Content uploaded by Farhad Ghiasvand
Author content
All content in this area was uploaded by Farhad Ghiasvand on May 15, 2024
Content may be subject to copyright.
Int J Appl Linguist. 2024;1–19. © 2024 John Wiley & Sons Ltd. 1wileyonlinelibrary.com/journal/ijal
Received: 10 January 2024 Revised: 8 April 2024 Accepted: 10 April 2024
DOI: 10.1111/ijal.12561
ORIGINAL ARTICLE
Is ChatGPT an evil or an angel for second language
education and research? A phenomenographic
study of research-active EFL teachers’ perceptions
Ali Derakhshan1Farhad Ghiasvand2
1Department of English Language and
Literature, Faculty of Humanities and Social
Sciences, Golestan University, Gorgan, Iran
2Department of English Language and
Literature, Allameh Tabataba’i University,
Tehran, Iran
Correspondence
Ali Derakhshan, Department of English
Language and Literature, Faculty of
Humanities and Social Sciences, Golestan
University, Gorgan, Iran.
Email: a.derakhshan@gu.ac.ir
Funding information
Golestan University, Grant/Award Number:
1571
Abstract
Artificial intelligence (AI) is influencing different aspects
of human life. An AI-powered technology, which has been
recently released, is ChatGPT. It is a cutting-edge tech-
nology that influences second/foreign language (L2) educa-
tion. Although there is increasing research on the benefits
and misfits of this chatbot in different disciplines, L2 edu-
cation lacks a thorough investigation. To fill this lacuna,
this phenomenographic study examined the perceptions of
research-active English as a Foreign Language (EFL) teach-
ers regarding the potentials and pitfalls of ChatGPT for
L2 learning, teaching, assessment, and research. To this
end, a semistructured interview was held with 30 Iranian
EFL teachers with varying educational backgrounds and AI
integration experiences. The results of content and the-
matic analysis indicated that ChatGPT is a double-edged
sword that can both benefit and hurt these areas of L2
education. The most notable potentials were augmenting
learner autonomy, providing personalized learning, reducing
teachers’ teaching workload, designing assessment rubrics,
and summarizing lengthy papers and theses to save L2
researchers’ time and energy. Concerning pitfalls, it was
reported that ChatGPT might kill creativity and academic
integrity, encourage cheating in online exams, spread fake
and misinformation into the world of research, and cherish
high-tech plagiarism. Some practical suggestions are made
2DERAKHSHAN and GHIASVAND
to empower L2 educators and researchers to survive in the
world of AI.
KEYWORDS
artificial intelligence, ChatGPT, educational technology,
phenomenographic analysis, research-active EFL teacher
ມڗW
۱ިD݁ݱٷިᇼሏߓߵۏٷٴ۬۱؇Ü
ෛٺܹژBࢾࣖᆦሏ7૭؇Vّ؊ٔଜଫᇧሏჲᄔ7A?ٞ܂ ڣٷ؇XAÜ݁ٴྥฃพߓߵ۱ިD݁ݱٷިᇼሏჳᄊ7༠ଜଫ7
[
݁ٷྥདྷނڎW7ݿبۚبᇚሏᆾሏᇆ
ሏ7ݿب7તદڣٷ؇XAÜུྱདྷڣٺ۬ߓߵ2݁ިBDBً؇V?XU༠؇Aᇚሏّ؊ٔଜଫᇧሏჲᄔ7A? 7ா༰ູ۬ොگچگ؇:ڣݞ7ೋڎW7Ü?A݁ިA?7ٞ؇Xَ؇۱ٷ༶؇AÜ۱؇Ü7તદ
ۚبً؇: ?AAނٺ۬۱؇Ü
ෛٺܹژ?A༡؇T7ຶ؇U7ݿب2݁ިBDBً؇V?XUڣ؇ڢڎٞ܂ ݁ޚ؇ܳأ۬ၖႤ݁ܭߓߵAXÜ7તદࢹࣖ࣎ࣖW݁چٴ؇ނڎߓߵ7Üߑߵா?V7તદ༠ఃఋ7તદ
݁ޚ؇ܳأ۬ ࢹࣖ࣎ࣖ7Aނٷ؇ۊฆพ7?A7¹݁أఒగ؇Vڣأ؇T?Aߑ߯X۱ݷBً؇V7َႅၯྱཏA7?A݁ިA? ಶ؇૭چܭ۱؇X݁ލႆ႙ఈ:ۚبᇚሏᆾሏᇆሏߓߵ7Üٞ؇?ܓଜଫÜ2݁ިBD
7ABٞ؇ᇀሏX ູොگچݑBً؇V?XU݁ިA?ߓߵAངڢݠ7A?7?ߓߵ7Ü7તદ݁ٷޙިA݁ݱ؇ۋٴ۬7Üയറ۬ ݿ؇ۊٺ؇Aٞ؇ڣٺً۬؇ᡂᡅ݁أ7ߦߵ7ᇃሏBً؇V7َႅၯྱཏً؇ݿ
ި7ًݑູොݱچউXູݠಸ؇:
ෛٺܹژ?AA7ًޚً۬؇7?༚؇U۱ިD݁ݱٷިᇼሏߓߵ7Aނڎ؇ຝ༇ູොܹچܭ
ොٺި7ᇓሏ X݁ިݪިᇼሏ૰؇V?7?ჳᄊۚبᇚሏᆾሏᇆሏٞ܂ᆍᆙލ
ଜଫ?Xܳٴ۬7ݿبჳᄊ ۱ܾᇧሏّި7ࢾًࣖ۬7તદۋިBW۱؇Ü2݁ިBDBً؇V?XUᅛᆇ܂ܔٷڎX۱ًܾ۬2ኡዛ؇ 2ݿྱصߓߵݿ؇ࢾࣖڢ؇ًܭّި༥۬ߙߵતદಶ؇૭چܭ۱؇7ڣݞ7๊ྸ7ݿٺ
گఃఈTٞ؇?ܓଜଫࢾࣖW7A7ف۬ ٞ؇?ܓଜଫÜ
ங
ෛཡၖႤ۱ݷً؇AࣁࣖA๊ྵ݁أఒగ؇Vޗݠ7ᇖሏڢ؇ܳص۱؇Ü7ABٞ؇ᇀሏ X༠ఃఈݬ۬ா?V݁گ؇:Xى؇ٞ؇Vَ؇݁۬۱؇Üޗި
ᇃሏߓߵ7Üཛྷڣ۬ۏިᇓሏ?A XڢبX7ߖߵ¨Ü
ොگگ؇VBً؇V?XUًި??A݁ިA?݁ލႆ႙ఈ:؇ຝ༇૰؇V ?7?ჳᄊۚبᇚሏᆾሏᇆሏᆘᆙܦ݆7ݿب༠ఃఈڢچبXٞിൖ؇A༰
ູොݱچউA77B ًඒඔًଊଫ?ّگܹص?A7݁ٺ༲؇َ؇:2َఃఈતદA77ڣݞ7๊ྸ?۱ڎ7ޗఃఈ༟؇:ۏأউX َ؇?AݿبA7?A?؇Üູොگچگ؇:݁ٷྥདྷܔٷڎXངڢب7?ᇀ
ሏً؇ّܦٷިܳި¨Ü ً؇A7ߙߵؗچصܔٷڎ?A7તદ݁ޚ؇ܳأ۬ۚٷڎུྱލዊዛ؇?ᅦᆇউߓߵ7Üّި7ஓٷڎݿ؇BÜ ಸ؇VX
ොگگ؇VBً؇V?XUXছষ؇A2݁ڎVً؇?
؇Ü۱ިD݁ݱٷިᇼሏ7A7ف۬ ނڎW7ݿب
۱ިD݁ݱٷިᇼሏڣٷ؇XAÜ2݁ިBཇۚبᇚሏᆾሏᇆሏ݁أBً؇V7َႅၯྱཏ ڣأ؇T?Aߑ߯X۱ݷູොܹچܭࢹࣖ࣎ࣖ7Aނٷ؇ۊฆพ
1INTRODUCTION
Despite dogmatic concerns and doubts, artificial intelligence (AI) technologies have recently made their way into edu-
cation with constant progress (Tlili et al., 2023). AI is the science of developing systems that are able to perform
complicated tasks as what humans do (Gibson et al., 2023). Attempts at the intersection of AI and natural language
processing (NLP) culminated in the designation of various intelligent bots and chatbots, which can comprehend and
produce human language (Wu & Yu, 2023). However, persistent efforts made by OpenAI, as an AI laboratory, resulted
DERAKHSHAN and GHIASVAND 3
in a cutting-edge tool on November 30, 2022, which was called ChatGPT or Chat Generative Pretraining Trans-
former. ChatGPT is an AI-powered chatbot that draws on large human conversation data and can produce human-like
responses in a few seconds (Susnjak, 2022). It is pretrained on written texts taken from books, articles, and websites
(Brown et al., 2020). With its revolutionary advent, ChatGPT gained the admiration of numerous scholars and edu-
cators during its first week of public release (Farrokhnia et al., 2023). They called it “scary good,” “insanely strong,”
“world changer,” and “the real thing” (Rudolph et al., 2023).
Thanks to its extraordinary potential to produce complex interactions and prose texts, ChatGPT attracted over
a million subscribers in a short time (Rudolph et al., 2023). Nevertheless, like other innovations, not everyone was
impressed by the injection of ChatGPT into education (Mhlanga, 2023). While the proponents of this AI-powered
chatbot highlight its positive role in creating an adaptive and personalized education (Zhai, 2022), its opponents
point to concerns related to data reliability, authenticity, security, decontextualization, and poor regulations (Haque
et al., 2022). The debates on the demerits of ChatGPT in education also underscored its negative effect on classroom
assessment, scientific integrity, creativity, higher-order thinking, and ethics (Mhlanga, 2023; Rudolph et al., 2023;
Susnjak, 2022). Regarding second language (L2) education, ChatGPT has been perceived as a robust tool that has
afforded new lines of research on the integration of technology and AI in English language education (Aljanabi et al.,
2023). Bin-Hady et al. (2023) found ChatGPT effective in improving EFL learners’ language competence, scaffolded
learning, and language use. Likewise, Yan (2023) identified the chatbot as useful in improving EFL students’ writing
skills. Another strand of research has focused on the potential of ChatGPT in facilitating the article-writing process,
especially literature review summarization and running systematic reviews (Aydin & Karaarslan, 2022;Zhai,2022).
Although these studies are illuminating, they have mostly drawn on state-of-the-art reviews and extensive lit-
erature reviews on AI and education in general. Such theoretical studies have facilitated the ground for future
empirical investigations (Farrokhnia et al., 2023). Nevertheless, the perceptions of research-active EFL teachers about
the potentials and weaknesses of ChatGPT in L2 teaching, learning, assessment, and research have been widely
overlooked. Much has been written about the merits and demerits of this AI-based chatbot, yet integrating teacher-
researchers’ views is an unaddressed angle. This amalgamation of views from EFL teachers, who are active researchers
in teaching, learning, and assessment, may provide a comprehensive image of ChatGPT in L2 education. Our study
draws on both practical and research perspectives to display whether ChatGPT is “evil or angel.” The strength of the
study lies in its use of a cohort of L2 experts, who have practiced ChatGPT in the classroom and conducted research
on L2 education making them aware of the advantages and disadvantages of ChatGPT for teaching, learning, assess-
ment, and research. The study brings new insights into L2 education by showcasing the contributions of generative AI
to various aspects of teaching, learning, assessment, and research. Specifically, the objective of this qualitativestudy is
to depict a comprehensive image of ChatGPT in the field of applied linguistics, which is a fledgling line of inquiry.
2LITERATURE REVIEW
2.1 Artificial intelligence (AI) and education
The emergence and infiltration of AI into the world of education have provided multiple solutions to today’s aca-
demic challenges (Bearman & Ajjawi, 2023; Derakhshan et al., 2024). AI can innovate education, add a new spirit to
teaching and learning practices, and speed up progression (Buriak et al., 2023; Fathi et al., 2024). It is the use of cutting-
edge technology to replicate and perform human actions and thinking patterns (Xu et al., 2021). After the COVID-19
pandemic, the interest in further integrating new technologies into education has vastly increased (Christopoulos &
Sprangers, 2021). One such serious attempt has been implementing AI in education. AI uses advanced, mass data and
analytics to improve educational practices. AI techniques can develop a personalized learning system and interweave
with the fabrics of learning (Zhai, 2022). While the contribution of AI to academic circles has recently boomed, its
inception dates back to the 14th century (Humble & Mozelius, 2019).
4DERAKHSHAN and GHIASVAND
This surge of interest in AI has obliged educators to become multiliterate in the postdigital era. Now, teachers and
students need to know the mechanisms of AI and its offshoots including the growth of machine learning, cognitive
computing, NLP, deep learning, and neural networks involved in learning processes (Tlili et al., 2023). AI provides light
for effective teachers to live in the future and become “more human” with new literacies in an intelligent era (Manyika
et al., 2017). While several benefits have been listed in the growing literature for “AI in education,” some scholars have
cast doubt on assigning teachers’ roles to computers and machines (Humble & Mozelius, 2019). The best choice is
to develop a teacher profile that is commensurate with these innovations (Wogu et al., 2019). It is asserted that the
injection of AI into educational settings transforms future education by calling for personalized and adaptive learning
experiences, equal access to education for minorities and excluded groups, and developing a dual-teacher model in
which the teacher is less bothered with routines and works on one-to-one interaction (Pedro et al., 2019). One of the
revolutionary manifestations of AI integration into education is ChatGPT, which is a powerful chatbot based on mass
data that can generate human-like responses to given tasks/queries (Susnjak, 2022).
2.2 ChatGPT: The game changer
ChatGPT is an AI-powered chatbot recently launched by the OpenAI lab that carried out innovative research on
machines and computers to benefit human life (Brown et al., 2020). ChatGPT aims to improve the world via strong AI
and machines that can perform intellectual tasks like human beings (Elkins & Chun, 2020). ChatGPT belongs to a new
generation of AI bots, namely GPT-3 which is supported and based on billions of corpus (Cooper, 2021). It is trained
to analyze and produce responses akin to those of human beings based on texts. This language model can take tests,
translate texts, write narratives and poems, produce computer codes, perform calculations, and autocomplete images
(Grossman, 2020). ChatGPT has been trained on mass data through an AI supercomputer called Microsoft Azure. Peo-
ple can ask questions or make requests from the chatbot that responds within seconds. This crazy–good innovation
has grasped the attention of millions of users all around the world of which over a million individuals subscribed only
5 days after its release (Murati, 2022).
ChatGPT does not search the net like search engines (e.g., Google) to find relevant information for responding
to a query. It can work offline and generate answers based on various domains of knowledge (Roose, 2022). Chat-
GPT works in different areas of specialty, yet it refrains from producing or suggesting illegal and immoral activities
(Rudolph et al., 2023). Drawing on “transformer architecture,” the chatbot identifies the relationship among words
and produces outputs (Vaswani et al., 2017). It is a leading revolution in AI that has been found influential in different
fields such as medical sciences, chemistry, higher education, and law. With its arrival, ChatGPT sparked sententious
debates among educators regarding its potentials and perils for education (Farrokhnia et al., 2023). While some
considered it insanely good, others regarded it as a demon and an existential threat (Rudolph et al., 2023). Despite
these concerns, it can be contended that the use of ChatGPT in education and L2 education all depends on how one
applies it.
2.3 The potentials and pitfalls of ChatGPT for (L2) education: What the research
says?
Like other forms of technology, ChatGPT has been regarded as a double-edged sword in education, in general, and
in L2 education, in particular. Bursting research has recently focused on the benefits and misfits of this AI-powered
tool. As for the potential of ChatGPT, research evidence shows that it can be momentous for learning by increasing
students’ motivation (Muñoz et al., 2023), problem-solving skills (Urban et al., 2024), classroom collaboration and
teamwork skills (Rudolph et al., 2023), information access (Cascella et al., 2023), autonomy (Agustini, 2023), classroom
engagement (Almusaed et al., 2023), and remote learning (Sharma & Yadav, 2022). Other benefits of ChatGPT for
DERAKHSHAN and GHIASVAND 5
learning concern its provision of real-time and immediate feedback, self-reflectivity development, and affordability
for all students (Baidoo-Anu & Ansah, 2023;Lo,2023). Considering teaching and pedagogy, research shows that
ChatGPT can reduce teachers’ teaching workload, establish a learner-centered pedagogy integrate technology into
education, make the teaching–learning process adaptive, design lesson plans for novice teachers, and provide useful
teaching materials for the class (Farrokhnia et al., 2023).
Concerning assessment, it has been found that ChatGPT can assist in designing tests and quizzes, assessing stu-
dents’ writing assignments, encouraging self and peer assessment, designing assessment rubrics, offering immediate
assessment feedback, providing a personalized, adaptive, and cheap testing system (Rudolph et al., 2023; Swiecki
et al., 2022;Zhai,2022). Additionally, there have been enumerated different potentials of ChatGPT for research.
Prior research reveals that this revolutionary AI technology can assist researchers in summarizing articles, especially
literature reviews (Aydin & Karaarslan, 2022), fostering academic writing and publishing (Zhai, 2022), generating
ideas (Dowling & Lucey, 2023), writing abstracts (Gao et al., 2022), and find pertinent data in a short time (Aydin &
Karaarslan, 2022). As noted, most of the current studies on ChatGPT belong to general education and hard sciences.
However, some studies have examined the uses of this AI chatbot in English language education. As a case in point,
Bin-Hady et al. (2023) argued that ChatGPT improves students’ language skills by offering them scaffolding and timely
feedback. Dörnyei (2020)andAlietal.(2023) corroborated the use of this AI-based technology in engaging and moti-
vating learners in the classroom. Moreover,in an exploratory study in China, Yan (2023) found that ChatGPT promoted
L2 students’ writing skills in a 1-week training course.
Despite these advantages, several problems and pitfalls have been reported in the increasing body of research in
this domain. As for learning, it has been contended that ChatGPT reduces students’ critical thinking skills (Susnjak,
2022); kills creativity (Buriak et al., 2023); promotes fake information and assignments (Tlili et al., 2023); provides
inaccurate, unreliable, and bias information for learners (Zhai, 2022); and overlooks the role of context, culture, and
emotions in learning (Farrokhnia et al., 2023). Concerning teaching, ChatGPT has been found dangerous for encourag-
ing superficial literacy and knowledge development (Dimitrov, 2023), ruining academic integrity (Cotton et al., 2023),
losing human interactions (Farrokhnia et al., 2023), limiting pedagogy to some aspects while ignoring others (Kasneci
et al., 2023). Another area of threat is assessment, which has been found negatively affected by ChatGPT. Research
reveals that this AI bot hurts security in online exams, enhances cheating among students, offers inaccurateand biased
assessment results in some cases like speaking and listening, and ignores nonverbal features of human communication
(Cotton et al., 2023). Lastly, the world of research and scientific publication has been found at risk thanks to the possi-
ble misuse of ChatGPT.Previous studies demonstrated that ChatGPT encourages students and researchers to commit
plagiarism democratically (Gašević et al., 2023); fabricate data (Welle, 2023); spread fake data and misinformation
in science (Liebrenz et al., 2023); copy and paste information from the chatbot (patchwork writing); and disregard-
ing reliability, validity, and ethical considerations of research (Mhlanga, 2023). It has also been claimed that ChatGPT
interpretations and results, in research, are limited in scope since they are based on pre-existing data, and overreliance
on this source hampers novelty in research (Dwivedi et al., 2023).
Although these scholarly works are momentous for providing a global picture of ChatGPT in education, research is
still in its infancy in the context of second/foreign language (L2) education and research. Most of the current stud-
ies have taken theoretical review approaches to enlist the challenges and opportunities of ChatGPT for different
disciplines, particularly in higher education. Additionally, the existing conceptualizations and interpretations have
been mostly made from a practical angle (teachers’ perspectives). Nevertheless, the potentials and pitfalls of this AI-
powered chatbot in L2 education and research from the perspective of research-active teachers have been ignored,
to date. This is very significant because ChatGPT is a language model, and unveiling its role in language education
seems promising and indispensable. Motivated by this gap, this phenomenographic study examined the perceptions of
EFL teachers, who were actively engaged in researching and practicing L2 education in light of technologies, regard-
ing the merits and demerits of ChatGPT for L2 learning, teaching, assessment, and research. This would complement
the practice-oriented picture with research-driven evidence. The following research questions guided the current
study:
6DERAKHSHAN and GHIASVAND
1. How do research-active EFL teachers perceive ChatGPT and its potentials for L2 learning, teaching, assessment,
and research?
2. What are the pitfalls of ChatGPT for L2 learning, teaching, assessment, and research from the perspective of
research-active EFL teachers?
3METHOD
3.1 Research design
This study used a phenomenographic research design, which appropriately reveals the perceptions of a group of peo-
ple about a specific phenomenon (Hajar, 2021). It portrays different ways through which people conceptualize and
perceive an event. This design takes naturalisticand explorative approaches to comprehend a real-world phenomenon
in situ (Sin, 2010). In contrast to phenomenology which takes the first-order perspective, phenomenography takes
the second-order perspective to describe one’s perceptions, conceptions, and understanding of a phenomenon (Stolz,
2020). Since EFL teachers may have different understandings and perceptions about ChatGPT and AI in education,
this qualitative design was employed to uncover research-active EFL teachers’ perceptions of ChatGPT potentials and
pitfalls. The choice of this design affected the instrument construction, participant recruitment, and data analysis in
that we looked for information-rich cases, which might have various understandings and experiencesof ChatGPT with
the aim of unraveling and describing their lived experiences rather than their simple perceptions.
3.2 Participants
The participants of this study were 30 EFL teachers, who were active in L2 research and practice. The sample included
both genders (males =17, 56 %, females =13, 43%) with their ages spanning from 30 to 51 years (M=39.65, SD =6.10).
They were teaching English in different state universities in Tehran majoring in applied linguistics and English lan-
guage literature. The participants were selected through purposive sampling, a nonprobability sampling technique in
which the most representative and information-rich samples are picked up based on some prespecified criteria (Pat-
ton, 2002). In this technique, certain characteristics, experiences, and willingness to attend research determine the
target sample of the study. In this study, the inclusion/selection criteria were being research-active EFL teachers, hav-
ing experience in implementing innovative technologies in the classroom, being familiar with AI and ChatGPT, and
doing research on technology-based L2 education. The reason for picking such cases up was to glean information from
both practical and research angles. The participants had been using ChatGPT in the four L2 domains for 1 year. They
declared to use ChatGPT to teach conversation, prepare lesson plans, provide test questions, give prompts for writing,
and summarizing research papers. Initially, 42 EFL teachers agreed to participate in the study of which 30 teachers
met the prespecified criteria for inclusion. They were contacted via email and Telegram application. Before commenc-
ing, the researchers fully described the purpose and process of the study. To observe ethical codes, formal consent
was signed and delivered to the researchers and the participants were assured of their freedom to withdraw, as they
desired, as well as their privacy, identity, and confidentiality concerns. Moreover, the researchers made sure that they
had no conflict of interest with the sample.
3.3 Instrument
In line with the research questions and objectives of the study, a semistructured interview was utilized in this study
to glean the data. To prepare the instrument, the researchers initially examined and followed the standard steps
DERAKHSHAN and GHIASVAND 7
and considerations in conducting qualitative interviews proposed by Creswell (2013). Then, they developed three
open-ended questions in compliance with the research objectives. The questions were then delivered to four experts
to ensure their content validity in terms of language suitability, pertinence, and clarity. After some comments and
revisions, the interview questions were approved to be applicable. The interview included two sections. The first
one was related to EFL teachers’ background information (age, gender, major, experience level, and familiarity with
ChatGPT). The second or main part asked the 30 respondents to explain and enumerate the possible benefits and
misfits of ChatGPT for L2 learning, teaching, assessment, and research (Appendix). The questions were developed
by the researchers and their content validity was ensured by a panel of four experts in qualitative research in L2
education. The interview was face-to-face, interactive, and audio-recorded in a quiet room through an iPhone.
3.4 Data collection
This study used a semistructured interview to collect the needed data. To this end, the researchers, initially, examined
and followed the standard steps and considerations in conducting qualitative interviews proposed by Creswell (2013).
Then, they developed three open-ended questions in compliance with the research objectives. The questions were
then delivered to four experts to ensure their content validity in terms of language suitability, pertinence, and clar-
ity. After some comments and revisions, the interview questions were approved to be applicable. Next, one-on-one
interviews were carried out with 30 research-active EFL teachers in a quiet room with no distractions. The interviews
were held in noninstructional times as suggested by the participants, who were mostly busy with teaching different
courses. An interview guide was used throughout the interviews. During the interviews, as suggested by Creswell
(2013), the researchers were respectful and good listeners. They did little talking and stimulated the teachers to
describe their perceptions of ChatGPT in detail. However, the researchers were reflective and aware of the inter-
view’s goals. To prevent bias and influence on data, the researchers strived to gain a “second-order perspective,” which
concerns how the participants characterize their experiences and conceptions of a phenomenon (Stolz, 2020). Each
interview lasted 25 min per participant. The data were then transcribed verbatim by the researchers before running
the final analyses. Then, the transcriptions were checked against audio files recorded by the smartphone. After ensur-
ing the data match, both researchers took part in the data analysis and coding phase of the study. They developed a
codebook in which the respondents’ quotes and extracted themes were pooled together per participant. Additionally,
relevant parts of the transcripts were signed for further interpretation. The findings were then demonstrated through
frequencies.
3.5 Data analysis
As pinpointed in the literature, there are different ways to run the data analysis of phenomenographic studies. In
this study, we qualitatively analyzed the obtained data through content and thematic analysis. In essence, content
analysis aims to determine the presence and frequency of certain words, themes, or concepts within qualitative
data, while thematic analysis allows researchers to extract high-level themes/patterns across the data and their
relationship (Braun & Clarke, 2006). Moreover, to follow a standard structure in phenomenographic qualitative
analysis (PQA), we used Stenfors-Hayes et al.’s (2013) model, which includes seven steps, namely data familiarization,
condensation, comparing, contrasting, grouping, articulating, and final labeling. The process was iterative and the
researchers moved back and forth among interview transcriptions, extracted codes, marked quotes, memos, and
interpretations. They created a codebook in which the most frequent themes were tabulated along with their sample
interview responses and the target respondents. Next, the principles of trustworthiness were ensured, as pivotal
parts of qualitative research (Lincoln & Guba, 1985). The first principle was member checking for which we asked the
respondents to examine their transcripts and our interpretations of their responses. Second, intercoder agreement
8DERAKHSHAN and GHIASVAND
was safeguarded by giving 50% of the data to another researcher, who had sufficient knowledge and experience in
running qualitative studies and data analyses. After cross-checking the data, an agreement of 0.97 was obtained by
Cohen’s Kappa coefficient. Third, the principle of confirmability was adhered to by asking another experienced L2
researcher to audit trial the data analysis phase thoroughly. Credibility was the next principle assured in this study by
using purposive sampling, interactive interviews, peer debriefing, and bracketing personal experiences/perceptions.
Fifth, the researchers augmented the dependability of the findings by running content and thematic analysis, iterative
data transcription and coding, and coherently describing the steps taken in the study. Finally, the transferability of the
findings was strengthened by offering adequate descriptive data for future researchers to replicate and recontextu-
alize the study. Concerning researcher positionality, it is critical to note that, in this qualitative study, the researchers
tried to remain neutral as much as possible, while it is almost impossible to be 100% objective in qualitative designs.
Therefore, they were just the data collectors and analyzers.
4FINDINGS
4.1 Perceptions and potentials of ChatGPT for L2 education and research
To answer the research question that set out to unpack the potentials of ChatGPT for L2 education and research,
content and thematic analysis were conducted on interviews. The findings illustrated a gamut of opportunities for L2
learning, teaching, assessment, and research as listed in Table 1.
Of the potentials for L2 learning, “learner autonomy in learning” (n=23, 77%), “availability all the time” (n=20,
67%), “interactive and personalized learning experience” (n=16, 53%), “immediate feedback on students’ works”
(n=14, 47%), and fostering “intrinsic motivation” (n=12, 40%) were the most frequently repeated themes/codes
across the interviews. The following interview responses represent the extracted themes/codes:
Well, in my perspective, one of the most important benefits of AI and ChatGPT for L2 learning is encouraging
students to be autonomous and independent in their learning process.(Teacher, 24)
This AI-induced chatbot is a great opportunity to learn English anytime, anywhere. It is available always.
(Teacher, 2)
I believe that using this chatbot is effective in L2 learning because it can provide an interactiveand personalized
learning experience for L2 students.(Teacher, 8)
Since ChatGPT provides swift and immediate feedback, it is useful for students and their academic work.
(Teacher, 23)
L2 students like to be free in learning. Hence, the use of ChatGPT allows them to become intrinsicallymotivated
to learn more and more from their heart, not by force .(Teacher, 29)
Of the list of potentials for L2 teaching and pedagogy, the capability of ChatGPT in “teaching workload reduction”
(n=25, 83%), “identifying and producing teaching materials” (n=23, 77%), “supplementing traditional teaching
approaches” (n=20, 67%), “encouraging adaptive language teaching” (n=12, 40%), and “technology-integration in
the class” (n=10, 33%) were the most recurrent themes/codes as evidenced by the coming excerpts:
Based on my own experience, the use of ChatGPT significantly reduces teaching workload and pres-
sures because we as L2 teachers can use this tool to design activities, materials, tests, and lesson plans.
(Teacher, 20)
DERAKHSHAN and GHIASVAND 9
TAB L E 1 The potentials of ChatGPT for L2 education and research.
Influencing domains and potentials Frequency
L2 learning
•Information accessibility
•Immediate feedback on students’ works
•Interactive and personalized learning experience
•Self-reflection development
•Learner engagement
•Intrinsic motivation
•Learner autonomy in learning
•Availability all the time
•Cost-effective learning
(7)
(14)
(16)
(8)
(6)
(12)
(23)
(20)
(4)
L2 teaching
•Quick lesson plan design
•Teaching workload reduction
•Identifying and producing teaching materials
•Learner-centered pedagogy
•Supplement traditional teaching approaches
•Real-life conversational practice
•Encouraging adaptive language teaching
•Promoting one-on-one instruction
•Technology integration in the class
(8)
(25)
(23)
(6)
(20)
(5)
(12)
(4)
(10)
L2 assessment
•Encouraging self and peer assessment
•Creating different tests and quizzes
•Designing assessment rubrics
•Evaluating and grading students’ assignments
•Providing feedback on students’ writing tasks
•Fostering authentic and alternative assessment practices
•Digital and automated assessment
•Objective and adaptive testing
•Fast and cheap evaluation
•Immediate assessment feedback
(6)
(12)
(22)
(3)
(16)
(5)
(8)
(18)
(10)
(14)
L2 research
•Summarizing lengthy papers and theses
•Providing a first draft of research projects
•Providing large data
•Encouraging linguistic and discourse analysis
•Supporting the writing process
•Saving time finding relevant information
(23)
(14)
(6)
(4)
(10)
(20)
EFL teachers can use ChatGPT to provide materials for the class like poems, short essays, and conversational
prompts and questions. .. (Teacher, 9)
I think the use of technologies; especially those powered by AI can support and complement our own traditional
ways of teaching.(Teacher,5)
In my opinion, the use of ChatGPT can encourage EFL teachers to use an adaptive system of teaching since the
chatbot provides personalized and immediate feedback in line with students’ level of proficiency.(Teacher,3)
I strongly believe that the use of ChatGPT and other AI tools shift old schools to become updated via technology
integration in the class . (Teacher, 13)
10 DERAKHSHAN and GHIASVAND
Concerning L2 assessment, “designing assessment rubrics” (n=22, 73%), “objective and adaptive testing” (n=18,
60%), “providing feedback on students’ writing tasks” (n=16, 53%), “immediate assessment feedback” (n=14, 47%),
“creating different tests and quizzes” (n=12, 40%), and fast and cheap evaluation’’ (n=10, 33%). The following
excerpts represent the abovementioned themes/codes from interviews.
As EFL teachers, ChatGPT is a great opportunity to have assessment rubrics for assessing different language
skills and sub-skills. It easily designs rubrics for us. (Teacher, 17)
One of the coolest features of ChatGPT for L2 testing is its potential to be objective and adaptive
indistinguishable from humans . (Teacher, 28)
I have used this chatbot to assess my students’ writing assignments. It can provide immediate feedback
regarding different aspects of writing competency.(Teacher,8)
This AI-powered tool is able to give students’ immediate assessment feedback, which is much faster and
cheaper than other methods . (Teacher, 19)
I think this tool is beneficial for L2 testing in that we can ask it to design sample multiple-choice and open-ended
questions for mid-term exams and pop-up quizzes . (Teacher, 10)
Lastly, the analysis of interview data demonstrated that ChatGPT is meritorious for L2 research mostly because
of its capacity to “summarizing lengthy papers and theses” (n=23, 77%), “saving time finding relevant informa-
tion” (n=20, 67%), “providing a first draft of research projects” (n=14, 47%), and “supporting writing process”
(n=10, 33%).
Well, the introduction of ChatGPT is a real boomfor L2 research. It can be used to summarize long papers and
theses and this saves researchers’ time and energy. (Teacher, 20)
I think this bot is good enough for research in that it can save our time findingre sources and information related
to our research topic.(Teacher,6)
To me, one of the first uses of ChatGPT in L2 research is educating basic researchers on how to create a
first draft of their essays and articles. Although it is not flawless, it can place the first foundation stones.
(Teacher, 16)
Honestly,I myself consult with this AI tool when I amwriting different sections of my articles. It really generates
ideas, which are highly comparable to those produce by highly intelligent human beings. (Teacher, 29)
To conclude, the results of interview analysis in this research question revealed that ChatGPT might have different
benefits for L2 education and research. As for L2 learning, the most frequent themes pertained to its potential to pro-
mote “learner autonomy,”“interactive and personalized learning,” providing “immediate feedback,” and its “availability
all the time.” Regarding L2 teaching and pedagogy, the participants maintained that ChatGPT reduces their “teaching
workload,” identifies and produces teaching materials’’ for them, and “supplements traditional teaching approaches.”
Furthermore, it was found that ChatGPT is meritorious for L2 assessment mostly by designing “assessment rubrics,”
having an “objective and adaptive testing,” “providing writing feedback,” and offering “immediate assessment feed-
back” on students’ performance. Finally, the results revealed that ChatGPT was perceived beneficial for L2 research
in that it can “summarize lengthy papers and theses” and “save L2 researchers’ time finding relevant information” for
their scientific works.
DERAKHSHAN and GHIASVAND 11
FIGURE 1 The pitfalls of ChatGPT for L2 education and research. [Color figure can be viewed at
wileyonlinelibrary.com]
4.2 Perceived pitfalls of ChatGPT for L2 education and research
To answer this research question that examined the pitfalls of ChatGPT for L2 learning, teaching, assessment, and
research from the perspective of research-active EFL teachers, the second interview question was thematically ana-
lyzed. The results of a deep analysis of interview responses culminated in a list of pitfalls and threats behind the
use of ChatGPT in L2 education and research as depicted in Figure 1with frequency of occurrence in front of
each theme.
From the perspective of research-active EFL teachers, ChatGPT was also harmful and actually a threat to L2 edu-
cation and research. As for L2 learning, the participants argued that this tool is most dangerous for “killing creativity”
(n=21, 70%), “promoting fabricated information and fake assignments” (n=20, 67%), “discouraging students’ critical
thinking” (n=18, 60%), “ignoring the role of culture and context in language learning” (n=15, 50%), “overlooking stu-
dents’ emotions and emotional support” (n=9, 30%). The following interview responses evince the frequently raised
perils of ChatGPT for L2 learning.
I think ChatGPT is a double-edged sword. While it can be beneficial for L2 learning, it can also kill students’
creativity and criticality .(Teacher,4)
One of the threats of this AI chatbot to the L2 learning process is that some students may get addicted to it and
over-rely on fabricated information and fake assignments. (Teacher, 26)
12 DERAKHSHAN and GHIASVAND
In my idea, this tool is dangerous for language learning because it pays no attention to the role of context,
culture, and even the emotional aspects of L2 learning. To me, it is absurd. (Teacher 12)
Among the six pitfalls extracted from interview data regarding L2 teaching, the most frequent perils concerned “dam-
aging academic integrity” (n=18, 60%), “encouraging shallow understanding and literacy”(n=13, 43%), “losing human
interaction in teaching speaking and listening” (n=8, 27%), and “decreasing teacher-student interactions” (n=5, 17%)
as represented in the succeeding interview extracts:
I guess the most dreadful pitfall of ChatGPT in L2 teaching is promoting dishonesty in academia. It can answer
exams like a human and this damages academic integrity. (Teacher, 25)
Well, I have worked and examined the language produced by this chatbot. The sentences and responses are
not deep but taken from a large corpus. Therefore, teaching with toolsencourages shallow understanding and
literacy. It cannot produce pure literacy . (Teacher, 11)
In my opinion, overreliance on this chatbot is dangerous, especially when teachingspeaking and listening skills.
These skills require human interaction,but ChatGPTlacks interaction. It exponentially redu ces teacher-student
interactions in the class, to me. (Teacher, 30)
With regard to L2 assessment, ChatGPT was considered hazardous mostly because it increases “cheating” (n=22,
73%), “overlooks nonverbal signs during speaking tests” (n=18, 60%), “endangers security in online exams” (n=15,
50%), and provides an “inaccurate assessment” (n=10, 33%). Others also argued that ChatGPT assessment is “bias”
(n=7, 23%), and “incomplete” since it disregards some aspects of speaking and listening skills (n=5, 17%). Below are
some sample interview responses encompassing the extracted themes/codes mentioned.
As research shows, ChatGPT is able to take online exams instead of students. Since it can generate human-like
sentences and even long paragraphs, in L2 testing it causes cheating and insecurity in online exams .(Teacher,
19)
I guess, ChatGPT is weak in assessment, especially in L2 speaking tests. It cannot realize or produce non-verbal
messages. Hence, its evaluation of speaking is inaccurate. The paralinguistic features of the language are
overlooked in this bot.(Teacher,3)
As this AI-powered chatbot is based on large data sets taken from native culture, context, and resources, its
evaluation of students’ performance may not be fair in non-native English-speaking countries. (Teacher, 17)
Finally, the analysis of interviews provided a list of threats that ChatGPT may have for L2 research. Of the nine
extracted pitfalls, the most frequent ones included “encouraging high-tech plagiarism” (n=23, 77%), “spreading mis-
information in the world of science” (n=20, 67%), “ignoring codes of ethics in research” (n=16, 53%), “promoting
patchwork writing” (n=12, 40%), and “reliability and validity concerns” (n=7, 23%).
One of the biggest problems of ChatGPT for L2 research is promoting asort of high-tech plagiarism trend
in academia as pinpointed by Noam Chomsky, too. This chatbot produces data based on pre-figured and
feedsinformation from published resources. Therefore, it may generate plagiarized content in articles and
assignments . (Teacher, 22)
The data produced by ChatGPT are not absolutely like human beings. They are based on algorithms gener-
ated by AI and machines. So, it is unlikely be representative of human’s behaviors and perceptions. This poses
challenges to the reliability and validity of data and results of studies powered by ChatGPT. (Teacher, 12)
DERAKHSHAN and GHIASVAND 13
Despite its opportunities, ChatGPT can really ruin the world of science by encouraging students to provide fake
and misinformation in their research projects. It also totally ignores the roleof ethical codes/considerations like
privacy, consent, deception, etc .(Teacher,5)
If you ask me, I would say that ChatGPT makes L2 students and researchers lazy because it really inspires them
to do patchwork writing rather than real, original writing of articles and writing assignments. (Teacher, 14)
In sum, the results of content and thematic analysis illustrated that the use and overreliance on ChatGPT pose several
threats to L2 learning, teaching, assessment, and research. The most frequently raised pitfalls for L2 learning included
“killing creativity,” “promoting fabricated information and fake assignments,” “discouraging students’ critical thinking,”
and “ignoring the role of culture and context in language learning.” Concerning L2 teaching, the participants main-
tained that ChatGPT is perilous to “academic integrity,” “deep and pure literacy,” and teacher–student interactions
in the classroom. Additionally, it was found that this AI chatbot hurts L2 assessment by “encouraging cheating, espe-
cially in online exams,” “ignoring paralinguistic, nonverbal features in speaking and listening tests,” and “providing an
inaccurate assessment” of students’ language competence. Finally, the participants pinpointed that the integration of
this tool into L2 research poses dangerous threats to the world of science by “spreading fake and misinformation,”
encouraging “high-tech plagiarism” and “patchwork writing,” and overlooking ethical considerations of research.
5DISCUSSION
This study was an endeavor to disclose the potentials and pitfalls of ChatGPT for L2 learning, teaching, assessment, and
research from the perspective of research-active EFL teachers. The results of content and thematic analysis revealed
that ChatGPT has different benefits and perils for L2 education and research. Concerning benefits, it was found that
this language bot could promote L2 learning by enhancing EFL students’ autonomy,creating an interactive and person-
alized learning experience for them, providing immediate feedback, and being convenient and available all the time.
This is in accord with previous studies (e.g., Agustini, 2023; Baidoo-Anu & Ansah, 2023;Lo,2023) which reported that
ChatGPT is influential in various aspects of learning including feedback, easy information access, autonomy, and per-
sonalization of learning processes. This finding can be explained by the idea that technologies and AI are cutting-edge
ways to immerse L2 students in their learning process (Bin-Hady et al., 2023). The affordances of ChatGPT and its
newness may engage students in autonomous learning where they can have immediate and personalized feedback
from the bot. Another alternative reason might be the nature of AI technologies that foster one’s access to information
drawing on mass data. Theoretically, the findings corroborate the utility of AI-powered bots in educational domains.
This study also revealed that ChatGPT is meritorious for L2 teaching in that it reduces teachers’ teaching work-
load, produces teaching materials, and supports traditional teaching approaches. These findings confirm and mirror
those reported in prior research (e.g., Farrokhnia et al., 2023) in which ChatGPT was perceived as influential in several
pedagogical aspects of education, as a whole. It is likely that these results are due to advancements in e-education,
especially with the sudden shift toward technology-based instruction during the pandemic. The participants seem
to be aware of L2 teaching complexities and workloads, which might have led them to consider AI technologies as
opportunities to facilitate language teaching. Since ChatGPT is by nature a language model, it can presumably benefit
language education and pedagogy more than other fields. Another justification can be the participants’ dual role as
teachers and researchers making them cognizant of the potential of this bot for soothing L2 teaching complications.
Another important finding was that ChatGPT could enlighten L2 assessment by assisting EFL teachers in designing
assessment rubrics, having an objective and adaptive testing, providing writing feedback, and offering immediate
assessment feedback on students’ assignments and tasks. This finding complies with several studies in the literature,
which endorsed the positive impact of generative AI (e.g., ChatGPT) on testing and assessment procedures (e.g.,
Rudolph et al., 2023; Swiecki et al., 2022;Zhai,2022). Such studies argued that this AI-powered bot helps L2 teachers
14 DERAKHSHAN and GHIASVAND
and assessors in designing tests and quizzes, assessing students’ writing assignments, encouraging self and peer
assessment, designing assessment rubrics, offering immediate assessment feedback, providing a personalized, adap-
tive, and cheap testing system. The findings shed more light on computer-assisted language testing (CALT) theories
and computer-adaptive tests (CAT), which corroborate the role of technologies in language testing and assessment.
This finding can be attributed to the participants’ language assessment literacy (LAL) and strong assessment identity
(Estaji & Ghiasvand, 2022). As the use of technologies in language testing is a common core of both LAL and assess-
ment identity, the research-active teachers have considered ChatGPT and AI as promising for L2 assessment from
various angles.
The last beneficiary domain was L2 research for which the results evinced that ChatGPT can summarize lengthy
papers and theses and save L2 researchers’ time finding relevant information for their scientific works. This find-
ing reflects those obtained by Aydin and Karaarslan (2022), who pinpointed that ChatGPT can summarize literature
reviews and find related information swiftly. Furthermore, the study is in keeping with Zhai (2022) and Dowling and
Lucey (2023), who found ChatGPT as a facilitator of article writing and idea generation.
Concerning the second research question, the results illustrated that ChatGPT is perilous for L2 learning in the
sense that it may kill creativity, promote fabricated information and fake assignments, discourage students’ critical
thinking, and ignore the role of culture/context in language learning. This finding matches those obtained by previous
scholars (e.g., Buriak et al., 2023; Farrokhnia et al., 2023; Susnjak, 2022; Tlili et al., 2023;Zhai,2022). The partici-
pants’ actual experience of using ChatGPT in L2 classes may be the reason behind this finding. Another explanation
can be the idea that ChatGPT is in its infancy and many of these problems have not been thought of by its develop-
ers. This intriguing finding might be a result of EFL teachers’ technological pedagogy content knowledge (TPACK) that
describes sets of knowledge that a teacher needs to successfully integrate technologies in educational milieus. The
participants seemed to know the blind spots of ChatGPT and AI integration in L2 learning.
Another finding, in this study, was that ChatGPT was perceivedas dangerous for L2 teaching by damaging academic
integrity, deep and pure literacy, and teacher–student interactions in the classroom. This agrees with the findings
reported by Cotton et al. (2023), Dimitrov (2023), and Farrokhnia et al. (2023), who, respectively, asserted that Chat-
GPT leads to superficial literacy, ruins academic integrity, and reduces human interactions in the classroom. This
finding might be attributable to the participants’ pedagogical reasoning, informed professionalism, and TPACK, which
form their thinking, actions, and pedagogical decisions. When instructional reasoning undergirds an EFL teacher’s
ambitious pedagogy, he/she can identify and tackle pedagogical dilemmas and uncertainties argumentatively. Hence,
it can be claimed that these qualities and professional judgment abilities of the participants have made them aware
of the challenges of ChatGPT for their pedagogy. Another alternative logic might be EFL teachers’ high digital and
academic literacies that formed their acceptance, adoption, and implementation of innovative technologies in their
classes
Another compelling finding, in this study, was that ChatGPT might damage L2 assessment by encouragingcheating,
ignoring paralinguistic features in aural tests, and providing an inaccurate picture of students’ language competence.
These pitfalls have been reported in previous studies (e.g., Cotton et al., 2023), as well. These studies also maintained
that ChatGPT endangers test security in online examinations and fairness concerns. This finding can be explained
by EFL teachers’ concern for test reliability, validity, and language competence components in testing through
technologies and AI. Another reason can be their high digital assessment literacy, which summons knowledge and
skills in conducting assessment practices in virtual settings.
Finally, this study indicated that the integration of ChatGPT into L2 research might jeopardize L2 research by
spreading fake and misinformation, encouraging high-tech plagiarism and patchwork writing, and neglecting codes
of ethics in research. The results support that of Gašević et al. (2023), who opined that ChatGPT misuse might
end in plagiarism. Likewise, this finding echoes those of Liebrenz et al. (2023), Zhuo et al. (2023), and Mhlanga
(2023), who contended that ChatGPT pollutes science and research by spreading fake information, encouraging
patchwork writing, and ignoring the essentiality of ethical concerns. The participants’ experience in teaching and
researching L2 education and educational technology may have made them aware of these dangerous traps. Another
DERAKHSHAN and GHIASVAND 15
explanation for this outcome may be EFL teachers developed scholarly identity over years of teaching and run-
ning/supervising numerous articles and theses related to educational technology and L2 education. While the study
can strengthen the interface of AI and L2 education, it is not clear if all these findings have come from teachers’
knowledge and identity or other intervening factors. This is left to future investigators to unravel other aspects and
territories.
6CONCLUSION AND IMPLICATIONS
Based on the findings, it can be inferred that, like other disciplines, the use of ChatGPT in L2 education and research is
a double-edged sword. It can produce both promising and disappointing outcomes in different areas of L2 education
depending on how it is used. It can also be concluded teachers’ pedagogical expertise and TPACK may enhance the ben-
efits of this AI-powered bot and minimize its perils. Knowledge in this area is still in its inception and L2 researchers and
practitioners have just initiated their long journey to integrate AI in L2 education. The demands of a shifting world that
calls for a new generation of cutting-edge technologies in education require EFL teachers to be digitally literate. Mod-
ern and professional teachers are now characterized by numerous types of literacies on top of which digital literacy
resides. This study, as one of the most comprehensive studies on the intersection of AI technology and L2 education,
offers several implications for research and practice in this domain.
Theoretically,this study contributes to the EFL community’s understanding of theories related to AI technology and
technology integration in L2 education. It can also enrich LAL and TPACK research by adding to them a new dimension
(AI technology competence). The results can also be practically momentous for EFL teachers and educators in that they
can take logical steps to leverage the interface of AI and L2 teaching and learning via evidence-informed professional
development courses. In today’s shifting world of education, EFL teacher trainers need to modify and reinforce their
programs in a way that traditional teaching, testing, and research practices of novice and experienced teachers are
supplemented by innovative AI-based technologies. Moreover, L2 educators may find this study beneficial by using
its ideas as a giant leap from old literacies toward modern, digital literacies. Now, EFL teachers need to connect their
professional knowledge, informed judgments, and instructional reasoning to new modes of education. Many of such
aspects may lose effect in a world governed by AI and superintelligent machines.
L2 researchers and test developers may also use the findings to develop strategies and examinations in which
the pitfalls of ChatGPT are kept at a minimum, while academic integrity and ethics are prioritized. Test developers
can assess EFL students’ academic performance by items, which require deep knowledge, criticality, communication,
and personal analysis. Open-ended and follow-up questions can be added to test questions to reduce the likelihood
of “high-tech plagiarism” by ChatGPT. Researchers can use this tool to facilitate the writing process by requesting
unclear points from it. They can use ChatGPT to translate abstracts to journals that require non-English abstracts
for submission. Academic staff in universities may also use this study and design courses in which proper uses of
this chatbot, as well as its misuses, are explained. They can use plagiarism detectors, structured guidelines, close
monitoring of task completion, and language analysis tools to see if the responses are produced by students or AI bots.
Finally, policymakers need to revise their conceptualizations and decisions regarding the integration of technology
and AI into L2 education. In case the benefits and misfits of ChatGPT are elucidated to all stakeholders, it can help the
whole educational system to flourish in the modern world.
ACKNOWLEDGMENTS
This study was sponsored by Golestan University under grant number: 1571. The authors are grateful to the insightful
comments suggested by the editor and the anonymous reviewers.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflict of interest.
16 DERAKHSHAN and GHIASVAND
DATA AVAILABILITY STATEMENT
The datasets generated and analyzed during the current study are available from the corresponding author on
reasonable request.
ORCID
Ali Derakhshan https://orcid.org/0000-0002-6639-9339
Farhad Ghiasvand https://orcid.org/0000-0002-6599-3838
PEER REVIEW
The peer review history for this article is available at https://publons.com/publon/10.1111/ijal.12561
INFORMED CONSENT
Informed consent letters were obtained from all the individual participants included in this study.
REFERENCES
Agustini, N. P. O. (2023). Examining the role of ChatGPT as a learning tool in promoting students’ English language learning
autonomy relevant to Kurikulum Merdeka Belajar. Edukasia: Journal Pendidikan Dan Pembelajaran,4(2), 921–934.
Ali, J. K. M., Shamsan, M. A. A., Hezam, T. A., & Mohammed, A. A. (2023). Impact of ChatGPT on learning motivation: Teachers
and students’ voices. Journal of English Studies in Arabia Felix,2(1), 41–49.
Aljanabi, M., & Ghazi, M., Ali, A. H., & Abed, S. A. (2023). ChatGPT: Open possibilities. Iraqi Journal for Computer Science and
Mathematics,4(1), 62–64.
Almusaed, A., Almssad, A., Yitmen, I., & Homod, R. Z. (2023). Enhancing student engagement: Harnessing “AIED”’s power in
hybrid education—A review analysis. Education Sciences,13(7), 632.
Aydin, Ö., & Karaarslan, E. (2022). OpenAI ChatGPT generated literaturereview: Digital twin in healthcare. In Ö. Aydin (Ed.), ,
Emerging computer technologies (pp. 22–31). ˙
Izmir Akademi Dernegi.
Baidoo-Anu, D., & Ansah, L. O. (2023). Education in the eraof generative artificial intelligence (AI): Understanding the potential
benefits of ChatGPT in promoting teaching and learning. Journal of AI,7(1), 52–62.
Bearman, M., & Ajjawi, R. (2023). Learning to work with the black box: Pedagogy for a world with artificial intelligence. British
Journal of Educational Technology,54, 1160–1173. https://doi.org/10.1111/bjet.13337
Bin-Hady, W. R. A., Al-Kadi, A., Hazaea, A., & Ali, J. K. M. (2023). Exploring the dimensions of ChatGPT in English language
learning: A global perspective. Library Hi Tech. https://doi.org/10.1108/lht-05-2023-0200
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology,3(2), 77–101.
Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell,
A., Agarwal, S., Herbert-Voss, A., Krueger, G., Henighan, T., Child, R., Ramesh, A., Ziegler, D., Wu, J., Winter, C., .. .
Amodei, D. (2020). Language models are few-shot learners. Advances in Neural Information Processing Systems,33, 1877–
1901.
Buriak, J. M., Akinwande, D., Artzi, N., Brinker, C. J., Burrows, C., Chan, W. C., Chen, C., Chen, X., Chhowalla, M., Chi, L., Chueh,
W., Crudden, C. M., Di Carlo, D., Glotzer, S. C., Hersam, M. C., Ho, D., Hu, T. Y., Huang, J., Javey, A., ... Ye, J. (2023). Best
practices for using AI when writing scientific manuscripts: Caution, care, and consideration: Creative science depends on
it. ACS Nano,17, 4091–4093.
Cascella, M., Montomoli, J., Bellini, V., & Bignami, E. (2023). Evaluating the feasibility of ChatGPT in healthcare: An analysis
of multiple clinical and research scenarios. Journal of Medical Systems,47(1), 1–5. https://doi.org/10.1007/s10916-023-
01925-4
Christopoulos, A., & Sprangers, P. (2021). Integration of educational technology during the Covid-19 pandemic: An analysis of
teacher and student receptions. Cogent Education,8(1), 1964690.
Cooper, K. (2021). OpenAI GPT-3: Everything you need to know. Springboard.https://www.springboard.com/blog/data-
science/machine-learning-gpt-3-open-ai/
Cotton, D. R., Cotton, P. A., & Shipway, J. R. (2023). Chatting and cheating: Ensuring academic integrity in the era of ChatGPT.
Innovations in Education and Teaching International,61, 228–239. https://doi.org/10.1080/14703297.2023.2190148
Creswell, J. W. (2013). Qualitativeinquiry and research design: Choosing among fiveapproaches. Sage Publications.
Derakhshan, A., Teo, T., & Khazaie, S. (2024). Is game-based language learning general or specific-oriented?
Exploring the applicability of a mobile virtual reality-based flipped classroom for medical English communi-
cation in the Middle East. Computers & Education,213, 105013. https://doi.org/10.1016/j.compedu.2024.10
5013
DERAKHSHAN and GHIASVAND 17
Dimitrov, M. (2023). What business leaders should know about using LLMS like ChatGPT.https://www.forbes.com/sites/
forbesbusinesscouncil/2023/02/07/what-business-leaders-should-know-aboutusing-llms-like-chatgpt/
Dörnyei, Z. (2020). Innovations and challenges in language learning motivation. Routledge.
Dowling, M., & Lucey, B. (2023). ChatGPT for (finance) research: The bananarama conjecture. Finance Research Letters,53,
103662. https://doi.org/10.1016/j.frl.2023.103662
Dwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., Baabdullah, A. M., Koohang, A., Raghavan, V., Ahuja,
M., Albanna, H., Albashrawi, M. A., Al-Busaidi, A. S., Balakrishnan, J., Barlette, Y., Basu, S., Bose, I., Brooks, L ., Buhalis, D., .. .
Wright, R. (2023). “So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implica-
tions of generative conversational AI for research, practice and policy. International Journal of Information Management,71,
102642.
Elkins, K., & Chun, J. (2020). Can GPT-3 pass a writer’s Turingtest? Journal of Cultural Analytics,5(2), 17212. https://doi.org/10.
22148/001c.17212
Estaji, M., & Ghiasvand, F. (2022). Teacher assessment identity in motion: The representations in e-portfolios of novice and
experienced EFL teachers. Issues in Language Teaching,11(2), 33–66. https://doi.org/10.22054/ilt.2022.70302.741
Farrokhnia, M., Banihashem, S. K., Noroozi, O., & Wals, A. (2023). A SWOT analysis of ChatGPT: Implications for educational
practice and research. Innovations in Education and Teaching International,61, 460–474. https://doi.org/10.1080/14703297.
2023.2195846
Fathi, J., Rahimi, M., & Derakhshan, A. (2024). Improving EFL learners’ speaking skills and willingness to communicate via
artificial intelligence-mediated interactions. System,121, 103254. https://doi.org/10.1016/j.system.2024.103254
Gao, C. A., Howard, F. M., Markov, N. S., Dyer, E. C., Ramesh, S., Luo, Y., & Pearson, A. T. (2022). Comparing scientific abstracts
generated by ChatGPT to original abstracts using an artificial intelligence output detector, plagiarism detector, and blinded
human reviewers. BioRxiv,12, 1–18. https://doi.org/10.1101/2022.12.23.521610
Gašević, D., Siemens, G., & Sadiq, S. (2023). Empowering learners for the age of artificial intelligence. Computers and Education:
Artificial Intelligence,4, 100130. https://doi.org/10.1016/j.caeai.2023.100130
Gibson, D., Kovanovic, V., Ifenthaler, D., Dexter, S., & Feng, S. (2023). Learning theories for artificial intelligence promoting
learning processes. British Journal of Educational Technology,54(5), 1125–1146. https://doi.org/10.1111/bjet.13341
Grossman, G. (2020). We’re entering the AI twilight zone between narrow and general AI. Venture Beat.https://venturebeat.
com/ai/were-entering-the-ai-twilightzone-between-narrow-and-general- ai/
Hajar, A. (2021). Theoretical foundations of phenomenography: A critical review. Higher Education Research & Development,
40(7), 1421–1436. https://doi.org/10.1080/07294360.2020.1833844
Haque, M. U., Dharmadasa, I., Sworna, Z. T., Rajapakse, R., & Ahmad, H. (2022). I think this is the most disruptive technol-
ogy: Exploring sentiments of ChatGPT early adopters using Twitterdata. ArXiv, 1–12. https://doi.org/10.48550/arXiv.2212.
05856
Humble, N., & Mozelius, P. (2019, October). Teacher-supported AI or AI-supported teachers. Paper presented at European Con-
ference on the Impact of Artificial Intelligence and Robotics (ECIAIR). The UK: Oxford. http://dx.doi.org/10.34190/ECIAIR.
19.007
Kasneci, E., Seßler, K., Küchemann, S., Bannert, M., Dementieva, D., Fischer, F., Gasser, U., Groh, G., Günnemann, S., Hüllermeier,
E., Krusche, S., Kutyniok, G., Michaeli, T., Nerdel, C., Pfeffer, J., Poquet, O., Sailer, M., Schmidt, A., Seidel, T., . .. Kasneci, G.
(2023). ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual
Differences,103, 102274.
Liebrenz, M., Schleifer, R., Buadze, A., Bhugra, D., & Smith, A. (2023). Generating scholarly content with ChatGPT: Eth-
ical challenges for medical publishing. Lancet Digital Health,5(3), e105–106. https://doi.org/10.1016/S2589-7500(23)
00019-5
Lincoln, Y. S., & Guba, E. G. (1985). Naturalisticinquiry.Sage.
Lo, C. K. (2023). What is the impact of ChatGPT on education? A rapid review of the literature. Education Sciences,13(4), 410.
Manyika, J., Chui, M., Miremadi, M., Bughin, J., George, K., Willmott, P., & Dewhurst, M. (2017). A future that works: Automation,
employment, and productivity. McKinsey Global Institute.
Mhlanga, D. (2023). Open AI in education, the responsible and ethical use of ChatGPT towards lifelong learning. In D., Mhlanga
(Ed.), FinTech and Artificial Intelligence for Sustainable Development: The Role of Smart Technologies in Achieving Development
Goals (pp. 387–409). Springer Nature.
Muñoz, S. A. S., Gayoso, G. G., Huambo, A. C., Tapia, R. D. C., Incaluque, J. L., Aguila, O. E. P., Cajamarca, J. C. R., Acevedo, J.E. R.,
Rivera, H. V. H., & Arias-Gonzáles, J. L. (2023). Examining the impacts of ChatGPT on student motivation and engagement.
Social Space,23(1), 1–27.
Murati, M. (2022). ([@miramurati].). Tweets [Twitter profile]. https://twitter.com/miramurati/status/1599796191243669504
Patton, M. Q. (2002). Qualitative research & evaluation methods. Sage Publications.
Pedro, F., Subosa, M., Rivas, A., & Valverde,P. (2019). Artificial intelligence in education: Challenges and opportunities for sustainable
development.UNESCO.
18 DERAKHSHAN and GHIASVAND
Roose, K. (2022, December 5). The brilliance and weirdness of ChatGPT. The New YorkTimes.https://www.nytimes.com/2022/
12/05/technology/chatgpt-ai-twitter.html
Rudolph, J., Tan, S., & Tan, S. (2023). ChatGPT: Bullshit spewer or the end of traditional assessments in higher education? Journal
of Applied Learning & Teaching,6(1), 242–263. https://doi.org/10.37074/jalt.2023.6.1.9
Sharma, S., & Yadav,R. (2022). Chat GPT–A technological remedy or challenge for education system. Global Journal of Enterprise
Information System,14(4), 46–51.
Sin, S. (2010). Considerations of quality in phenomenographic research. International Journal of Qualitative Methods,9(4), 305–
319.
Stenfors-Hayes, T., Hult, H., & Dahlgren, M. A. (2013). A phenomenographic approach to research in medical education. Medical
Education,47(3), 261–270. https://doi.org/10.1111/medu.12101
Stolz, S. A. (2020). Phenomenology and phenomenography in educational research: A critique. Educational Philosophy and
Theory,52(10), 1077–1096.
Susnjak, T. (2022). ChatGPT: The end of online exam integrity? ArXiv, 1–21. https://doi.org/10.48550/arXiv.2212.09292
Swiecki, Z., Khosravi, H., & Chen, G. (2022). Assessment in the age of artificial intelligence. Computers and Education: Artificial
Intelligence,3, 100075. https://doi.org/10.1016/j.caeai.2022.100075
Tlili, A., Shehata, B., Adarkwah, M. A., Bozkurt, A., Hickey, D. T., Huang, R., & Agyemang, B. (2023). What if the devil is my
guardian angel: ChatGPT as a case study of using chatbots in education. Smart Learning Environments,10(1), 15. https://
doi.org/10.1186/s40561-023-00237-x
Urban, M., Dˇ
echtˇ
erenko, F., Lukavský, J., Hrabalová, V., Svacha, F., Brom, C., & Urban, K. (2024). ChatGPT improves creative
problem-solving performance in university students: An experimental study. Computers & Education, 105031.
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017, December 4–9).
Attention is all you need. Proceedings of NeurIPS, Longbeach, California USA. 5998–6008.
Welle, D. (2023). ChatGPT is changing education, AI experts say — but how? https://news.abs- cbn.com/spotlight/01/25/23/
chatgpt-is-changing-education-ai-experts-say-but-how
Wogu, I. A. P., Misra, S., Olu-Owolabi, E. F., Assibong, P. A., & Udoh, O. D. (2019). Artificial intelligence, artificial teachers and
the fate of learners in the 21st century education sector: Implications for theory and practice. International Journal of Pure
and Applied Mathematics,119(16), 2245–2259.
Wu, R., & Yu, Z. (2023). Do AI chatbots improve students learningoutcomes? Evidence from a meta-analysis. British Journal of
Educational Technology,55(1), 10–33. https://doi.org/10.1111/bjet.13334
Xu, L. D., Lu, Y., & Li, L. (2021). Embedding blockchain technology into IoT for security: A survey. IEEE Internet of Things Journal,
8(13), 10452–10473. https://doi.org/10.1109/JIOT.2021.3060508
Yan, D. (2023). Impact of ChatGPT on learners in a L2 writing practicum: An exploratory investigation. Education and
Information Technologies,28, 1–25.
Zhai, X. (2022). ChatGPT user experience: Implications for education. SSRN Electronic Journal, 1–18. https://doi.org/10.2139/
ssrn.4312418
Zhuo, T. Y., Huang, Y., Chen, C., & Xing, Z. (2023). Exploring AI ethics of ChatGPT: A diagnostic analysis. arXiv: 2301.12867.
https://doi.org/10.48550/arXiv.2301.12867
How to cite this article: Derakhshan, A., & Ghiasvand, F. (2024). Is ChatGPT an evil or an angel for second
language education and research? A phenomenographic study of research-active EFL teachers’ perceptions.
International Journal of Applied Linguistics, 1–19. https://doi.org/10.1111/ijal.12561
APPENDIX
(Interview Questions)
Before attending the interview, I, hereby, confirm my agreement and willingness to take part in this research study.
Agree
Disagree
DERAKHSHAN and GHIASVAND 19
Part 1: Background information
1. Age: ............
2. Gender:............
3. Teachingexperience ....................years.
4. Fieldofstudy:............................
5. Degree of familiarity with AI and ChatGPT: low average high
6. Have you ever practiced ChatGPT in your classes?
7. Have you been trained on how to use this chatbot?
Part 2: Teachers’ perceptions of ChatGPT potentials and pitfalls
1. What is your idea about the integration of technologies, especially AI technology into L2 education? Can you
explain your view?
2. What are some possible benefits or potentials of ChatGPT as an AI-powered chatbot for L2 learning, teaching,
assessment, and research? Can you list them?
3. What is your opinion about the threats and pitfalls of using ChatGPT in L2 education and research? Would you
please provide a list of such perils for L2 learning, teaching, assessment, and research separately?
AUTHOR BIOGRAPHIES
Ali Derakhshan is professor of applied linguistics at the English Language and Literature Department, Golestan
University, Gorgan, Iran. He has been a member of the Iranian Elites Foundation since 2015. He has also been
selected as a distinguished researcher by the Teaching English Language and Literature Society of Iran in 2021.
His name appeared in Stanford University’s list of world’s top 2% of most influential scientists in 2022 and
2023. He has published in accredited international journals, including Computers and Education,Language Teaching
Research,System,Assessing Writing,Applied Linguistics Review,Studies in Second Language Learning and Teaching,Jour-
nal of Multilingual and Multicultural Development,ELT Journal,International Review of Applied Linguistics in Language
Teaching, International Journal of Applied Linguistics,Thinking Skills and Creativity,Current Psychology,Asia Pacific Edu-
cation Researcher,Educational Studies,Pragmatics and Society,Journal of Psycholinguistic Research,Porta Linguarum,
Revista Ibérica, and so on. His monograph The “5Cs” positive teacher interpersonal behaviors: Implications for learner
empowerment and learning in an L2 context was published by Springer in 2022. His co-authored book Instructed
second language pragmatics for the speech acts of request, apology, and refusal: A meta-analysis has been recently
published by Springer. His research interests are positive psychology, teacher education, learner individual differ-
ences, cross-cultural interpersonal factors in educational psychology, interlanguage pragmatics, and intercultural
communication.
Farhad Ghiasvand is a postdoctoral research fellow of applied linguistics at Allameh Tabataba’i University, Tehran,
Iran. He was selected as an outstanding MA and Ph.D. student by Iran’s National Elites Foundation. He also won
Allameh Tabataba’i University’s Top Researcher Award in 2023 based on his annual publications. His main areas
of research include teacher identity, language testing and assessment, washback effect, English for academic
purposes, and positive psychology.
