Unlocking the Learning Potential: ChatGPT as a Virtual Platform for Cross-Interaction in English Language Learning ^†

Abstract

We explored how participants learning English as a foreign language (EFL) perceive the use of ChatGPT for English communication in combination with voice control for GPT, a Google Chrome extension, both of which are AI chatbot functionalities. A revised version of the ChatGPT Engagement and Perception Questionnaire (CEPQ) was distributed to 70 participants. The results indicated a generally positive response: participants recognized ChatGPT’s beneficial effect on their confidence in English communication, their indirect interaction experiences with people from different cultural backgrounds, and their familiarity with the system. Overall, 66% of the participants had prior experience with ChatGPT, 65% noted an increase in curiosity, and 77% reported improved confidence in communicating in English. However, 33% experienced frustration due to ChatGPT’s occasional misunderstandings, pointing to the need for enhanced natural language processing capabilities. While 31% expected easy communication with an AI chatbot, the same proportion struggled with pronunciation comprehension. The installation of the voice control extension was straightforward for 57% of participants. A notable challenge was time management, with 47% finding the time allotted for practice tasks insufficient, underscoring the importance of addressing this to improve student engagement and learning outcomes. These results provide a reference on user experiences with AI chatbots in the context of English language learning.

1. Introduction

This study aims to explore the integration of artificial intelligence (AI) in English as a foreign language (EFL) teaching. By incorporating AI tools into language learning, we explored how to enhance communication skills in EFL. One such AI tool is ChatGPT, an AI chatbot designed to simulate human conversations and provide language practice opportunities. We investigated the perceptions of students of EFL regarding the use of ChatGPT with the voice control extension, a Google Chrome plugin, for English conversation purposes. The findings of this research contribute to the understanding of EFL students’ perceptions of AI chatbot communication in English. By understanding the positive impacts, challenges, and usability aspects of ChatGPT, educators and developers can optimize the integration of AI technologies in language learning environments. Then, it is possible to enhance language proficiency and foster practical communication skills in EFL through innovative AI-driven tools such as ChatGPT.

Integrating AI into education demands a user-centered approach focusing on comprehending students’ attitudes and customizing the functions accordingly [1]. Hence, when integrating AI into education, it becomes vital to consider students’ perceptions to ensure user-oriented implementation and optimize its functionalities. This study provides pedagogical insights on leveraging ChatGPT, emphasizing critical considerations for users to harness its full potential. As the utilization of ChatGPT gains recognition as a promising method to enhance English acquisition, it becomes imperative to explore the sentiments and viewpoints of EFL learners regarding ChatGPT, thereby assisting researchers in gaining deeper insights into its efficacy in language learning given the scarcity of studies addressing the perceptions of Taiwanese university students regarding ChatGPT. We conducted a questionnaire survey to collect resource data to explore the following:

• The participants’ perceptions regarding the use of ChatGPT, an AI chatbot, for English communication purposes in EFL;
• Frustrations or challenges that participants encounter when using ChatGPT, particularly regarding understanding and pronunciation comprehension.

2. Literature Review

We adopted two key educational theories: constructivist theory (CT) and autonomous learning theory (ALT). These theories provide a robust theoretical foundation for application. We fostered constructive and autonomous processes, enabling students to navigate their learning journey with the aid of ChatGPT.

2.1. CT

CT was proposed by Piaget and later expanded by Vygotsky, who posited that knowledge is constructed through active engagement with the world, rather than passive reception [2,3]. This theory was epitomized in this study as students interacted with ChatGPT, creating their understanding through problem-solving activities. Through problem-solving activities and dynamic interaction with ChatGPT, students constructed their understanding and embodied the essence of constructivist theory in their learning process. In this study, ChatGPT served as a facilitator of this active learning approach. By engaging with ChatGPT, students were encouraged to explore, question, and reflect—fundamental aspects of the constructivist learning paradigm. Through this dynamic interaction, students acquired specific knowledge and developed authentic language usage, improved their problem-solving skills, and engaged in conversations. The dialogs with ChatGPT encouraged students to actively construct meaning, fostering an enriched, personalized learning journey.

2.2. ALT

ALT suggests that when students assume responsibility for their learning, they are likely to develop effective learning strategies and become motivated as independent learners. The concept of autonomy, which originates from political and moral philosophy, is multifaceted and has been extensively discussed in academia [4]. Initially, learning autonomy focused on completely independent learning [5,6]. This approach advocates for learners to take full responsibility for decision making in their learning journey. According to Holec, teachers play a crucial role in closely monitoring and attending to the pedagogical aspects for their students [7]. This perspective highlights the shared roles of teachers and students within the autonomous learning model, with teachers acting as facilitators and students as experts in their learning process. Similarly, Lewis believed that fostering learner autonomy needs guidance and interdependence, as learners cannot develop autonomy alone [8]. In this context, ChatGPT and the teacher act as guides to assist students in their English learning journey.

Furthermore, autonomous learning offers EFL learners several advantages. Lu, Jiang, and Throssell suggested that autonomous learning improves language proficiency, critical thinking skills, lifelong learning habits, and adaptability to diverse learning contexts [9]. Additionally, Mehdiyev explored the advantages of autonomous learning as perceived by EFL students. These advantages include increased motivation, the opportunity for personalized learning experiences, greater flexibility, and enhanced empowerment in making decisions about their learning process [10]. Finally, computer technology facilitates autonomous learning by EFL university students, promoting their active engagement, self-directedness, and accountability in the learning process [11].

2.3. An AI Chatbot in an EFL Context

Technology continues to evolve and transform many aspects of our lives, and education is no exception. As a result, English teaching is also changing owing to technological advancements. One innovative approach that has garnered increasing interest is the use of ChatGPT. This tool is recognized as a promising solution for teaching English with greater efficiency and effectiveness [12]. Incorporating ChatGPT into oral and aural English teaching allows numerous significant advantages. These benefits heighten student engagement, improve confidence in English communication, enhance language skills, and increase student autonomy, creating an authentic and interactive language-learning environment that features intelligent communication.

2.3.1. Engagement

Tatsenko and Donik explored the impact of using ChatGPT in the classroom, specifically in English language learning [12]. Their findings indicated that incorporating ChatGPT into the learning process increased student engagement and equipped them with practical skills for future careers.

2.3.2. Confidence

Fryer and Carpenter found that ChatGPT reduced shyness among language learners [13]. Students expressed greater confidence in their English communication skills after using ChatGPT. Moreover, the chatbot provided a supportive environment where they could practice language skills without fear of judgment, leading to increased self-assurance [14].

2.3.3. Language Skill Enhancement

The thoughtful suggestions of the scholars help harness the full potential of ChatGPT for the expansion of EFL education. By integrating ChatGPT into the classroom, learners can enhance English language skills in fluency, comprehension, and communication skills. Vocabulary acquisition and pronunciation also improve through regular interactions with ChatGPT [14].

2.3.4. Student Autonomy

ChatGPT provides numerous benefits for EFL communication teaching, as it can function as a virtual assistant, simulate authentic conversations, and offer accessible language learning opportunities anytime and anywhere [12].

2.3.5. Authentic and Interactive Environment

Kohnke, Moorhouse, and Zou emphasized the significance of rich input in achieving successful second language acquisition [15]. They highlighted that AI-powered chatbots offer linguistic input, enable daily conversation practice, and create an authentic and interactive language-learning environment. ChatGPT facilitates an authentic and interactive language-learning environment [16].

2.3.6. Intelligent Communication

A notable benefit of using AI-powered chatbots for language teaching and learning is their ability to communicate intelligently, learn from prior interactions, and act as tireless language-learning assistants [15,17,18]. Ultimately, integrating ChatGPT into English language teaching provides benefits such as increased student engagement, enhanced confidence in English communication, improved language skills and learning autonomy, an authentic and interactive learning environment, and intelligent communication capabilities. These findings highlight the potential of ChatGPT as a valuable tool for practical and accessible language education in the digital era.

2.4. Authentic English Conversation

Speaking skills in language learning are difficult to acquire due to the need to integrate multiple language abilities [19]. Daily conversations are the most effective way to practice and enhance oral and aural skills. Students require meaningful conversation practice to enhance their language skills instead of simply reading aloud from textbooks. Meaningful conversation practice involves actively engaging in conversations where students can apply and develop their speaking abilities in authentic and relevant contexts [16]. Hwang, Guo, Hoang, Chang, and Wu emphasized the importance of authenticity in language learning [20]. In this approach, the value of offering learners opportunities to practice English in real-life situations is acknowledged as a way to engage in meaningful dialogues and interact with genuine language resources, which contributes to more effective language acquisition. Authentic material consist of spoken or written language data generated in real communication rather than created explicitly for language teaching purposes [21,22]. Hence, EFL teaching must be authentic, reflecting real-world language usage through sources. It must also be task-based, engaging learners in practical language use through role-plays, problem-solving, discussions, and simulations [21,23]. To improve EFL learners’ English proficiency, it is essential to (1) prioritize speaking practice and provide opportunities for active participation through activities such as discussions, debates, role-plays, and presentations; (2) emphasize authentic communication by encouraging real-life conversations; and (3) utilize technology tools and multimedia resources to enhance oral English skills [24].

3. Research Methodology

To collect data, we distributed a revised version of the ChatGPT Engagement and Perception Questionnaire (CEPQ) to 70 participants. This questionnaire was designed to gather information on the participants’ perspectives on diverse facets of using ChatGPT including the following aspects: the influence of ChatGPT on their confidence in communicating in English, their indirect experiences of interacting with native English speakers or people from different cultural backgrounds, the students’ level of familiarity with the system, the challenges they faced, and the practicality of features such as the voice control extension.

3.1. Context and Participants

Seventy freshman EFL students enrolled in the Everyday English course reported general satisfaction with their speaking experiences using ChatGPT. The course’s curriculum featured a conversation activity with ChatGPT as a key element of the language practice exercises. Throughout the course, students participated in a variety of English learning activities. These included installing the ChatGPT voice control extension for Google Chrome, engaging in casual conversations, participating in guessing games, completing problem-solving exercises, and posing daily questions to ChatGPT. The students participated in these activities either individually or in pairs. Upon completion of these activities, they engaged in a brief oral reflection session to share their experiences and insights from the conversations with ChatGPT.

3.2. Instruments

We utilized a modified version of the 13-item CEPQ, which was adapted from Seraj, Habil, and Hasan [25]. The modifications were made to better align the questionnaire with the goals and context of this study, ensuring its relevance and applicability. The value of Cronbach’s alpha for the questionnaire items was 0.834, indicating a high level of internal consistency and supporting the reliability, validity, and overall robustness of the collected data (

Table 1. Reliability of questionnaire survey.

Cronbach’s Alpha	Lower Bound	Upper Bound	Number of Items
0.834	0.770	0.886	13

).

4. Results and Discussion

We adopted analysis of variance (ANOVA) to analyze the data of male and female participants. The results showed no significant difference between the two groups, indicating that gender did not play a significant role in the observed variables analyzed (

Table 2. Top five items of CEPQ.

	Mean (Standard Deviation)
Items	Male	Female	Total	Rank
6. This lesson helped me get more confident using English for communication.	4.13	4.13	4.13	1
	(1.14)	(0.797)	(0.916)
13. I don’t usually have a chance to talk to foreigners, but I could experience it indirectly through AI chatbots such as ChatGPT.	3.87	3.96	3.93	2
	(0.815)	(0.955)	(0.906)
1. I tried ChatGPT before so I am familiar with the way how to use it.	3.96	3.87	3.9	3
	(0.878)	(1.172)	(1.079)
2. This lesson aroused my interest in learning more about ChatGPT.	3.87	3.83	3.84	4
	(0.920)	(0.842)	(0.862)
3. I didn’t have any problems logging in to ChatGPT and installing the voice control extension.	3.91	3.72	3.79	5
	(0.949)	(1.097)	(1.048)

).

The students’ top five responses to the questionnaire demonstrated that most participants firmly acknowledged the positive impact of using ChatGPT in enhancing their English learning. It was found that the lesson significantly boosted the individuals’ confidence in utilizing English to communicate effectively. The second-ranked response reflected the limited opportunities to engage directly with native English speakers or people from different cultural backgrounds, but the participants indirectly experienced such interactions through AI chatbots such as ChatGPT. The third-ranked response indicated that the participant had prior experience with ChatGPT, resulting in their familiarity with its usage and functionality. The fourth-ranked item in the questionnaire results indicated that the lesson sparked curiosity among the students, stimulating their desire to explore and delve deeper into the intricacies of ChatGPT. The fifth item indicated that students reported a seamless experience logging into ChatGPT and installing the voice control extension without any issues (Table 2).

The questionnaire data showed noteworthy results. First, 66% of participants reported having a prior experience with ChatGPT, showcasing their familiarity with the system’s operations. Additionally, 65% of the participants acknowledged that the recent instructional session piqued their interest and motivated them to learn more about ChatGPT. Furthermore, 77% of the students felt that the session enhanced their confidence in using English for effective communication, suggesting a positive influence on their linguistic skills and overall confidence in English as a tool for interpersonal interaction. Most strikingly, 93% of the participants recognized that they lacked opportunities to converse with native English speakers or people from different cultural backgrounds, yet they were able to gain indirect exposure to such interactions through AI chatbots such as ChatGPT.

The results also indicated challenges faced by the respondents. For example, 33% of them experienced frustration when ChatGPT failed to understand them, emphasizing the need for ongoing improvement in natural language processing technologies for smoother and more accurate user interactions with AI-based language models. Concurrently, 31% of the respondents thought that engaging in a conversation with an AI chatbot such as ChatGPT was simple. However, the same percentage of participants reported difficulties with ChatGPT’s ability to comprehend their pronunciation. Overall, 57% of the participants reported no issues in accessing ChatGPT and installing the voice control extension, indicating that a significant number of users found the platform user-friendly and the integration of the voice control feature straightforward. This suggests that the interface is intuitive and well-supported, aiding in easy installation and use. Such successful technical execution of ChatGPT ensures a seamless and efficient user experience, particularly in the integration of the voice control extension within the ChatGPT system.

Overall, 47% of participants indicated that the time allotted was insufficient to complete all the assigned practice tasks during the class. In contrast, 23% of respondents confirmed that the provided time was adequate to finish all the tasks. This variation in time management underscores the differences in students’ availability and time allocation, which may affect their ability to engage with and complete the tasks. Educators must take these findings into account and mitigate potential time constraints to improve student participation and enhance learning outcomes. Consequently, the survey results presented important aspects such as the participants’ familiarity with ChatGPT, the effectiveness of instructional sessions, the challenges faced, and the importance of managing time effectively. These are important in refining technological and teaching strategies to improve the utility of ChatGPT in educational contexts and catering to the varied needs of EFL learners (

Table 3. Students’ perception.

	Core
Items	1	2	3	4	5
1. I tried ChatGPT before so I am familiar with the way how to use it.	3%	7%	24%	29%	37%
2. This lesson aroused my interest in learning more about ChatGPT.	0%	6%	29%	41%	24%
3. I didn’t have any problems logging in to ChatGPT and installing the voice control extension.	1%	9%	33%	24%	33%
4. It was great to speak English whether my sentence was perfect or not.	1%	9%	34%	33%	23%
5. I think it’s possible to speak comfortably to an AI chatbot such as ChatGPT.	1%	21%	46%	31%	0%
6. This lesson helped me get more confident using English for communication.	1%	3%	19%	36%	41%
7. I was frustrated when ChatGPT didn’t understand me.	9%	14%	44%	19%	14%
8. ChatGPT often didn’t understand my pronunciation.	13%	29%	27%	20%	11%
9. I didn’t have enough time to finish all the practice tasks in this class.	21%	26%	30%	14%	9%
10. I didn’t have enough time to explore all the possibilities the AI chatbot offers.	11%	20%	20%	29%	20%
11. I didn’t feel like I was communicating with people, so I couldn’t keep talking for long.	6%	31%	36%	16%	11%
12. AI often doesn’t understand my pronunciation, so I think it would be better to do text-chatting.	6%	24%	40%	20%	10%
13. I don’t usually have a chance to talk to foreigners, but I could experience it indirectly through AI chatbots such as ChatGPT.	0%	7%	23%	40%	30%

).

5. Conclusions

We examined the experiences of EFL learners using ChatGPT to practice English speaking. The findings enriched the existing findings on how users interact with AI chatbots in English language learning. These insights are valuable in guiding the development and refinement of AI chatbot technologies, ultimately enhancing their role in supporting student communication, engagement, and overall learning success in EFL teaching methodologies.

The experience of the participants was overwhelmingly positive. Thanks to their interaction with ChatGPT, they showed confidence in boosting their English communication after they became familiar with the system. A significant proportion of participants had prior experience with ChatGPT, indicating its prevalence. Meanwhile, several participants experienced frustration due to the chatbot’s lack of understanding, which implies that improvements in natural language processing are essential. Additionally, pronunciation comprehension challenges were faced by several participants, highlighting the need for further attention. Nevertheless, the installation of the voice control extension was easy for a majority of the participants. Time constraints were another factor, with a substantial percentage of participants reporting insufficient time for completing practice tasks. The study’s findings are consistent with prior scholarly research [14], showing that a majority of Taiwanese EFL students recognize the potential of AI tools, such as ChatGPT, in enriching their English learning experience. ChatGPT was recognized as a highly effective tool for their educational progress.

To maximize the benefits of using ChatGPT in education, Kohnke, Moorhouse, and Zou highlighted the need for teachers to acquire digital competencies, which include mastery of technology, alignment with teaching methods, and social consciousness [15]. This necessitates a thorough understanding of ChatGPT’s features and their periodical updates, ChatGPT’s operational mechanics, the design of practical activities, the ability to tackle challenges, and mindfulness of ethical issues. Additionally, Mohamed proposed strategies to foster professional development, curriculum design, and the application of effective teaching methods, all aimed at utilizing ChatGPT to improve EFL teaching quality [26].

Indeed, the effectiveness of ChatGPT in education depends significantly on the digital proficiency of both educators and students. Equipping both parties with essential technological know-how and an understanding of educational principles enables them to exploit the advantages of ChatGPT in enhancing the language teaching and learning experience.