See discussions, stats, and author profiles for this publication at: [Link]

net/publication/384732061

Enhancing L2 Writing with Generative AI: A Systematic Review of Pedagogical

Integration and Outcomes

Preprint · October 2024

DOI: 10.13140/RG.2.2.19572.16005

CITATIONS READS
2 1,176

2 authors, including:

Hui Wang
University of Arizona
9 PUBLICATIONS 23 CITATIONS

SEE PROFILE

All content following this page was uploaded by Hui Wang on 08 October 2024.

The user has requested enhancement of the downloaded file.

 Enhancing L2 Writing with Generative AI: A
Systematic Review of Pedagogical Integration and
Outcomes
Hui Wanga , Anh Danga
a
Second Language Acquisition and Teaching, University of Arizona, Tucson, USA

Abstract
The rapid development of Generative Artificial Intelligence (GenAI) opens up
new possibilities for the fields of education. GenAI tools can serve as writing
assistants, supporting student writers by producing human-like text and sug-
gestions. With the growing adoption of these tools, this study examines how
instructors have integrated GenAI in second language (L2) writing classes to
support students’ writing skills. We focus on the AI-assisted learning expe-
riences, including student population, teaching contexts, pedagogical design,
measurements, and learning outcomes. 21 empirical studies were identified
for this systematic review. Results indicate that GenAI tools are primarily
applied for brainstorming ideas, assessing and providing feedback, generat-
ing writing samples for analysis, scaffolding writing process, raising GenAI
literacy, and developing critical thinking skills. The research also emphasizes
learning challenges and critical factors by reviewing the post-implementation
assessment, including reliance on GenAI tools, inaccurate information, cul-
tural bias, and language assimilation for L2 students. The findings inform
future AI-assisted design by highlighting teacher mediation and supervision
as well as teacher training on the use of GenAI in L2 writing activities. Future
practices and research should be encouraged to further explore the impact
and potential of AI-assisted writing instruction across a broader range of
academic levels and domains.
Keywords: Generative AI, Artificial Intelligence, ChatGPT, L2 writing,
AI-assisted instruction, AI-enhanced education

Preprint submitted to System October 8, 2024

1. Introduction
The rise and advancement of Generative Artificial Intelligence (GenAI),
such as ChatGPT, has achieved profound impacts on educational fields. This
powerful virtual assistant can enable human-like conversations and provides
thorough writing support, including generating ideas, composing essays, of-
fering personalized feedback, and polishing language to meet a range of needs
from writers [1, 2, 3]. Previous research has highlighted the inevitable pres-
ence of GenAI in students’ writing practices [4], although its benefits on
teaching and learning writing remain inconclusive [5]. For second language
(L2) writing, the absence of cultural diversity and inclusion in GenAI’s re-
sponses may exacerbate biases to the underrepresented L2 student group [6].
This begs a question of what it means for L2 students to integrate GenAI in
their writing processes, given the different positions that they’re being put
in. This complex nature of GenAI has led to a need for a systematic review of
actual AI-assisted L2 writing instruction implementation and its impacts on
students’ learning. However, limited systematic reviews have focused specif-
ically on the application of GenAI in writing education, particularly in the
context of L2 writing.
Driven by the current research gap, this systematic review focuses on
the integration of GenAI technologies in L2 writing courses to support L2
students’ writing skills. We will delve into the application of GenAI tools
through an examination of 21 empirical studies in diverse L2 teaching con-
texts. This study aims to explore the current application of GenAI in L2
writing pedagogy and inform future AI-assisted L2 writing instructional de-
signs.

2. Literature review
Generative Artificial Intelligence (GenAI) has experienced rapid develop-
ment and brought profound impacts in the field of education. AI-powered
chatbots, such as ChatGPT, are designed to provide immediate human-like
conversations to help people solve problems in writing tasks ranging from lin-
guistic errors to genre-specific patterns [7, 8]. Numerous studies have been
conducted on the implications of GenAI in education. For writing-intensive
courses, GenAI has been employed to achieve a range of goals, including
evaluating student writing [9, 10], offering corrective feedback [11, 12], gen-
erating samples for analysis [13, 14], fostering student collaboration [15], and

2
enhancing students’ learning experiences [16]. On the other hand, due to the
nature of these GenAI tools, there are also discussions regarding the risks of
integrating them in writing courses. As banning GenAI from schools is not
an option [17, 18], ethical concerns regarding plagiarism and academic in-
tegrity have become central of these discussions [19, 20, 21], along with other
unignorable concerns, such as overreliance and inaccuracy [22, 23]. When it
comes to the context of L2 writing, more complex perspectives and nuanced
voices tend to emerge, highlighting the additional challenges faced by L2
student writers. For example, according to Bender et al. [6], GenAI models
were primarily trained on the internet-based language dataset, where white
supremacist and misogynistic, ageist, etc. views are often overrepresented.
Training on these datasets worsens diversity and inclusion in their responses
and in turn enforces White Mainstream English and amplifies biases and
harms to the underrepresented L2 student population [6]. Additionally, cur-
rent GenAI detectors often misclassify L2 writing as AI-generated texts, rais-
ing bias, fairness, and robustness concerns for L2 students [24, 25]. Overall,
GenAI’s impacts on L2 writing education remain inconclusive and compli-
cated. Such diverse discussion leads to a pressing need for a systematic
review of previous empirical research on AI-assisted L2 writing instruction
with a focus on the diverse student populations, guiding concepts, GenAI
application approaches, and learning outcomes.
Several existing systematic review studies have explored GenAI’s impacts
in the educational area. For instance, Dempere et al. [26] reviewed 143
studies on the use of ChatGPT in higher education, identifying significant
benefits and risks, and recommended thoughtful and responsible integration
of GenAI within the higher education context. Li et al. [27] analyzed the
design of learning tasks for AI education in K-12 contexts and revealed its
effectiveness in fostering understanding and engagement as well as learning
challenges in developing AI literacy and solving technical issues. Labadze
et al. [28] investigated the ways in which teachers and students used AI
chatbots in their teaching and learning and addressed its important benefits
in providing grammatical corrections and syntactic suggestions.
Given the potential of GenAI, educators have expressed interest and like-
lihood to use GenAI in writing courses while acknowledging that there will
be hesitation, and uneasiness regarding AI-assisted writing instruction [29].
However, there have been few systematic reviews focusing on how GenAI
was applied in writing education specifically. This scarcity is even more
prominent in the context of L2 writing. Dang & Wang [17] examine over

3
100 universities’ publicly available statements on the use of GenAI for the
teaching of writing, and suggest educators to consider diverse student pop-
ulations in their GenAI policies and advocate for more related research on
the relationship between L2 writing and GenAI to amplify the voices of L2
students in the era of GenAI. Pedagogical implications and accommodation
strategies for L2 students should receive more attention in discussions about
the integration of GenAI in education.
Therefore, driven by the existing research gap and the need for a compre-
hensive understanding of AI-assisted L2 writing instruction, our systematic
review analyzes both pedagogical approaches and implications of integrating
GenAI technologies in L2 writing courses. We aim to explore student popu-
lations and contexts, pedagogical strategies, learning focuses, measurements,
and learning outcomes. Our research questions are as follows:
1. What were the student population and teaching contexts in the empir-
ical studies of integrating GenAI into L2 writing classes?
2. In what ways did instructors incorporate GenAI tools into their L2
writing classes? What AI-assisted pedagogical strategies were applied
to teach L2 writing?
3. What were the learning focuses that guide AI-assisted teaching prac-
tices in L2 writing contexts?
4. How did researchers measure AI-assisted L2 writing instruction, and
what were the learning outcomes and perceptions of L2 students after
engaging with AI-assisted L2 writing instruction?

3. Method
3.1. Data collection
Data collection was conducted in May 2024 by using educational databases
including Google Scholar, ProQuest, Scopus, Science Direct, Web of Science,
and JSTOR. To align with the research questions, we used a set of key-
words as search terms, such as “GenAI”, “AI”, “L2 writing”, “education”,
and etc. The search results were evaluated by the research team based on
the inclusion criteria. Eligible papers should meet the inclusion criteria: 1)
empirical studies, 2) focus on the application of GenAI in writing education,
3) implemented in L2 contexts, 4) written in English and full-text available.
We considered both published and unpublished studies to reduce publication

4
 Table 1: Student population and contexts.
Parent code Child code
Sample size
Student original country
First language
Population information Age
Proficiency
Academic levels
Target language
Duration
Implementation country
Context information Class mode
GenAI tools
Writing task genres

bias [30]. Participants cover a range of age groups from K-12 to graduate stu-
dents. After an examination with the research team, twenty one studies met
the inclusion criteria, including 17 published journal articles, 2 proceeding
articles, and 2 articles from conference books (see Data availability).

5
 Table 2: Pedagogical application of GenAI in L2 writing.
Code Explanation
Evaluation and feedback GenAI was used as a resource for students to
receive feedback, and conduct automatic as-
sessments on their drafts.
Brainstorming GenAI was used to help students brainstorm
initial writing ideas and outlines.
GenAI was used for students to:
1. Predict topics;
Predicting 2. Continue writing after the given partial
sentences/paragraphs/drafts;
3. Continue developing ideas.
Revision Instructors asked students to revise and edit
AI-generated information.
GenAI was used for students to:
1. Generate writing samples as models for
Modeling analysis and evaluation;
2. Generate writing samples and compare
them with human writing.
Summarizing GenAI was used for students to summarize
long or difficult texts.
GenAI was used for students to:
Content generation 1. Generate longer texts or first drafts;
2. Produce images for multimodal writing.
GenAI was used for students to:
1. Scaffolding the writing process;
Intellectual writing support
2. Analyze data;
3. Search for more resources.
GenAI was used for teachers to:
1. Generate writing prompts, and other
Teaching support teaching materials;
2. Conduct learning analytics;
3. Identify learning patterns.

6
3.2. Data analysis
To answer the research questions, we developed a coding scheme focusing
on four dimensions of the studies: 1) student population and contexts (see
Table 1); 2) pedagogical application of GenAI in L2 writing (see Table 2); 3)
learning focuses and guiding concepts in AI-assisted classes (see Table 3); 4)
measurements and learning outcomes (see Table 4). The first dimension was
designed to answer the first research question about student population and
contexts. It included sample size, student original country, first language, age
groups, proficiency, academic levels, target language, duration, implementa-
tion country, class mode, GenAI tools, and writing task genres. Following
Wang et al. [18], the second dimension explored how GenAI tools were ap-
plied in L2 writing classes. We categorized the applications into evaluation
and feedback, brainstorming, predicting, revision, modeling, summarizing,
content generation, intellectual writing support, and teaching support. The
third dimension was informed by Hyland [31] and Long & Magerko [32] on
learning focuses and guiding concepts in AI-assisted L2 writing classes. It
covered language structure, text functions, creative expression, writing pro-
cess, content knowledge, genre and contexts of writing, GenAI literacy, and
critical thinking. The fourth dimension investigated measurements and learn-
ing outcomes, including several measurement methods, writing performances,
and students’ affective responses.
Thematic analysis [33] was conducted to identify the main themes in
writing performances and affective responses. Specifically, we found that the
emerging themes can be categorized into linguistic accuracy, genre conven-
tions, content, and structure. Regarding students’ affective responses, we
considered the primary emerging themes, such as writing support, useful-
ness, motivation, confidence, GenAI features, reliance on AI, creativity, and
critical thinking.

7
 Table 3: Learning focuses and guiding concepts in AI-assisted classes.
Code Explanation
Language structure The implementation focused on essential lin-
guistic knowledge and vocabulary choices.
Text functions The implementation focused on relating
structures to meanings and using language
to express target functions.
Creative expression The implementation focused on encouraging
students to express their personal experi-
ences and opinions. Writing was considered
a creative act of self-discovery.
Writing process The implementation focused on developing
students’ metacognitive awareness of their
writing process, helping them reflect on the
strategies they use to write.
Content knowledge The implementation involved subject mat-
ter themes and topics, such as pollution and
health.
Genre and contexts of writing The implementation focused on genre
knowledge, including awareness of audience,
adaptation to contexts, and clarity of pur-
poses to successfully meet the expectations
of specific text types.
The implementation involved:
1. Play around and get familiar with tools;
2. Understand the strengths and
GenAI literacy
weaknesses of GenAI tools;
3. Define prompts, learn prompt
engineering.
Critical thinking The implementation included developing
students’ critical thinking abilities.

8
 Table 4: Measurements and learning outcomes.
Parent code Child code
Surveys
Interviews
Students’ productions and artifacts:
1. Implementation vs. Control
Measurements
2. Diagnostic writing tests
Journals, think-aloud, stimulated recalls
Observation
Student reflection
Linguistic accuracy
Genre conventions
Writing performances
Content
Structure
Duration
Implementation country
Context information Class mode
GenAI tools
Writing task genres
Writing support
Writing support
Usefulness
Motivation
Confidence
GenAI features:
Affective responses 1. Quick response
2. Ease of use
3. Convenience
4. User-friendly
Reliance on AI
Creativity
Critical thinking

9
 Figure 1: Countries and the number of studies.

4. Results
4.1. Results for RQ1: student population and teaching contexts
To answer RQ1, this section introduces student demographics and teach-
ing contexts, including duration, technologies, and writing tasks of the 21
empirical studies. On average, the included empirical studies of AI-assisted
L2 writing instruction had a mean student sample size of 31 but with a large
variance (SD = 27.29), which suggests the sample sizes vary widely from
each other. Eleven studies were small-scale exploratory research, with en-
rollment of no more than 30 participants. Two studies did not report the
number of participants. The remaining studies included fewer than 100 stu-
dent participants. Regarding student demographics, as shown in Table 5,
more studies aimed at teaching advanced adult learners, who came from di-
verse L1 backgrounds. Notably, four studies (19%) did not report students’
L1 background, although it may be related to the country where the imple-
mentation was conducted. Ten studies (48%) did not disclose the students’
language proficiency. Twelve studies (57%) did not provide information on
their students’ ages.

10
 Table 5: Student demographics.
Theme Code Count Percentage
Low 1 5
Low-intermediate 2 10
Intermediate 1 5
Proficiency
High-intermediate 2 10
Advanced 5 24
Not reported 10 48
Mandarin 3 14
Cantonese 2 10
German 1 5
Indonesian 3 14
Dutch 1 5
First language (L1) Korean 1 5
Japanese 1 5
Arabic 2 10
English 1 5
Mixed 2 10
Not reported 4 19
>=18 6 29
13-17 3 14
Age
<=12 0 0
Not reported 12 57

Despite the small number of empirical studies in this emerging AI-assisted

instruction field, research has been conducted across 13 countries in Europe,
Asia, the Middle East, Oceania, and North America (see Fig. 1). Asia
leads with the highest number of studies (n = 13, 62%) with Hong Kong
and Indonesia each conducting four studies (19%). Europe follows, with four
studies (19%) conducted across four different countries. Two studies were
from the Middle East countries. North America and Oceania each had one
study in this domain. This distribution reflected a notable concentration of
interest and research activity in AI-assisted L2 writing instruction in Asia,
alongside a growing global engagement with this educational technology.

11
 Table 6: Teaching contexts.
Theme Code Count Percentage
Elementary school 0 0
Middle school 2 10
High school 4 19
Institution
Undergraduate 12 57
Graduate and post-graduate 2 10
Language learning center 1 5
Foreign language 19 90
Language specialty
Second language 2 10
English 18 86
Target language Chinese 1 5
German 2 10
Less than 1 week 2 10
2-4 weeks 7 33
Duration 5-8 weeks 3 14
More than 8 weeks 2 10
Not reported 7 33
ChatGPT 21 100
GenAI tools
Mixed 4 19

In terms of more details on teaching contexts, the institutions examined

span from middle school to graduate levels, most (n=12, 57%) were situated
in the undergraduate programs, while K-12 setting came in second with 4
studies (19%) in high school context and two studies in middle school context
(10%) (see Table 6). Most studies (n=19, 90%) were conducted in foreign
language learning settings where English was the target language for learning
(n=18, 86%). As to the duration of the studies, seven studies (33%) were
short-term implementations that lasted 2-4 weeks. Meanwhile, seven studies
(33%) did not report their length of teaching.
As revealed at the bottom of Table 6, we found that so far ChatGPT
served as the prevalent GenAI tool in L2 writing classrooms. Almost all
selected studies used ChatGPT as their main GenAI tool for teaching and
learning L2 writing. Four studies (19%) also combined with other GenAI
tools, such as Bing Chat [14] and Microsoft Aim writing [34].
Regarding the writing tasks that these studies focused on (see Fig. 2),

12
 Figure 2: Writing task genres and the number of studies.

Figure 3: AI-assisted pedagogical applications in L2 writing contexts.

we notice a wide range of genres, with creative story (n=4, 19%) and argu-
mentative writing (n=3, 14%) being used the most, with a learning focus on
forms and functions. We used the term “genre” here to indicate. However,
seven studies did not specify the genres of the writing tasks they had, instead
categorizing them as academic writing or L2 writing in general.

4.2. Results for RQ2: AI-assisted pedagogical applications in L2 writing con-

texts
Research question 2 examines the variety of AI-assisted pedagogical ap-
plications that were applied to L2 writing classes. The 21 studies included
comprehensive descriptions of AI-assisted language learning activities as well
as the roles GenAI performed during the implementation process to support

13
L2 students in developing their writing and writing-related skills. We cate-
gorized them into 9 categories developed based on Crompton & Burke [35]
and Wang et al. [18]. Fig. 3 reflects the analysis of AI-assisted pedagogical
applications.
The result shows that evaluation and feedback were the most frequent
application of GenAI in L2 writing classrooms (n=14, 67%). This category
includes using GenAI tools as a resource for students to evaluate their drafts
and receive feedback. For L2 student writers who face extra language difficul-
ties, numerous studies have employed GenAI as a tool to facilitate students’
self-evaluation and provide instant corrective feedback. It helped L2 students
avoid long-time waiting for teacher feedback, keeping them engaged and mo-
tivated throughout the writing process, as well as promoting self-regulated
learning [35, 36]. Regarding the content of evaluation and feedback, the ma-
jority of studies (n=14, 67%) focused on correcting linguistic errors in L2
writing [37, 12, 38, 39], with a few studies (n=2, 10%) also mentioned feed-
back on genre features and conventions using proper prompts [36, 39]. In
addition, it is worth noting that compared with traditional teacher feedback,
which often highlights language errors in texts and allows students to make
changes accordingly, GenAI usually automatically corrects and edits original
texts and presents rewritten texts without further explanation [40].
Modeling was in second place as a common application of GenAI in L2
writing classrooms (n=11, 52%). It refers to 1) generating writing samples
as models for analysis and evaluation; and 2) generating writing samples
and comparing them with human writing. Considering the L2 learning back-
ground, L2 students often face extra disadvantages in some unfamiliar cul-
tural contexts, exposure to a variety of genres can accelerate the processes of
situated acquisition [41]. Leveraging its access to an extensive corpus of data,
GenAI can generate writing samples of different genres, which can serve as
writing samples for L2 students to analyze genre and language conventions.
For example, in the study conducted by Liu et al. [14], the drafts produced by
ChatGPT have functioned as model texts to improve L2 students’ writing
skills regarding organizational structures, language and content. However,
Kim et al. [42] observed that AI-generated model texts may often provide
rigid, machine-like templates that led L2 students to passively follow these
structures and write with strict organization without independent thoughts
and varieties.
Using GenAI as intellectual writing support was identified as another
useful AI-assisted pedagogical application (n=8, 38%). Three subcodes were

14
recognized in how GenAI supported the L2 students’ writing process: scaf-
folding the writing process, analyzing data, and searching for resources. Our
extant reviews of the selected studies highlighted the use of GenAI for scaf-
folding (n=3, 14%). Scaffolding support is defined as a strategy that prompts
students to start, connect, and develop their thoughts step by step [41],
guiding them to construct their texts through ”gradual approximation” [43].
This approach helps L2 learners understand writing as a process and de-
velop greater confidence [41]. An example of this is the study by Harunasari
[44], students were assigned to write a short story where ChatGPT broke
down the whole task into smaller manageable pieces, guiding them through
developing an initial storyline, identifying key plot points, and composing
narrative effectively. Besides scaffolding guidance on the writing process, in
a study by Liu et al. [14], students interacted with Bing Chat to search for
academic references, reflecting the application of GenAI in resource searching
for academic writing.
Another code that emerged in data analysis was applying GenAI for brain-
storming (n=6, 29%). This includes brainstorming initial writing ideas and
outlines at the beginning of the writing process. Generating creative and
unique ideas can be some students’ writing blocks. The integration of Chat-
GPT can spark students’ creativity and enhance their idea generation abil-
ities [13, 44, 37]. Instead of automatically generating writing content, this
approach aimed at fostering students’ own critical thinking, planning and
creativity at the initial stages of the writing process.
Aside from the four most frequent applications discussed above, a smaller
portion of studies explored other pedagogical uses of GenAI, as illustrated in
Fig. 3. These applications included integrating GenAI for students to gen-
erate longer texts for their first drafts and to produce images for multimodal
writing projects, documented in three studies (14%). Another three stud-
ies (14%) addressed how students revised and edited AI-generated content
to improve their language awareness and writing skills. Additionally, three
studies (14%) focused on GenAI’s role in teaching preparation and manage-
ment. These studies found that GenAI can serve as a teaching assistant to
generate writing prompts, create teaching materials, analyze learning data,
and identify learning patterns. Two studies (10%) introduced that students
can use GenAI to predict and continue their writing from partial sentences,
paragraphs, or drafts. One study (5%) reported that they used GenAI as a
learning tool to help L2 students summarize long and difficult texts.

15
 Figure 4: Learning focuses in AI-assisted L2 writing teaching.

4.3. Results for RQ3: Learning focuses in AI-assisted L2 writing teaching

RQ3 explores different learning focuses that guided AI-assisted teaching
practices in L2 writing contexts. Learning focuses support teachers in decid-
ing what students need to learn and what they need to provide for an effective
AI-assisted writing class. According to Hyland [31], L2 writing teaching is
often guided by six focuses: language structures, text functions, creative
expression, composing processes, content, and genre and contexts of writ-
ing. In the context of integrating GenAI into L2 writing pedagogy, we have
added two pertinent themes: GenAI literacy and critical thinking. Accord-
ing to Long and Magerko [32], GenAI literacy refers to a set of competencies
that support us to understand, engage, and evaluate GenAI applications.
It is worth noting that teachers often draw on multiple focuses rather than
strictly adhering to one of those, but it is common for one focus to prevail
in shaping their teaching design and implementation [45, 31]. Therefore, al-
though these focuses are seldom reflected as distinct teaching approaches,
examining the diversity and disparity of learning focuses on AI-assisted L2
writing classrooms can shed light on the features of AI-assisted L2 writing
instruction and provide insights into some potential ways to further develop
our pedagogical approaches.
As revealed in Fig. 4, language structure emerged as a prominent learn-
ing focus, with 14 studies (67%) emphasizing this goal [40, 13, 37, 14, 34]. A
big portion of AI-assisted L2 writing instruction involved developing L2 stu-
dents’ linguistic knowledge and grammar competence which were the most
immediately obvious learning obstacles of L2 student writers [46, 31, 47].
GenAI literacy came as the second prevalent learning focus, appearing in 11
studies (53%), as observed in addressing the strengths and weaknesses of AI-

16
 Figure 5: Measurement methods in AI-assisted L2 writing teaching.

generated texts and helping L2 students enhance their prompt engineering

skills [13, 48, 39, 49, 34]. Writing process was another prevalent learning
focus (n = 10, 48%) which fostered a process-based writing approach to em-
phasize developing students’ awareness and abilities to review, revise, and
reflect in the process of writing [10, 12, 50]. Text functions were seen in
8 studies (38%). They focused on teaching structures and relating struc-
tures to meanings. Some relevant learning activities included using GenAI
to generate an outline [44, 42, 14] and comment on paragraph structure to
better scaffold L2 students in writing [10]. Seven studies (33%) focused on
genre and contexts of writing, drawing L2 students’ attention to awareness
of audience, adaptation to contexts, and clarity of purposes [14, 38, 50, 51].
These areas might be the most deficiencies of L2 students due to cultural
differences in rhetoric and the unfamiliarity of rhetorical conventions in their
second language [52, 31, 53, 54]. Additionally, both critical thinking and
content knowledge have been the learning focus in 4 studies (19%). Creative
expression was addressed in 3 studies (14%). Four studies (19%) did not
present specific learning focuses or objectives in their AI-assisted L2 writing
instruction.

4.4. Results for RQ4: Measurements, learning outcomes and perceptions

To address RQ4, we noticed six types of measurement methods in the col-
lected articles: student artifact evaluation, surveys, observations, interviews,
student reflection, and journal reports (see Fig. 5). These measurements
were used to analyze two dimensions of the effects of AI-assisted L2 writ-
ing instruction: (1) students’ writing performances, such as linguistic and
genre-specific features; and (2) students’ affective responses to AI-assisted
L2 writing instruction, including perceptions and feedback.

17
 For the first dimension, most of the studies (n = 14, 67%) included a
mixed analysis of students’ writing performances through collected student
productions and artifacts [40, 55, 51]. Among these studies, eleven studies
(52%) applied a pre-test and post-test design to assess students’ writing
performances [44, 37, 48, 38], and 3 studies (14%) compared the drafts from
implementation and control groups [10, 13, 39]. To understand more about
the effects of GenAI on students’ writing performances, we delved deep into
their findings in terms of artifact analysis. The results widely acknowledged
that GenAI was effective in correcting linguistic errors and thus improving
the accuracy of student writing [13, 49]. Some studies also addressed that
GenAI helped L2 students notice and adopt appropriate styles and registers
in their language use, such as converting informal expressions into formal
academic language in scientific writing genres [50]. However, when it came
to the use of examples, results became inconsistent in the collected studies.
Some studies highlighted that GenAI can provide a variety of examples during
writing process [14, 34], however, the provided examples might not always be
concrete, specific, or relevant enough to support the claims effectively [13, 14].
As many studies were focused on teaching academic writing, citations were
often one of the major elements evaluated in student texts. The results of
citations tended to be negative, as they can be missing or inaccurate when
students engaged with GenAI tools. Studies also revealed that the structure
of students’ essays and paragraphs did not significantly improve and even
declined in some cases [13].
Regarding examining students’ affective responses to AI-assisted L2 writ-
ing instruction (the second dimension), self-reported evaluation was often
used with surveys as the most frequent data collection instrument (n = 13,
62%). Other instruments included observation (n = 6, 29%), interviews (n
= 6, 29%), student reflection (n = 3, 14%), and journals (n = 2, 10%). As
for perceptions of L2 students after engaging with AI-assisted L2 writing
instruction, many students across different studies affirmed the effectiveness
of using GenAI in lexicogrammatical revision, which greatly alleviated their
anxiety and boosted their confidence and motivation in English writing as
foreign language learners [37, 55, 36, 39]. Some students expressed a sense
of empowerment in their second language acquisition attributed to GenAI’s
feature of quick response, ease of use, convenience [56, 48, 12, 39]. Addition-
ally, the ease of understanding the responses was highlighted as user-friendly
for L2 learners [10, 14].
The most frequent concern that emerged from student responses and ob-

18
servations was overreliance on GenAI. With the adoption of GenAI in teach-
ing and learning, a number of students became excessively dependent on it,
which hindered their brainstorming ability, critical thinking and creativity
[13, 55, 51]. Since the default tone of ChatGPT was usually White Main-
stream English [6], L2 students in the study of Song & Song [39] realized
that the suggestions and revised writing did not always align with their own
writing style or were not contextually accurate. Moreover, the individuality,
human nuances, and authenticity in writing were reported reduced or elimi-
nated after AI revision, resulting in writing that tended to be homogeneous
and standardized, lacking stances, originality, and creativity [10, 50].

5. Discussion
In the following sections, we highlight the main findings and trends in the
collected studies, discuss pedagogical implications on both the design and
implementation of AI-assisted L2 writing instruction, and explore potential
gaps and opportunities for future research in this area.

5.1. Contexts of studies

The systematic review of the 21 studies highlights a prevalent focus on AI-
assisted L2 writing instruction for adult English language learners in foreign
language contexts (EFL). Together with the development of GenAI, this focus
raises concerns about the underrepresentation of K-12 students, ESL learners,
and learners of languages other than English. This phenomenon calls for
future empirical research to explore a wider range of student demographics
in various contexts, including non-English learners and countries where the
target language is the dominant language, to address the needs of L2 writing
educators who work with diverse student populations globally.
Over half of the collected studies had less than 30 participants. The imple-
mentation was frequently in small-scale short-term contexts. More empirical
research is needed to explore the design and implications of AI-assisted L2
writing instruction with a larger participant group and a longer duration.
The predominance of using ChatGPT as their GenAI tool may be at-
tributed to its ease of use and stability across education contexts [57]. Fewer
studies have tested other GenAI tools or combined ChatGPT with other dig-
ital tools [14, 34]. As large language models continue to evolve, we expect
more opportunities to explore and compare various tools in future educa-
tional practices and research.

19
5.2. AI-assisted pedagogical approaches
With regard to how teachers applied GenAI to assist their L2 writing
classes, we identified key pedagogical approaches emerging from the studies,
including evaluation and feedback, modeling, intellectual writing support,
brainstorming, etc. A salient pattern from the findings was the prevalent
use of GenAI to address linguistic errors, which was an efficient and reli-
able approach for L2 students who need extra language support for writing.
Beyond linguistic errors, GenAI was also useful in generating sample texts
for L2 students to observe and identify genre conventions [14, 55, 38]. For
L2 students who are unfamiliar with naturalistic language settings, learning
about genres can be a foundational approach to engaging with cultural and
textual practices [41]. Furthermore, GenAI was integrated to scaffold the
writing process by breaking large projects into smaller pieces and providing
ideas to inspire students’ thoughts.
While we learned various approaches to integrate GenAI in L2 writing
education, it is important to highlight that GenAI often automatically edits
original texts [40], allowing L2 students to directly copy and paste the re-
vised text into their drafts. This should bring our attention to the question
of whether it merely improves a single essay per se or genuinely enhances stu-
dents’ writing skills. Further teaching practices and research are needed to
evaluate students’ authentic self-writing skills without GenAI after engaging
with AI-assisted instruction.
In addition, we argue that it is crucial to emphasize that the role of GenAI
should be a writing assistant offering support for consideration rather than
an alternative writing editor for students or an alternative teacher imparting
writing knowledge. The findings in the studies present that texts generated
by GenAI may contain fake citations and rigid, machine-like templates, which
could be misleading and confusing for L2 student writers. Thus, it is essential
for L2 writing teachers to explicitly discuss these risks and increase students’
awareness of GenAI. Specifically, teachers can emphasize the assistant role
of GenAI and remind students not to overrely on GenAI. Students should
be cautious not to accept all AI-generated content as accurate and should
avoid treating AI-generated texts as standards or requirements. Exemplary
designs can be found in some studies, such as Dwina [58], Harunasari [44],
and Strobl et al. [50]. They had students analyze and compare AI-generated
texts with human writing with prompts guiding students to assess content.
This approach can promote students’ abilities of critical thinking, evaluation,

20
fact-checking, and verifying information when using GenAI in L2 writing
contexts.

5.3. Learning content in GenAI-based class

In terms of learning content in AI-assisted L2 writing instruction, stud-
ies in this domain widely focused on using GenAI to improve L2 students’
language structure, which signaled its effectiveness in addressing the founda-
tional aspect of L2 writing [40, 13, 37, 14, 34]. The second salient trait was
developing students’ GenAI literacy, including understanding GenAI capabil-
ities, delving into ethical considerations, and enhancing prompt engineering
skills [13, 48, 39, 49, 34]. It indicated that building GenAI awareness and
literacy has been listed as one of the learning objectives in some classes and
tended to become a core component of AI-assisted L2 writing instruction.
This pattern resonates with the discussion in Chiu and Sanusi [59] and Stolpe
and Hallström [60], emphasizing cultivating GenAI literacy across broader
educational domains and preparing students for a future where GenAI per-
meates everyday and professional environments.
While it is important to explore the new technology, we suggest that the
integration of GenAI in writing education should align with the learning ob-
jectives and curriculum content, recognize its pedagogical affordances, and
ensure coherent and responsive courses. As a kind of digital tool, TPACK
(Technological Pedagogical Content Knowledge) framework is also applicable
to the use of GenAI and relevant teacher training. Mishra et al., [61] claim
that “TPACK focuses on the integration of tools with content and pedagogy,
not on the specific tools themselves” (p.239). In order to use GenAI effec-
tively, teachers need to be equipped with the skills to integrate these tools
with pedagogy knowledge (PK)—the understanding of teaching strategies
and management—and content knowledge (CK)—the understanding of con-
tent and learning objectives to be taught. In other words, this integration
should go beyond only using GenAI for innovation per se; it requires an in-
tentional focus on the pedagogical and content objectives that the technology
is meant to support [62, 63, 18]. For example, in the context of L2 writing,
teachers may ask students to observe and analyze the language patterns used
in AI-generated texts with prompts leading their attention to authorial voice
and audience awareness. This allows students to concentrate on the content
learning of the rhetorical situation and writing techniques. As this example
indicates, responsive teacher training is essential to ensure that instructors

21
can strategically and purposefully leverage GenAI to enhance students’ writ-
ing skills.

5.4. Measurement and outcomes

The results of this systematic review reveal a wide range of measurement
methods used in research on AI-assisted L2 writing instruction, including
qualitative and quantitative analysis of student productions and artifacts,
surveys, observations, interviews, reflections, and journals. Student work
revealed that GenAI was effective in improving the accuracy of student writ-
ing as well as raising their genre awareness. However, noteworthy is the
fact that in some cases, the structure of students’ essays and paragraphs
did not evidently improve and even declined after engaging with GenAI.
This might be due to some students submitting drafts generated exclusively
by ChatGPT with minimal self input [13]. AI-generated writing often lacked
concrete examples and citations [13] and tended to share a uniform template-
like structure which made students’ writing lack effectiveness and creativity
[49]. Additionally, since GenAI often adopts White Mainstream English in
its responses, using it to assist L2 writing education may reinforce the idea
of standard English. This approach can assimilate L2 student writing and
exacerbate hegemonic biases against L2 students [6]. These findings suggest
greater teacher mediation and supervision on the use of GenAI in AI-assisted
writing practices. In teaching practices, teachers can encourage L2 students
to value linguistic diversity, understand diverse linguistic structures, and crit-
ically engage with AI-generated texts.
Analyzing students’ perceptions and experiences, we found that GenAI
can increase L2 students’ confidence and reduce anxiety in writing tasks,
which suggested that GenAI may be more effective for facilitating the initial
stages of language acquisition, particularly in lexicogrammatical revision.
However, dependence on GenAI raises concerns about the development of
critical thinking and creative writing skills. For L2 students, GenAI may
pose bias and language assimilation for L2 students. According to Hawanti
and Zubaydulloevna [37], ChatGPT learns from text data found on the inter-
net, which can include biases from biased or offensive sources. This may lead
the model to show cultural or regional biases toward certain norms or lan-
guages and therefore assimilate L2 students in writing practices. The results
from students’ perceptions and experiences once again emphasize the impor-
tance of teacher supervision and mediation on GenAI in L2 writing classes.
Effective strategies may include establishing clear policies and guidelines for

22
GenAI use and requiring students to submit their chat history of GenAI,
which could help manage if it is used effectively and ethically [44, 18]. More
research is needed to continue to explore the broader implications of GenAI
integration from the perspective of instructors who are implementing AI-
assisted activities in L2 writing courses.

6. Conclusion
By examining 21 studies that integrated GenAI in L2 writing classrooms
across various language proficiencies and countries, the study mapped out
the current student demographics, pedagogical applications, learning focuses,
measurements, and learning outcomes on AI-assisted L2 writing courses. The
findings showed that a primary focus on integrating ChatGPT for adult En-
glish language learners in EFL contexts, reflecting the need for attention
to more diverse student groups and broader educational contexts. The re-
viewed studies presented diverse pedagogical applications of GenAI in lan-
guage learning, with instant corrective feedback for linguistic errors emerg-
ing as the most prevalent application, alongside other applications such as
evaluation, modeling, intellectual writing support, and brainstorming. This
highlighted GenAI’s effectiveness in assisting L2 students who require addi-
tional linguistic support. In AI-assisted L2 writing classrooms, the learning
objectives also included exploring the new technology and cultivating GenAI
awareness and literacy. Understanding strengths and weaknesses of GenAI
tools as well as prompt engineering skills would help students learn how to
interact with GenAI tools to enhance their learning without limiting their
creative thinking. However, the review also found important limitations of
integrating GenAI in L2 writing education. From the reviewed studies, we
noticed that automatic edits from GenAI may contain fake and unnatural
content, which could be misleading and confusing for L2 students. Addi-
tionally, while qualitative and quantitative analysis revealed that GenAI was
generally effective in improving student writing and boosting L2 students’
confidence, a few reviewed studies reported that there were student works
that did not improve or even declined. This phenomenon might be attributed
to the excessive reliance on GenAI tools and the biases and language assim-
ilation that these tools may present to L2 students. These findings reflected
the importance of developing students’ critical thinking skills to evaluate AI-
generated content and cultivating students’ awareness to avoid over-reliance.
The findings of this study carried pedagogical and research implications

23
for the future integration of GenAI in L2 writing courses. For teaching prac-
tices, this systematic review summarizes a variety of pedagogical approaches
and learning focuses that could serve as valuable resources for teacher scholar
who are interested in implementing AI-assisted L2 writing classes. In addi-
tion, the study emphasizes the important role of teacher training on this new
technology, advocating for frameworks such as the TPACK model to ensure
that GenAI integration aligns with specific learning objectives and course
content. Moreover, the findings also address the importance of teacher su-
pervision and mediation in AI-assisted L2 writing environments to guide
students toward the ethical and effective use of GenAI tools. Specifically,
teachers’ responsibilities in this new era may extend to raising L2 students’
awareness about potential pitfalls of AI-generated content, such as misinfor-
mation, cultural bias, and language assimilation. Establishing clear policies
and guidelines may also become essential to regulate students’ GenAI usage,
and thereby building a responsible and effective L2 writing classroom.
For a research-oriented perspective, this study contributed to a compre-
hensive overview of GenAI’s role in L2 writing education. As this technology
rapidly evolves and becomes more widely used, more empirical studies are
needed to explore its impact across diverse student demographics and edu-
cational contexts. For example, future studies could explore the engagement
and benefits of AI-assisted writing instruction for learners of different lan-
guage proficiencies and L1 backgrounds. Additionally, more research should
further explore student performance in AI-assisted L2 writing classes and
teachers’ role in scaffolding these activities. For instance, students’ authentic
independent writing skills without GenAI after participating in AI-assisted
instruction might be a worthy area of investigation, as it could provide in-
sights into the transferability of skills acquired through AI-assisted instruc-
tion. Finally, while most of the reviewed studies implemented GenAI over
short periods (within two months), it would be beneficial to examine its
long-term effects on L2 students’ writing skills and language development.

Declarations
Declaration of Competing Interest: The authors declare no financial
or personal relationships that could inappropriately influence their work.
Acknowledgements: Not applicable. This research did not receive any
specific grant from funding agencies in the public, commercial, or not for
profit sectors.

24
References
[1] M. Abdullayeva, Z. M. Musayeva, The impact of chat gpt on student’s
writing skills: An exploration of ai-assisted writing tools, in: Interna-
tional Conference of Education, Research and Innovation, Vol. 1, 2023,
pp. 61–66.

[2] K. Fuchs, Exploring the opportunities and challenges of nlp models in

higher education: is chat gpt a blessing or a curse?, in: Frontiers in
Education, Vol. 8, Frontiers Media SA, 2023, p. 1166682.

[3] J. Rudolph, S. Tan, S. Tan, Chatgpt: Bullshit spewer or the end of

traditional assessments in higher education?, Journal of applied learning
and teaching 6 (1) (2023) 342–363.

[4] M. Sullivan, A. Kelly, P. McLaughlan, Chatgpt in higher education:

Considerations for academic integrity and student learning (2023).

[5] M. M. Rahman, Y. Watanobe, Chatgpt for education and research: Op-

portunities, threats, and strategies, Applied Sciences 13 (9) (2023) 5783.

[6] E. M. Bender, T. Gebru, A. McMillan-Major, S. Shmitchell, On the

dangers of stochastic parrots: Can language models be too big?, in:
Proceedings of the 2021 ACM conference on fairness, accountability,
and transparency, 2021, pp. 610–623.

[7] C. K. Y. Chan, W. Hu, Students’ voices on generative ai: Perceptions,

benefits, and challenges in higher education, International Journal of
Educational Technology in Higher Education 20 (1) (2023) 43.

[8] G.-G. Lee, L. Shi, E. Latif, Y. Gao, A. Bewersdorf, M. Nyaaba, S. Guo,

Z. Wu, Z. Liu, H. Wang, et al., Multimodality of ai for education:
Towards artificial general intelligence, arXiv preprint arXiv:2312.06037
(2023).

[9] A. Barrett, A. Pack, Not quite eye to ai: student and teacher perspec-
tives on the use of generative artificial intelligence in the writing process,
International Journal of Educational Technology in Higher Education
20 (1) (2023) 59.

25
[10] J. Escalante, A. Pack, A. Barrett, Ai-generated feedback on writing:
insights into efficacy and enl student preference, International Journal
of Educational Technology in Higher Education 20 (1) (2023) 57.

[11] A. McGuire, W. Qureshi, M. Saad, A constructivist model for leverag-

ing genai tools for individualized, peer-simulated feedback on student
writing., International Journal of Technology in Education 7 (2) (2024)
326–352.

[12] R. Schmidt-Fajlik, Chatgpt as a grammar checker for japanese english

language learners: A comparison with grammarly and prowritingaid,
AsiaCALL Online Journal 14 (1) (2023) 105–119.

[13] K. Glahn, Using chatgpt to teach english for academic purposes: Per-
spectives from the classroom, Journal of Education, Innovation and
Communication 5 (2) (2024) 63–75.

[14] M. Liu, L. J. Zhang, C. Biebricher, Investigating students’ cognitive

processes in generative ai-assisted digital multimodal composing and
traditional writing, Computers & Education 211 (2024) 104977.

[15] L. Law, Application of generative artificial intelligence (genai) in lan-

guage teaching and learning: A scoping literature review, Computers
and Education Open (2024) 100174.

[16] H. Li, Y. Wang, S. Luo, C. Huang, The influence of genai on the effec-
tiveness of argumentative writing in higher education: evidence from a
quasi-experimental study in china, Journal of Asian Public Policy (2024)
1–26.

[17] A. Dang, H. Wang, Ethical use of generative ai for writing practices: Ad-
dressing linguistically diverse students in u.s. universities’ ai statements,
Journal of Second Language Writing (2024).

[18] H. Wang, A. Dang, Z. Wu, S. Mac, Seeing chatgpt through universi-

ties’ policies, resources and guidelines, arXiv preprint arXiv:2312.05235
(2023).

[19] J. Luo, A critical review of genai policies in higher education assessment:

A call to reconsider the “originality” of students’ work, Assessment &
Evaluation in Higher Education (2024) 1–14.

26
[20] M. A. Yeo, Academic integrity in the age of artificial intelligence (ai)
authoring apps, Tesol Journal 14 (3) (2023) e716.

[21] A. Yusuf, N. Pervin, M. Román-González, Generative ai and the fu-

ture of higher education: a threat to academic integrity or reformation?
evidence from multicultural perspectives, International Journal of Edu-
cational Technology in Higher Education 21 (1) (2024) 21.

[22] W. Alharbi, Ai in the foreign language classroom: A pedagogical

overview of automated writing assistance tools, Education Research In-
ternational 2023 (1) (2023) 4253331.

[23] C. K. Y. Chan, K. K. Lee, The ai generation gap: Are gen z students

more interested in adopting generative ai such as chatgpt in teaching
and learning than their gen x and millennial generation teachers?, Smart
learning environments 10 (1) (2023) 60.

[24] J. Campino, Unleashing the transformers: Nlp models detect ai writing

in education, Journal of Computers in Education (2024) 1–29.

[25] W. Liang, M. Yuksekgonul, Y. Mao, E. Wu, J. Zou, Gpt detectors are

biased against non-native english writers, Patterns 4 (7) (2023).

[26] J. Dempere, K. Modugu, A. Hesham, L. K. Ramasamy, The impact of

chatgpt on higher education, in: Frontiers in Education, Vol. 8, Frontiers
Media SA, 2023, p. 1206936.

[27] L. Li, Y. Fengchao, E. Zhang, A systematic review of learning task design

for k-12 ai education: Trends, challenges, and opportunities, Computers
and Education: Artificial Intelligence (2024) 100217.

[28] L. Labadze, M. Grigolia, L. Machaidze, Role of ai chatbots in educa-

tion: systematic literature review, International Journal of Educational
Technology in Higher Education 20 (1) (2023) 56.

[29] H. Nguyen Thi Thu, Efl teachers’ perspectives toward the use of chat-
gpt in writing classes: A case study at van lang university, Nguyen,
TTH (2023). EFL Teachers’ Perspectives toward the Use of ChatGPT
in Writing Classes: A Case Study at Van Lang University. International
Journal of Language Instruction 2 (3) (2023) 1–47.

27
[30] R. Rosenthal, The file drawer problem and tolerance for null results.,
Psychological bulletin 86 (3) (1979) 638.
[31] K. Hyland, Second language writing, Cambridge university press, 2019.
[32] D. Long, B. Magerko, What is ai literacy? competencies and design
considerations, in: Proceedings of the 2020 CHI conference on human
factors in computing systems, 2020, pp. 1–16.
[33] V. Braun, V. Clarke, Using thematic analysis in psychology, Qualitative
research in psychology 3 (2) (2006) 77–101.
[34] D. Yan, Impact of chatgpt on learners in a l2 writing practicum: An ex-
ploratory investigation, Education and Information Technologies 28 (11)
(2023) 13943–13967.
[35] H. Crompton, D. Burke, Artificial intelligence in higher education: the
state of the field, International Journal of Educational Technology in
Higher Education 20 (1) (2023) 22.
[36] L. M. Silitonga, S. Hawanti, F. Aziez, M. Furqon, D. S. M. Zain, S. An-
jarani, T.-T. Wu, The impact of ai chatbot-based learning on students’
motivation in english writing classroom, in: International Conference on
Innovative Technologies and Learning, Springer, 2023, pp. 542–549.
[37] S. Hawanti, K. M. Zubaydulloevna, Ai chatbot-based learning: allevi-
ating students’ anxiety in english writing classroom, Bulletin of Social
Informatics Theory and Application 7 (2) (2023) 182–192.
[38] S. Shaikh, S. Y. Yayilgan, B. Klimova, M. Pikhart, Assessing the usabil-
ity of chatgpt for formal english language learning, European Journal of
Investigation in Health, Psychology and Education 13 (9) (2023) 1937–
1960.
[39] C. Song, Y. Song, Enhancing academic writing skills and motivation:
assessing the efficacy of chatgpt in ai-assisted language learning for efl
students, Frontiers in Psychology 14 (2023) 1260843.
[40] S. Athanassopoulos, P. Manoli, M. Gouvi, K. Lavidas, V. Komis, The
use of chatgpt as a learning tool to improve foreign language writing in a
multilingual and multicultural classroom, Advances in Mobile Learning
Educational Research 3 (2) (2023) 818–824.

28
[41] K. Hyland, Genre pedagogy: Language, literacy and l2 writing instruc-
tion, Journal of second language writing 16 (3) (2007) 148–164.

[42] S. Kim, J. Shim, J. Shim, et al., A study on the utilization of ope-

nai chatgpt as a second language learning tool, Journal of Multimedia
Information System 10 (1) (2023) 79–88.

[43] H. G. Widdowson, Teaching language as communication, Oxford Uni-

versity Press, 1978.

[44] S. Y. Harunasari, Examining the effectiveness of ai-integrated approach

in efl writing: A case of chatgpt, International Journal of Progressive
Sciences and Technology (IJPSAT) 39 (2) (2023) 357–368.

[45] A. Cumming, Experienced esl/efl writing instructors’ conceptualizations

of their teaching: Curriculum options and implications, Exploring the
dynamics of second language writing (2003) 71–92.

[46] E. Hinkel, Teaching academic ESL writing: Practical techniques in vo-

cabulary and grammar, Routledge, 2003.

[47] A. C. Purves, Writing across languages and cultures: Issues in con-

trastive rhetoric, (No Title) (1988).

[48] X. Li, B. Li, S.-J. Cho, Empowering chinese language learners from
low-income families to improve their chinese writing with chatgpt’s as-
sistance afterschool, Languages 8 (4) (2023) 238.

[49] D. J. Woo, Evaluating a learning design for efl writing using chatgpt.

[50] C. Strobl, I. Menke-Bazhutkina, N. Abel, M. Michel, Adopting chat-

gpt as a writing buddy in the advanced l2 writing class, Technology in
Language Teaching & Learning 6 (1) (2024) 1–19.

[51] D. J. Woo, K. Guo, S. Z. Salas-Pilco, Writing creative stories with ai:

learning designs for secondary school students, Innovation in Language
Learning and Teaching (2024) 1–13.

[52] U. Connor, Contrastive rhetoric, Cross cultural aspects of second lan-

guage writing/Cambridge University (1996).

29
 [53] Y. Kachru, Cultural meaning and contrastive rhetoric in english educa-
tion, World Englishes 16 (3) (1997) 337–350.
[54] R. B. Kaplan, Cultural thought patterns in inter-cultural education,
Language learning 16 (1-2) (1966) 1–20.
[55] S. Mahapatra, Impact of chatgpt on esl students’ academic writing
skills: a mixed methods intervention study, Smart Learning Environ-
ments 11 (1) (2024) 9.
[56] M. A. Ahmed, Chatgpt and the efl classroom: Supplement or substitute
in saudi arabia’s eastern region, Information Sciences Letters 12 (7)
(2023) 2727–2734.
[57] N. Saif, S. U. Khan, I. Shaheen, F. A. ALotaibi, M. M. Alnfiai, M. Arif,
Chat-gpt; validating technology acceptance model (tam) in education
sector via ubiquitous learning mechanism, Computers in Human Behav-
ior 154 (2024) 108097.
[58] L. Dwina, Chatgpt: Empowering self-directed writing through mind
mapping and ai-assisted composition, Nusantara Science and Technol-
ogy Proceedings (2024) 33–45.
[59] T. K. Chiu, I. T. Sanusi, Define, foster, and assess student and teacher
ai literacy and competency for all: Current status and future research
direction (2024).
[60] K. Stolpe, J. Hallström, Artificial intelligence literacy for technology
education, Computers and Education Open 6 (2024) 100159.
[61] P. Mishra, M. Warr, R. Islam, Tpack in the age of chatgpt and generative
ai, Journal of Digital Learning in Teacher Education 39 (4) (2023) 235–
251.
[62] S. Leahy, P. Mishra, Tpack and the cambrian explosion of ai, in: Soci-
ety for Information Technology & Teacher Education International Con-
ference, Association for the Advancement of Computing in Education
(AACE), 2023, pp. 2465–2469.
[63] P. Perera, Preparing to revolutionize education with the multi-model
genai tool google gemini? a journey towards effective policy making
(2023).

30

View publication stats