful prompts, he doesn't rely on it to provide fully fleshed-out ideas. He said:

"ChatGPT often gives me general suggestions, but I still need to find more specific examples and refine the ideas before I incorporate them into my writing."

Fatemeh recognized both the strengths and limitations of ChatGPT in the ideation process. She pointed out:

"It’s not always about groundbreaking new ideas, but rather about helping me think of things I hadn’t considered before. I find it useful when I'm stuck."

Amir, who often uses ChatGPT for linguistic-related assignments, noted that the tool excels in offering a variety of perspectives, which aids in broadening his thinking. He explained:

"ChatGPT helps me see different ways of approaching a topic. It’s especially useful when I have a vague idea but need assistance in refining it."

For Niloofar, ChatGPT's role was similar: while it generated useful directions for research, the final conceptualization of the essay required independent work. She shared:

"ChatGPT helps me outline my main ideas, but I always take time to develop them further by reading related texts and finding concrete examples."

Leila also used ChatGPT to help her with academic writing, especially when she needed inspiration for introducing a new topic or argument. She explained:

"It’s like a brainstorming partner. I’ll ask it for suggestions, and then I’ll take those ideas and dive deeper into them to create my argument."

Meanwhile, Mina and Farhad reflected on the role of ChatGPT in generating ideas. Ali argued that while ChatGPT made the process quicker, it wasn't groundbreaking. He remarked:

"Before ChatGPT, I used Google to search for ideas. Now, with ChatGPT, it's just faster and more organized."

Mohammad, in his experience, observed that while ChatGPT was helpful, it provided a similar experience to what he had previously encountered when brainstorming with other tools. He said:

"ChatGPT doesn’t necessarily offer new ideas, but it does provide a more streamlined process for organizing thoughts and narrowing down topics."

In summary, while all participants acknowledged the usefulness of ChatGPT in the ideation process, they recognized its limitations. ChatGPT is lauded for its provision of prompt and diverse suggestions; however, the students underscored that these concepts necessitate additional development and refinement through independent research and critical thinking. The collective responses underscore that while ChatGPT exhibits superior efficiency compared to conventional tools such as Google, it does not supplant the necessity for personal engagement in the generation of ideas and academic work.

4.3. Optimizing Prompts to Enhance Learning Outcomes

The group unanimously agreed that regularly revising prompts was crucial for obtaining accurate and effective responses. Participants likened ChatGPT to a tool requiring clear and structured instructions for optimal use.

Ali, who frequently uses ChatGPT to discuss language learning strategies, noted:

"It's not enough to just ask the software to help with an essay in English. You need to provide detailed instructions—mentioning the academic style, key ideas, expectations, and specific topics. With this clarity, ChatGPT can deliver more tailored and accurate responses."

This finding underscores the necessity of providing ChatGPT with precise instructions to ensure the attainment of optimal results.

Zahra, who uses ChatGPT for enhancing her English skills for IELTS preparation, highlighted that breaking down tasks into smaller parts was effective. She shared:

"When preparing for my IELTS writing tasks, I often ask ChatGPT to evaluate specific sections of my essay—like grammar or cohesion. It performs better with focused and step-by-step instructions."

Mohammad, who applies ChatGPT in English literature assignments, observed:

"By refining my prompts to focus on specific literary analyses, I can receive more relevant insights and examples, which saves time and improves the quality of my work."

Farhad, using ChatGPT for preparing legal documents in his law courses, emphasized the importance of specificity:

"ChatGPT is effective when I provide clear legal scenarios and ask it to draft documents based on defined legal terms. Broad prompts don't work as well in technical subjects like law."

Mina, practicing English pronunciation and speaking skills with ChatGPT, mentioned:

"For speaking exercises, I often request ChatGPT to simulate real-life conversations. Including detailed context, such as specific topics or tone, makes the interactions more practical."

Reza, who focuses on nursing-related vocabulary, stated:

"ChatGPT helps me build a medical terminology glossary. Providing detailed prompts about specific nursing contexts ensures accurate and relevant vocabulary suggestions."

Fatemeh, relying on ChatGPT to review grammar in her writing tasks, noted:

"I receive the best feedback when I submit my writing in small segments. The software is more effective when I specify what areas—like grammar or sentence structure—it should focus on."

Amir, using ChatGPT for understanding English syntax in linguistics, emphasized the importance of iterative feedback:

"When I experiment with different syntax-focused prompts, ChatGPT provides detailed explanations that deepen my understanding of complex structures."

Leila, exploring ChatGPT's feedback mechanisms for academic writing, shared:

"I often test different ways of requesting feedback, like asking for constructive criticism on coherence or argument development. This trial-and-error approach has helped me improve my writing significantly."

In order to optimize prompts, participants employed a trial-and-error approach, shared strategies with their peers, or drew inspiration from Iranian platforms such as Balad. These platforms frequently offer practical examples of effective prompts that users can adapt to suit their specific needs.

This collaborative endeavor highlights the perception of ChatGPT as a technology-enhanced instrument, necessitating active user engagement to optimize its functionality. By refining their prompts, participants optimize the software's potential to support their learning objectives across diverse disciplines, including language learning as well as specialized fields such as law and nursing.

4.4. Formulating Critical Assessments of Information Generated by ChatGPT

While ChatGPT has been demonstrated to be a beneficial tool for the generation of ideas and the provision of assistance, students have exhibited a critical approach to the assessment of its outputs, often reflecting on the reliability of the information it provides. This discerning attitude is exemplified by Farhad, who routinely employed ChatGPT to draft legal documents:

"I once asked ChatGPT to draft a legal memorandum, but the results were overly generic and lacked precision. This made me question its ability to handle the specific nuances of legal language. I ended up revising the draft extensively using authoritative legal sources."

Farhad's response emphasizes the necessity of verifying the outputs of artificial intelligence (AI) by employing reliable references.

In a similar vein, Leila, who concentrated on enhancing academic writing, underscored the significance of critical evaluation.

"ChatGPT's suggestions for improving my arguments were helpful but often lacked depth. To address this, I cross-referenced its feedback with scholarly articles and ensured that the ideas aligned with my research goals."

This approach underscores the manner in which students employ ChatGPT as an ancillary resource rather than as a sole authority.

Reza, who employed ChatGPT for nursing-related vocabulary, also demonstrated caution:

"While ChatGPT provided accurate definitions, I found it necessary to consult medical textbooks to ensure that the terminology was contextually appropriate. This cross-checking process is crucial in a field where precision is key."

In the domain of language acquisition, Mina employed a comparable approach when utilizing ChatGPT for pronunciation training:

"Although ChatGPT helped simulate conversations, I often verified its pronunciation guidance with native speakers or trusted language learning platforms. This ensured that my practice was authentic and accurate."

Amir shared the critical perspective, leveraging ChatGPT to facilitate his comprehension of English syntax:

"ChatGPT provided useful explanations of grammatical structures, but occasionally it failed to address the subtleties of advanced syntax. I found that combining its responses with academic resources led to a more comprehensive understanding."

Ali, a frequent user of ChatGPT to discuss language learning strategies, has identified its limitations in writing that is characterized by subtlety and complexity:

"When I asked ChatGPT to refine a paragraph, its suggestions were generic and lacked cultural or stylistic nuance. This prompted me to seek additional feedback from my peers and mentors."

Fatemeh, who employed ChatGPT for grammar correction, also noted:

"While ChatGPT caught many errors, it sometimes missed subtle stylistic issues. I realized the importance of combining its feedback with my judgment."

Despite its apparent utility, students, including Farhad, Reza, and Leila, have reported that ChatGPT frequently exhibits deficiencies in contextual awareness and encounters challenges when confronted with complex or highly specialized tasks. Consequently, they adopted strategies such as iterative questioning, consulting authoritative sources, and leveraging teacher guidance to validate ChatGPT's outputs.

For example, Zahra, preparing for the IELTS exam, mentioned:

"My instructor encouraged me to critically evaluate ChatGPT's suggestions and focus on improving specific areas, such as coherence and cohesion. This combination of AI and human feedback was particularly effective."

Similarly, Niloofar, who used ChatGPT for translation research, shared that her teacher advised comparing AI-generated translations with human ones to identify contextual gaps:

"This process made me realize that while ChatGPT offers direct translations, it often overlooks cultural and contextual nuances."

A consensus emerged among participants that ChatGPT is a potent instrument, yet its utilization is predicated on users maintaining a critical stance toward its outputs. The efficacy with which users can leverage the software is contingent upon three factors: a comprehensive understanding of ChatGPT's limitations, the exploration of diverse prompts, and the acquisition of expert guidance from educators. By integrating these strategies, students were able to optimize their utilization of ChatGPT, thereby ensuring that its outputs complemented their learning objectives without replacing traditional methods.

4.5. ChatGPT and Plagiarism Issues

About concerns regarding plagiarism that have emerged in the context of the utilization of ChatGPT, all participants exhibited a high degree of caution and awareness concerning this potential risk. Fatemeh articulated this sentiment, stating:

"Some students might think that if they’ve already brainstormed ideas, they don't need to write the essay themselves. I can understand how this approach might clash with traditional educational goals. However, students who are mainly focused on finishing their homework might be inclined to take this route."

Ali shared a similar viewpoint, suggesting that the appropriateness of using ChatGPT might depend on how students perceive the significance of the task:

"It’s understandable that students who are less motivated might be tempted to use ChatGPT as a shortcut since it offers a quick way to complete tasks."

Despite this concern, a consensus emerged among all participants that the implementation of an outright ban on ChatGPT would not constitute a viable solution. Instead, a consensus emerged that instructors should adapt their pedagogical approaches to align with the advancements in technology, guiding the responsible use of these tools.

Farhad, who used ChatGPT for legal documents, said:

"ChatGPT can be a skill, similar to how we use Google for research. I don’t think many would question the legitimacy of getting ideas from Google searches. I believe ChatGPT has the potential to be as ubiquitous as Google in the future. This raises the question of whether we still need to learn how to draft documents or write resumes ourselves."

Farhad's commentary underscored the complexity of determining the appropriate curriculum in an era of pervasive integration of artificial intelligence (AI) tools, such as ChatGPT, into professional and academic domains. A significant number of students have noted that AI can execute certain tasks with greater expediency and efficacy compared to humans, including complex coding tasks. This observation suggests the potential for a future in which AI becomes a significant contributor to technical professions.

Mina, a student focusing on pronunciation practice with ChatGPT, added:

"It's probably not realistic to prohibit students from using ChatGPT. Teachers could instead play an essential role in guiding students on how to use it productively."

In a similar vein, Leila, who had previously investigated feedback mechanisms with ChatGPT, proposed the following:

"Perhaps we could encourage students to report how they use ChatGPT. It might be beneficial to suggest students use it, as long as they are transparent about their process. This could also be a valuable reflective exercise."

These responses lend credence to the notion that a pivotal strategy for combating plagiarism entails the effective and ethical guidance of students in utilizing ChatGPT. Rather than perceiving ChatGPT as a threat, students have advocated for its incorporation into educational practices, ensuring that its use remains responsible and aligned with learning objectives. By adopting these practices, educators and students can adapt to the evolving technological landscape while maintaining academic integrity.

5. Discussion and Conclusions

This study offers one of the first empirical examinations of students' experiences with ChatGPT in a university setting, with a particular focus on its potential impact on language learning. Drawing from the responses of ten students from the Azad University of Varamin, the following three key benefits emerged from the use of ChatGPT:

• Learning Partner or Tutor: ChatGPT was regarded as a valuable instrument for providing customized, accessible, and adaptive feedback. Students like Ali, who frequently utilize ChatGPT as a language learning strategy, underscored its capacity to function as a supplemental tutor, thereby supporting their independent study endeavors.
• Language Proficiency: Students underscored the impact of ChatGPT on the enhancement of their language skills. For instance, Zahra employed ChatGPT to refine her English for IELTS preparation, illustrating how the AI can assist with language proficiency when students proactively engage with it, modify prompts, train the model, and verify its accuracy. Mina, who utilizes ChatGPT to refine her pronunciation and spoken language skills, attested to its efficacy in promoting proper language usage.
• Idea Generation: ChatGPT has also demonstrated its efficacy as a brainstorming tool. Niloofar, who utilizes the AI for research in English translation, noted its efficacy in generating ideas for various academic tasks, thereby demonstrating its role in stimulating creative thought and providing a foundation for more complex projects.

It is noteworthy that the students exhibited a high degree of critical thinking regarding the content generated by ChatGPT. For instance, Fatemeh employed ChatGPT to rectify grammatical errors in her written assignments, while also ensuring that the output was reviewed and refined to align with her personal academic standards. Similarly, Reza's approach entailed a meticulous examination of the AI's responses, particularly in the context of nursing-related terminology, to ensure their accuracy by referencing medical dictionaries.

This ability to modify prompts, train the model, verify outputs, and selectively accept information underscores the pivotal role that critical engagement plays in leveraging AI tools effectively. The students' experiences suggest that ChatGPT can serve as a complementary resource, but it requires active involvement from users to ensure the quality and appropriateness of the generated content.

In conclusion, this study underscores the potential benefits of ChatGPT in supporting language learning and academic development. However, it is equally crucial to underscore the significance of critical thinking, verification, and judicious selection in the utilization of AI outputs. It is incumbent upon educators to adopt AI tools such as ChatGPT while providing guidance to students in their judicious and effective utilization, thereby facilitating the enhancement of their learning outcomes.

5.1. Contributions to Language Learning

This study addresses a significant topic in the realm of technology-enhanced AI tools for language teaching and learning. While opinions vary regarding ChatGPT’s impact on language acquisition (Kohnke et al., 2023; Yan, 2023), our findings suggest it can provide immediate feedback and personalized learning experiences (Chen et al., 2022; Kim, 2018), functioning effectively as a personal language tutor.

Participants found ChatGPT valuable for tasks such as text revision, structural and content suggestions for essays, and providing examples to support arguments. This is particularly relevant in large classrooms, where teacher accessibility may be limited.

Although direct evidence from students’ writings was not collected, participants suggested that ChatGPT's assistance likely improved textual quality (Yan, 2023). For example, Ali and Zahra engaged in reflective essay revisions as they prepared for international assessments. Such reflections have the potential to enhance language skills and foster student autonomy (Kasneci et al., 2023; Kohnke et al., 2023).

5.2. Critical Perspectives and Limitations of ChatGPT

A key finding of this study is students’ development of a critical perspective on ChatGPT’s role in English language learning, a topic not extensively addressed in existing literature. Rather than accepting AI-generated content uncritically, participants evaluated its accuracy, relevance, and specificity. This aligns with findings from Kasneci et al. (2023), who emphasize fostering critical thinking skills, though our results suggest students may naturally develop these abilities through trial and error.

Participants highlighted the potential limitations of AI in learning, emphasizing the importance of assessing the utility of information generated by ChatGPT. These findings contrast with concerns about overreliance on ChatGPT (Kasneci et al., 2023) and suggest that students in this study balanced risks with critical assessments of benefits.

Unlike Yan’s (2023) findings, where students expressed concerns about ChatGPT’s impact on academic integrity, participants in this study displayed a more positive outlook. Most supported universities embraced ChatGPT, with only one dissenting. This openness may stem from participants’ rich experiences with ChatGPT and their capacity for critical evaluation.

5.3. Implications for Teaching

Our findings suggest integrating AI tools like ChatGPT into language teaching may be beneficial. As we transition into an era of artificial intelligence, existing pedagogical tools must adapt to meet evolving needs. It may be more effective to guide students in responsible and productive use of ChatGPT rather than banning it outright.

5.4. Study Limitations and Future Research

The study's methodology was constrained by its reliance on a limited sample shortly after ChatGPT's introduction. It is plausible that students' perceptions might evolve through systematic training or guidance rather than self-directed exploration. To address these limitations, future research should aim to develop comprehensive training programs and involve larger, more diverse samples to provide a more comprehensive understanding of students' experiences and perceptions.