International Journal of TESOL Studies (2025)

250522, 1-21 https://doi.org/10.58304/ijts.250522

Article

Critical Thinking in the Age of Generative AI: Effects of a Short-Term

Experiential Learning Intervention on EFL Learners

Ngo Cong-Lem*
Thang Tat Nguyen
Khanh Nhat Hoang Nguyen
Dalat University, Vietnam

Received: 20 March, 2025/Received in revised form: 10 May, 2025/Accepted: 16 May, 2025/

Available online: 28 May, 2025

Abstract
The integration of generative AI tools such as ChatGPT has transformed English as a Foreign
Language (EFL) education, offering new opportunities for supporting writing, research, and critical
inquiry. However, unguided use of AI may foster cognitive passivity and over-reliance, highlighting
the need for targeted pedagogical interventions. Grounded in experiential learning theory, this quasi-
experimental study, employing a pretest–posttest control group design, evaluated the effectiveness
of a 90-minute workshop, the Critical AI Engagement Cycle, in enhancing EFL learners’ critical
thinking skills when using ChatGPT. Despite the short duration, the workshop included multiple
scaffolded activities designed to stimulate immediate critical reflection. Seventy-two undergraduate
and graduate students at a Vietnamese public university participated, with 38 assigned to the
experimental group and 34 to the control group. Participants were selected using convenience
sampling based on course enrollment and availability. Pre- and post-test results demonstrated
statistically significant improvements in overall critical thinking and each of the four subdomains—
analytical skills, logical reasoning, evidence evaluation, and open-mindedness—among participants
in the experimental group. Notably, the consistent and large effect sizes across all critical thinking
subdomains (Cohen’s d = 0.94 to 1.23) underscore the robust impact of the intervention. The
experimental group significantly outperformed the control group in post-intervention critical
thinking scores, even after controlling for pretest scores, gender, prior AI knowledge, and AI skill
level, as confirmed by ANCOVA analyses. The results suggest that even brief, theoretically grounded
interventions can significantly enhance critical thinking skills in AI-mediated EFL environments.
These findings underscore the importance of evidence-informed practices and highlight the need for
explicit critical thinking training to ensure sustainable and responsible educational practices in the
age of generative AI.

Keywords
ChatGPT, critical thinking, EFL, experiential learning, intervention, Vietnam

*Corresponding author. Email: [email protected]

Online First View

2 International Journal of TESOL Studies

1 Introduction

The integration of generative artificial intelligence (AI) tools such as ChatGPT has rapidly reshaped
English as a Foreign Language (EFL) education, providing unprecedented opportunities for enhancing
research skills, writing proficiency, and critical thinking (Abdelhalim, 2024; Alshammari, 2024; Zou et
al., 2024). ChatGPT has been found to support learner autonomy, facilitate personalized feedback, and
foster curiosity-driven inquiry (Alshammari, 2024; De La Puente et al., 2024). In particular, studies have
noted its potential to stimulate critical engagement with content when pedagogically scaffolded (Chen,
2024; Furze et al., 2024).
However, this transformation is accompanied by significant pedagogical challenges. While AI
tools can enhance metacognitive reflection and research competency (Abdelhalim, 2024; Chen, 2024),
concerns persist regarding students’ over-reliance on AI outputs, limited verification of content accuracy,
and diminished higher-order thinking when AI use is unguided (Darwin et al., 2024; Liang & Wu, 2024;
Teng, 2023). Furthermore, while learners appreciate the convenience and immediate support provided
by AI, they often lack the critical skills needed to discern bias, inaccuracies, and ethical risks embedded
in AI-generated content (Avsheniuk et al., 2024; Hu, 2025). Therefore, pedagogical models that promote
strategic and ethical engagement with AI are urgently needed.
Among these challenges, ethical reasoning represents a particularly underdeveloped dimension in
current pedagogical responses to AI. Learners need to be equipped not only to question content accuracy
but also to recognize and respond to ethical issues such as unverified claims, biased outputs, or fabricated
citations (Hu, 2025; Avsheniuk et al., 2024). These risks demand a deliberate instructional focus on
transparency, academic integrity, and evaluative judgment in AI-mediated tasks. Accordingly, the present
study integrates discussions on ethical awareness and responsible tool use as part of its intervention
design.
To address both the cognitive and ethical dimensions of critical AI engagement, this study draws on
experiential learning theory (Kolb, 1984) as a guiding framework. This theory offers a useful foundation
for designing AI-integrated activities that encourage learners to move beyond passive consumption. By
engaging learners in cycles of experience, reflection, conceptualization, and active experimentation,
experiential approaches can cultivate deeper metacognitive awareness and critical thinking skills
necessary for responsible AI use (Fullana et al., 2016; Daradoumis & Arguedas, 2020).
Despite a burgeoning interest in AI-assisted education, empirical evidence quantifying the effects
of structured AI interventions on EFL learners’ critical thinking remains scarce. Much of the existing
research has relied on qualitative designs (Darwin et al., 2024; Liang & Wu, 2024) or self-reported
perceptions (Abdelhalim, 2024; Almazrou et al., 2024), limiting objective assessment of cognitive
outcomes. Recent experimental studies have demonstrated that AI integration can foster learner
engagement (Huang & Teng, 2025), and higher-order thinking (Deng et al., 2025; Liu & Wang, 2024),
yet there is a notable paucity of short-term, targeted interventions grounded in established pedagogical
frameworks. Moreover, few studies have employed validated quantitative measures to assess changes in
critical thinking following AI-focused interventions in EFL contexts (Yusuf et al., 2024; De La Puente et
al., 2024).
Additionally, while some research highlights the role of metacognitive strategies and structured
prompting in maximizing AI’s educational benefits (Chen, 2024; Teng, 2025), practical models that
operationalize these strategies within brief, scalable interventions are lacking. Particularly in resource-
constrained environments like Vietnam, where digital literacy training remains emergent, there is an
urgent need for accessible, effective pedagogical models that foster critical AI engagement (Furze et al.,
2024).
This study aims to address these gaps by evaluating the effectiveness of a 90-minute workshop, the
Critical AI Engagement Cycle, designed to enhance EFL learners’ critical thinking skills in interacting

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 3

with ChatGPT. Grounded in experiential learning theory (Kolb, 1984) and critical thinking frameworks
(Facione, 2011), the intervention provides scaffolded opportunities for learners to engage, reflect,
analyze, and strategize AI use.
The study is guided by the following research question:
• T
 o what extent does a structured workshop grounded in experiential learning enhance EFL learners’
critical thinking skills in using ChatGPT, as measured by a validated critical thinking scale?
This study contributes to the evolving field of AI-assisted language education by offering empirical
evidence on the impact of a structured, short-term intervention designed to foster critical thinking in EFL
learners. By focusing on a theoretically grounded and practically scalable workshop, the study addresses
key gaps in the literature related to the cognitive and ethical dimensions of AI use in language education.
The findings have important implications for curriculum design, particularly in resource-constrained
contexts where time-efficient and pedagogically sound approaches are urgently needed to equip learners
with the critical competencies required to engage with AI tools responsibly and reflectively.

2 Literature Review

2.1 Generative AI and Its Role in EFL Education

Recent research on generative AI tools such as ChatGPT has moved beyond general debates over
their benefits and risks, focusing instead on how pedagogical design shapes their effectiveness in EFL
education. While early discourse emphasized AI’s ability to personalize learning and offer real-time
support, more recent studies highlight the importance of guided interaction, critical verification, and
ethical reflection to avoid surface-level engagement and misinformation (Abdelhalim, 2024; Xu & Liu,
2025). This section reviews empirical evidence on how structured use of AI tools impacts EFL learners’
cognitive, metacognitive, and affective outcomes.
Emerging research suggests that structured pedagogical interventions are crucial for realizing
AI’s educational potential. Abdelhalim (2024) found that EFL undergraduates’ use of ChatGPT was
significantly shaped by their metacognitive awareness, underscoring the need for explicit training.
Similarly, Chen (2024) and Teng (2025) showed that AI tools can enhance reflective performance and
writing self-efficacy when used with scaffolding and clear objectives. Strobl et al. (2024) demonstrated
that using ChatGPT as a writing model in advanced L2 German classrooms promoted higher-order
thinking during revision, as students critically evaluated AI-generated texts and improved their own
writing. These findings highlight the need for structured pedagogical guidance to balance perceived
usefulness with critical engagement.
Studies in the speaking and writing domains further demonstrate the value of AI tools when used
reflectively. Hapsari and Wu (2022) found that AI chatbots reduced speaking anxiety and stimulated
critical thinking during oral practice. Shen and Tao (2025) revealed that AI-based writing feedback reduced
anxiety and improved metacognitive strategy use. Dizon et al. (2025) found that although university-
level Japanese EFL students perceived ChatGPT as a useful tool for translation and summarization, they
expressed concerns about over-reliance and the potential hindrance to authentic language development.
However, these studies primarily examine specific skill domains or affective outcomes, often relying on
self-reported data and lacking rigorous measurement of cognitive gains.
Despite promising results, several gaps remain. First, most prior research adopts qualitative designs
or perception-based surveys (e.g., Darwin et al., 2024; Alshammari, 2024), with few studies employing
validated instruments to objectively assess gains in critical thinking (see Deng et al., 2025; Liu & Wang,
2024). Second, ethical reasoning—an essential component of critical AI engagement—receives limited
attention, despite emerging risks such as misinformation, bias, and over-reliance (Avsheniuk et al.,
2024; Hu, 2025). Third, few interventions are brief, replicable, and theoretically grounded, making them

Online First View

4 International Journal of TESOL Studies

difficult to scale in resource-constrained EFL contexts (Yusuf et al., 2024; Furze et al., 2024). Fourth,
comparative analyses across subdimensions of critical thinking remain rare, leaving open the question
of which cognitive skills benefit most and why (Imjai et al., 2025). As Todd (2025) argued, generative
AI may act as a disruptive force in language education, requiring reconceptualized teaching models that
prioritize human oversight, critical thinking, and innovation.
In response, the present study proposes a short-term, structured intervention grounded in experiential
learning theory (Kolb, 1984) and aligned with a multi-dimensional model of critical thinking (Facione,
2011; Imjai et al., 2025). Rather than focusing solely on academic performance or technical skills, the
intervention aims to cultivate reflective judgment, ethical reasoning, and evaluative thinking through
scaffolded interactions with ChatGPT. By doing so, this study contributes to current literature by offering
a theoretically informed and practically feasible model of AI-integrated instruction for EFL learners.

2.2 Critical Thinking Development in EFL Contexts

Critical thinking is a core academic skill and a central outcome in modern EFL education, particularly
in AI-mediated learning environments. Defined by Facione (2011, p. 27) as “purposeful, self-regulatory
judgment” encompassing analysis, evaluation, inference, and open-mindedness, critical thinking
underpins essential abilities such as academic literacy, autonomous learning, and the responsible use of
information.
Recent scholarship has increasingly emphasized the importance of embedding critical thinking within
practical English language skills, especially in contexts where generative AI tools are integrated into
learning. For instance, Darwin et al. (2024) reported that EFL learners conceptualize critical thinking
as a process of questioning, contextual analysis, and evidence-based reasoning—skills essential for
academic reading and writing. In AI-supported writing contexts, Teng (2025) found that students with
higher metacognitive awareness made more effective use of ChatGPT for feedback, showing improved
self-efficacy and evaluative capacity. Similarly, Hapsari and Wu (2022) demonstrated that AI chatbot use
reduced speaking anxiety while enhancing critical thinking in oral interaction. Almazrou et al. (2024)
confirmed these trends in a broader context, finding that students perceived ChatGPT as beneficial in
generating diverse perspectives and encouraging reflection. Yang et al. (2024) further observed that
both students and teachers must navigate ChatGPT as a pervasive “ghostwriter,” calling for stronger
focus on ethical judgment and critical interpretation in academic writing classrooms. Collectively, these
findings affirm that critical thinking can be cultivated through targeted English language tasks such
as argumentative writing, text analysis, speaking practice, and source evaluation—particularly when
mediated through structured AI use.
Systematic reviews further reinforce these findings. Wei and Li (2024) noted that interactive and
constructive uses of AI—rather than passive consumption—are most effective in promoting learners’
analytical and inferential reasoning. Shen and Teng (2024) identified a reciprocal relationship between
AI-assisted writing, self-directed learning, and critical thinking, noting that these skills mutually reinforce
one another. Teng’s (2024) review emphasized that while ChatGPT supports writing development, its
effective integration requires deliberate scaffolding to prevent cognitive dependency. Such findings
suggest that AI tools, when embedded within purposeful and scaffolded pedagogical designs, can
enhance rather than diminish students’ cognitive engagement.
To guide both instructional design and assessment in this study, four key dimensions of critical
thinking were adopted from Imjai et al. (2025): Analytical Skills, Logical Reasoning, Evidence
Evaluation, and Open-Mindedness. These dimensions were selected due to their empirical grounding
and alignment with the specific cognitive tasks posed by AI interaction in English learning. Analytical
Skills refer to the capacity to break down complex input (e.g., AI-generated texts) and identify relevant
connections. Logical Reasoning denotes the ability to make consistent, well-supported decisions based

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 5

on structured argumentation. Evidence Evaluation involves the judgment of information credibility,

accuracy, and source validity—critical in contexts where AI may hallucinate or fabricate data. Open-
Mindedness captures the willingness to consider alternative viewpoints and revise assumptions in light of
new information, particularly relevant in learner-AI dialogue. These dimensions reflect both the cognitive
and dispositional facets of critical thinking and are particularly salient in AI-mediated academic tasks.
Accordingly, the current study builds on this growing body of evidence by evaluating how a brief,
structured and theoretically grounded intervention—guided by Kolb’s experiential learning theory—
can support the development of these four interrelated aspects of critical thinking. The intervention
aimed not only to build cognitive skill but also to foster reflective awareness and ethical discernment
in AI-supported English academic communication. As AI becomes increasingly embedded in language
learning ecosystems, developing EFL students’ critical capacities remains essential—not only for
academic success but also for navigating the complex, evolving digital landscape with autonomy and
responsibility.

2.3 Theoretical Framework: Experiential Learning as a Theoretical Foundation

Experiential learning theory (Kolb, 1984; Kolb & Kolb, 2009) provides a foundational lens for
understanding how learners actively construct knowledge through iterative cycles of doing, reflecting,
conceptualizing, and applying. Grounded in the work of Dewey, Piaget, and Lewin, experiential learning
conceptualizes learning not as the passive absorption of information, but as a continuous process
whereby knowledge emerges from the transformation of lived experience. This model offers a potentially
valuable pedagogical framework in the context of emerging technologies, where learners are increasingly
expected to navigate unfamiliar tools, assess ambiguous information, and develop the cognitive and
ethical judgment required for autonomous learning.
Kolb’s model is composed of four interconnected stages. The first, Concrete Experience, refers to
direct engagement with a task or phenomenon that initiates the learning process. Learners encounter
new content, perform an action, or interact with an environment that disrupts existing mental models
and invites inquiry. The second stage, Reflective Observation, involves critical reflection on that
experience—what occurred, what was observed, and what outcomes were unexpected. Learners begin
to notice patterns, contradictions, or gaps in understanding. The third stage, Abstract Conceptualization,
requires the learner to integrate those reflections into broader theoretical insights. This may involve
revising assumptions, generating hypotheses, or articulating generalizable principles. Finally, in Active
Experimentation, learners test their new understandings in novel contexts, applying strategies, adapting
behavior, and producing new experiences that continue the cycle.
This recursive process supports the development of metacognitive awareness, self-regulated learning,
and higher-order thinking skills. Kolb and Kolb (2009) argue that experiential learning fosters the ability
to shift flexibly between modes of action and reflection—an essential capacity in today’s complex,
AI-mediated educational environments. Similarly, Fullana et al. (2016) found that reflective learning
contributes not only to students’ academic development, but also to their motivation and sense of self
as learners. More recently, studies have applied this framework to digital contexts. For instance, Lin et
al. (2025) demonstrated that experiential learning cycles enhanced reflective thinking in AI-supported
STEM activities, while Hu (2025) and Ward et al. (2025) highlight how guided interaction, reflection,
and ethical questioning can cultivate critical AI literacy in communication and multicultural education
contexts.
Importantly, the experiential learning framework underscores the human-centered nature of
critical engagement—particularly when learners interact with technologies that produce fluent yet
fallible outputs. As Lewis and Sarkadi (2024) caution, while generative AI systems like ChatGPT can
generate plausible responses, they do not possess reflective or ethical capacity. Thus, it is important that

Online First View

6 International Journal of TESOL Studies

pedagogical designs seek to scaffold learners’ capacity to interrogate, verify, and ethically interpret AI-
generated content. Experiential learning theory provides a structured yet flexible model for fostering
these dispositions, enabling learners to build understanding not only of what AI produces, but of how and
why those outputs should be critically evaluated.
In this study, experiential learning theory functions as both a conceptual foundation and a pedagogical
rationale. It guided the design of the Critical AI Engagement Cycle, a brief, structured workshop aimed at
enhancing EFL learners’ critical thinking when interacting with ChatGPT. Rather than teaching technical
AI use in isolation, the intervention sought to embed learners in cycles of inquiry, reflection, abstraction,
and reapplication, thereby operationalizing experiential learning in the service of AI literacy. This
theoretical foundation also informs the study’s research questions and analysis, as we investigate whether
experiential engagement can lead to measurable gains in critical thinking across cognitive and ethical
domains. In doing so, the study contributes to broader efforts to align instructional design with both the
affordances and limitations of generative AI in English language education.

2.4 Positioning the Present Study

Building on the limitations identified in recent literature, this study offers a practical model for enhancing
EFL learners’ critical thinking in AI-supported contexts. Unlike many existing interventions, which are
long-term or exploratory in nature (e.g., Teng, 2025; Liu & Wang, 2024), the present research tests a
brief, structured workshop—the Critical AI Engagement Cycle—designed to promote both cognitive and
ethical engagement with AI.
Grounded in experiential learning theory, the workshop provides scaffolded opportunities for
reflection, analysis, and application of AI tools in academic tasks. The study employs a validated
scale (Imjai et al., 2025) to assess changes in four critical thinking dimensions and targets a resource-
constrained Vietnamese EFL setting. In doing so, it contributes empirical evidence to ongoing
conversations about scalability, measurement, and pedagogy in AI-mediated English education.

3 Methods

3.1 Participants

The study was conducted at a public university in Vietnam’s Central Highlands, where English as a
Foreign Language (EFL) programs have begun integrating generative AI tools such as ChatGPT. A total
of 72 students (38 in the experimental group; 34 in the control group) participated. Participants were
enrolled in undergraduate and graduate programs in English Language Studies and Applied Linguistics.
Their English proficiency ranged from upper-intermediate to advanced, based on institutional placement
tests and successful completion of English-medium academic coursework. This background helps ensure
that participants had sufficient linguistic competence to engage critically with AI-generated English
content.
Convenience sampling was employed due to logistical constraints, and participants were selected
based on course enrollment and willingness to participate. The experimental and control groups were
drawn from two intact class sections within the same program but taught separately, with no overlapping
instruction, group work, or scheduled interaction. Both groups were instructed by the same teacher using
the same syllabus, ensuring consistency in instructional delivery across groups.
The intervention was implemented as a brief, 90-minute workshop delivered only to the experimental
group, with the posttest administered immediately afterward. This design reduced the likelihood of
contamination, as students had limited opportunity to discuss the intervention content across groups
before data collection concluded. In-class discussions were confined to each section, and participants
were not informed of the study conditions assigned to the other group. Baseline characteristics such

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 7

as age, gender, and AI familiarity were similar across groups, and descriptive statistics confirmed their
general comparability, supporting the internal validity of between-group comparisons.

3.2 Intervention

A 90-minute workshop titled Developing Critical Thinking Usage of Generative AI was delivered
to the experimental group. The aim of the workshop was to raise participants’ awareness of both the
affordances and limitations of ChatGPT and to cultivate their ability to use the tool critically, reflectively,
and ethically in academic contexts. The workshop design was informed by constructivist, experiential,
and reflective learning theories (Kolb, 1984; Schön, 2017), emphasizing learning through active
experimentation, dialogic interaction, critical evaluation, and metacognitive self-awareness. These
principles were embedded in both the overall structure of the workshop and the nature of the tasks, which
systematically encouraged students to test assumptions, scrutinize outputs, and reflect on responsible AI
use.
The workshop was developed based on a structured instructional model that served as a practical
lesson plan. Its content and sequencing were informed by experiential learning theory and empirical
studies on AI literacy in EFL education. Prior to implementation, the intervention was reviewed by
two experts in TESOL and instructional design to ensure theoretical coherence and pedagogical clarity.
Although no formal pilot test was conducted, feedback from these reviewers was used to refine the
timing, complexity, and flow of tasks.
The intervention followed a five-phase instructional sequence, conceptualized as the Critical AI
Engagement Cycle, which scaffolded the learning experience around progressively deeper engagement
with ChatGPT. Students engaged primarily in individual tasks supplemented by guided peer discussions
to foster collaborative reflection. A visual overview of the Critical AI Engagement Cycle is provided to
illustrate the pedagogical structure (see Figure 1).

Figure 1
The Critical AI Engagement Cycle: A Five-Phase Structure for Guiding Critical and Reflective Use of
Generative AI

Online First View

8 International Journal of TESOL Studies

In the first phase, Awareness-Building, students were introduced to ChatGPT’s development process
and inherent limitations, drawing on documentation from OpenAI that described reinforcement learning
processes, fine-tuning procedures, and common issues such as factual hallucinations and response
variability. In the second phase, Empirical Grounding, participants engaged with synthesized findings
from recent empirical research (Cong-Lem et al., 2024), which highlighted ChatGPT’s fabrication of
references, factual inaccuracies, inconsistent domain-specific performance, and ethical risks in academic
settings.
The third phase, Experiential Testing, involved a series of guided, hands-on tasks requiring
participants to directly evaluate ChatGPT’s outputs. Activities included verifying AI-generated
references, evaluating ChatGPT’s accuracy in summarizing scholarly research, assessing its performance
on logical reasoning prompts, and evaluating its mathematical accuracy. For example, in one task,
students prompted ChatGPT to generate a paragraph about a common language learning theory and then
asked it to list references with DOI links. They then attempted to locate these references via Google
Scholar, discovering that several were fabricated. In another activity, students crafted prompts based on
real studies (e.g., “Tell me about the findings of the 2025 study by Yu on vocabulary acquisition”) and
evaluated whether ChatGPT could identify the correct content. Additional exercises involved asking
ChatGPT to compose a 25-word sentence and checking for precision in word count, or prompting it with
a logic problem involving multiple possible causes of failure to assess the AI’s reasoning consistency.
Students were also instructed to challenge ChatGPT’s responses—e.g., by refuting its initial suggestion
on the best method for vocabulary learning—and observe whether it maintained or revised its position.
Through these exercises, students developed firsthand awareness of the model’s cognitive strengths and
limitations. In the fourth phase, Strategy Development, students received explicit instruction and practice
in refining prompts, validating information sources, and treating ChatGPT’s responses as preliminary
starting points for further human inquiry rather than as definitive answers.
Finally, in the fifth phase, Reflective Synthesis, students participated in a concluding instructor-
facilitated discussion and completed a written reflection activity. They were encouraged to articulate their
evolving perspectives on the responsible use of AI tools in academic work, focusing on critical judgment,
recognition of AI’s limitations, and ethical considerations. The instructor served as a facilitator and
critical thinking coach throughout, guiding students’ reflections without prescribing answers.
Approximately 15 to 20 minutes were allocated to each phase to ensure structured progression
within the 90-minute session. The workshop was conducted in a computer-equipped classroom to
enable immediate interaction with ChatGPT during activities. This design ensured that students moved
through Kolb’s experiential learning cycle—concrete experience, reflective observation, abstract
conceptualization, and active experimentation—while building competencies across analytical reasoning,
evidence evaluation, logical inference, and open-mindedness.

3.3 Instruments

Participants completed a two-part survey administered pre- and post-intervention. Part 1 collected
demographic information (age, gender, AI use frequency, prior AI knowledge, and self-reported AI skill
level). Part 2 consisted of a 12-item critical thinking scale adapted from Imjai et al. (2025), assessing
four subdomains: Analytical Skills (AS), Logical Reasoning (LR), Evidence Evaluation (EE), and Open-
Mindedness (OM). Minor contextual adjustments were made to reflect the study’s focus on AI-assisted
English learning, such as replacing general terms (e.g., “data” or “information”) with references to “AI-
generated content” or “tools like ChatGPT.”
These surface-level wording changes preserved the conceptual integrity and original structure of
each subscale. For example, the item “You consistently analyse complex data…” was revised to “I
consistently analyze complex AI-generated content…” to situate the statement within an educational

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 9

AI context. To improve clarity and balance across dimensions, one additional item was added to each
subscale, resulting in three items per dimension.
Each item was rated on a 5-point Likert scale (1 = Strongly Disagree; 5 = Strongly Agree). Internal
consistency reliability was high, with Cronbach’s alpha coefficients of .90 (pretest) and .85 (posttest) in
this study. The slight decrease in posttest alpha likely reflects natural variation in participants’ response
patterns following the intervention, and both values remain within the range indicating strong reliability.

3.4 Data Collection and Analysis

Pre- and post-tests were administered in controlled classroom settings immediately before and after the
intervention. No missing data were reported; all participants completed both assessments. Descriptive
statistics (means, standard deviations) were calculated for all critical thinking measures. Paired-samples
t-tests or Wilcoxon signed-rank tests were employed to analyze within-group changes, depending on data
normality. Cohen’s d was computed to evaluate effect sizes. Between-group comparisons at post-test
were analyzed using Analysis of Covariance (ANCOVA), controlling for pretest scores and covariates
(gender, AI use frequency, AI skill level, and prior AI knowledge). Homogeneity of regression slopes and
normality of residuals were tested to validate ANCOVA assumptions. All analyses were conducted using
R version 4.4.2, with a significance threshold set at p < .05.

3.5 Ethical Considerations

The study received approval from the president of the university where the participants were enrolled.
Informed consent was obtained from all participants. Participation was voluntary, with the option to
withdraw at any time without penalty. Data were anonymized using unique participant codes. Beyond
procedural ethics, the workshop explicitly emphasized responsible AI use, encouraging students to
develop critical thinking not only for cognitive advancement but also for ethical engagement with AI-
generated content, addressing concerns about misinformation, bias, and academic integrity.

4 Results

This section presents the findings related to the impact of the Critical AI Engagement Cycle workshop
on EFL learners’ critical thinking skills. Descriptive statistics, inferential analyses (within- and between-
group comparisons), pretest equivalence testing, and ANCOVA results are reported systematically to
address the research question: To what extent does a structured workshop grounded in experiential
learning enhance EFL learners’ critical thinking skills in using ChatGPT, as measured by a validated
critical thinking scale?

4.1 Participant Background Characteristics

Table 1 presents descriptive statistics for participant background variables. The sample consisted of 72
participants (experimental group: n = 38; control group: n = 34), with a mean age of 22 years (SD =
4.28). The majority of participants were female (83.3%). Regarding AI familiarity, 56.9% of participants
reported moderate to high AI use frequency. Mean scores for AI skill level and prior knowledge were
comparable across groups, indicating similar levels of technological readiness at baseline. To account for
potential confounding effects, all background variables—including gender, AI use frequency, prior AI
knowledge, and self-rated AI skill level—were included as covariates in subsequent ANCOVA analyses.

Online First View

10 International Journal of TESOL Studies

Table 1
Participant Background Characteristics
Variable Experimental (n = 38) Control (n = 34) Total (N = 72)
Age (Mean ± SD) 22.1 ± 4.2 21.9 ± 4.4 22.0 ± 4.3
Gender (Female %) 84.20% 82.40% 83.30%
AI Use Frequency (Mod-High %) 58% 55% 56.90%
AI Skill Level (Mean ± SD) 3.45 ± 0.62 3.41 ± 0.59 3.43 ± 0.61
Prior Knowledge (Mean ± SD) 3.39 ± 0.67 3.37 ± 0.65 3.38 ± 0.66

4.2 Pretest Equivalence Testing

To examine baseline equivalence between groups, Mann-Whitney U tests were conducted comparing
pretest critical thinking scores between the experimental and control groups. No significant differences
were found for Analytical Skills (U = 680, p = .702), Logical Reasoning (U = 637, p = .923), Evidence
Evaluation (U = 714, p = .440), Open-Mindedness (U = 660.5, p = .872), or Overall Critical Thinking (U
= 664.5, p = .839). These results suggest that the two groups were statistically comparable at the outset of
the study.

4.3 Critical Thinking Scores: Descriptive Statistics

As shown in Table 2, the experimental group exhibited notable improvements across all measures of
critical thinking. Overall Critical Thinking (Overall_CT) scores increased from a pretest mean of 3.33 (SD
= 0.69) to a posttest mean of 4.02 (SD = 0.37). Similarly, substantial gains were observed across the four
critical thinking subdomains, particularly in Evidence Evaluation (EE) and Logical Reasoning (LR). In
contrast, the control group showed smaller gains across all measures, with Overall_CT increasing from
3.29 (SD = 0.80) to 3.61 (SD = 0.63).

Table 2
Descriptive Statistics for Critical Thinking Scores
Measure Group Pretest Mean (SD) Posttest Mean (SD)
Overall Critical Thinking Experimental 3.33 (0.69) 4.02 (0.37)
Control 3.29 (0.80) 3.61 (0.63)
Analytical Skills Experimental 3.27 (0.73) 3.87 (0.49)
Control 3.25 (0.75) 3.50 (0.69)
Logical Reasoning Experimental 3.33 (0.78) 4.07 (0.45)
Control 3.29 (0.98) 3.64 (0.70)
Evidence Evaluation Experimental 3.30 (0.79) 4.07 (0.49)
Control 3.23 (0.86) 3.56 (0.86)
Open-Mindedness Experimental 3.40 (0.81) 4.07 (0.47)
Control 3.38 (0.82) 3.75 (0.73)

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 11

4.4 Within-Group Pretest and Posttest Comparisons

To examine the workshop’s impact within the experimental group, paired-samples tests were conducted.
Overall Critical Thinking significantly improved, as indicated by the Wilcoxon signed-rank test (W = 19,
p < .001, Cohen’s d = 1.23), reflecting a large effect size. Analytical Skills (AS) also showed a significant
increase (W = 37, p < .001, d = 0.94). Logical Reasoning (LR) improved significantly, with a paired t-test
yielding t(37) = -6.07, p < .001, d = 1.11. Evidence Evaluation (EE) exhibited significant gains as well
(t(37) = -4.78, p < .001, d = 1.19). Lastly, Open-Mindedness (OM) significantly improved (W = 45, p <
.001, d = 1.00). In comparison, the control group showed modest improvements with smaller effect sizes,
suggesting limited development without the intervention.

4.5 Between-Group Comparisons at Posttest

Further analyses using Mann-Whitney U tests were conducted to compare posttest critical thinking
scores between the experimental and control groups. Table 3 summarizes the results.

Table 3
Between-Group Posttest Comparisons (Mann-Whitney U Tests)
Measure U Statistic p-value Cohen’s d
Overall Critical Thinking 384 .003 0.80
Analytical Skills 443 .018 0.62
Logical Reasoning 379.5 .002 0.74
Evidence Evaluation 407.5 .005 0.75
Open-Mindedness 483 .057 0.52

The experimental group significantly outperformed the control group on Overall Critical Thinking,
Analytical Skills, Logical Reasoning, and Evidence Evaluation. Differences in Open-Mindedness
approached significance.

4.6 ANCOVA Results

To further validate the effects of the workshop, Analysis of Covariance (ANCOVA) was conducted.
ANCOVA is a statistical technique that compares posttest scores between groups while controlling for
potential confounding variables known as covariates. Tests of homogeneity of regression slopes and
normality of residuals confirmed that ANCOVA assumptions were not violated. In this study, covariates
included pretest critical thinking scores, gender, AI use frequency, AI skill level, and prior knowledge
related to AI. Controlling for these variables ensures that differences observed at posttest are more
confidently attributed to the intervention rather than pre-existing group differences.

Table 4
ANCOVA Results Controlling for Covariates
Measure F (1, 62) p-value Partial η²
Overall Critical Thinking 17.60 <.001 .22
Analytical Skills 9.07 .004 .13
Logical Reasoning 12.50 .001 .17
Evidence Evaluation 11.74 .001 .16
Open-Mindedness 3.68 .060 .06

Online First View

12 International Journal of TESOL Studies

The results confirm significant differences favoring the experimental group across Overall Critical
Thinking and three critical thinking subdomains after controlling for baseline scores and covariates.
In summary, the findings provide robust evidence that the 90-minute Critical AI Engagement
Cycle workshop significantly enhanced EFL learners’ critical thinking skills when using ChatGPT.
Improvements were consistent across Overall Critical Thinking and its subdomains, particularly
in Evidence Evaluation and Logical Reasoning, supporting the efficacy of short-term, structured
interventions grounded in experiential learning theory for fostering critical AI literacy.

5 Discussion

5.1 Summary and Interpretations of Key Findings

This study investigated the effectiveness of a 90-minute Critical AI Engagement Cycle workshop in
enhancing EFL learners’ critical thinking skills when interacting with ChatGPT. The results demonstrated
that the experimental group achieved significant improvements across Overall Critical Thinking and
four subdomains (Analytical Skills, Logical Reasoning, Evidence Evaluation, and Open-Mindedness),
whereas the control group showed only modest gains. ANCOVA analyses, controlling for pretest scores,
gender, AI use frequency, AI skill level, and prior knowledge, supported the robustness of these findings.
Several features of the intervention may help explain its effectiveness. Grounded in experiential
learning theory, the design guided learners through a sequence of active engagement, reflection,
conceptual analysis, and experimentation. This structure appears to have supported deeper metacognitive
processing than passive exposure to AI tools. Participants critically examined AI-generated outputs,
identified inconsistencies, and practiced verification strategies—behaviors aligned with higher-order
thinking development.
Modest improvements were also observed in the control group across several critical thinking
subdomains. These gains, while smaller than those of the experimental group, could be attributed to
general cognitive development during the academic semester, increased test familiarity, or indirect
exposure to critical thinking tasks through regular coursework. However, as confirmed by ANCOVA
results, the improvements in the experimental group remained significantly greater even after controlling
for pretest scores and background variables.
The differential improvements observed across the four critical thinking subdomains can be
understood in light of existing empirical evidence and the design features of the intervention. Specifically,
Evidence Evaluation and Logical Reasoning showed the greatest gains, which may be attributed to
the hands-on, task-based activities that explicitly required students to verify AI-generated outputs and
challenge flawed logic in ChatGPT’s responses. This corresponds with findings from Abdelhalim (2024)
and Teng (2025), who emphasize the importance of metacognitive awareness and evaluative reasoning
in AI-mediated learning. Similarly, Almazrou et al. (2024) reported that students perceived ChatGPT as
especially beneficial for generating diverse perspectives and prompting analytical thinking, especially
when guided by structured tasks. In contrast, more dispositional subdomains like Open-Mindedness
may require repeated practice and deeper reflective engagement to see substantial change (Darwin et al.,
2024; Fullana et al., 2016). Thus, the results suggest that experiential, feedback-rich interactions with AI
can differentially foster the development of specific critical thinking facets, depending on how they are
scaffolded in instructional design.

5.2 Comparison with Previous Literature

The findings are broadly consistent with previous studies highlighting the positive influence of structured
AI use on critical thinking development in EFL contexts (Abdelhalim, 2024; Deng et al., 2025; Liu &

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 13

Wang, 2024). The substantial gains, particularly in Evidence Evaluation and Logical Reasoning, reinforce
the importance of embedding experiential cycles of reflection, testing, and strategy development into AI-
integrated learning environments. This study extends prior research by demonstrating that even a single-
session intervention, grounded in experiential learning theory (Kolb, 1984), can produce meaningful
cognitive outcomes when pedagogically scaffolded.
However, these results contrast with findings by Liang and Wu (2024), who reported limited critical
thinking gains despite ChatGPT use. One plausible explanation lies in the design of the intervention.
Unlike unguided AI exposure, the Critical AI Engagement Cycle deliberately operationalized experiential
learning stages. Students engaged in concrete interaction (generating outputs), reflective observation
(identifying limitations), abstract conceptualization (developing theories about AI behavior), and active
experimentation (strategizing AI use). This structured experiential design likely enabled deeper and more
sustained critical engagement.
Moreover, the present findings resonate with broader research emphasizing strategic and ethical AI
literacy development in EFL education (Chen, 2024; Darwin et al., 2024). Students learned not only to
identify logical fallacies or evaluate evidence but also to question AI outputs’ credibility and reliability—
skills essential for navigating English academic communication contexts.
Compared to more traditional expository or lecture-based instruction, which often emphasizes
procedural use of AI tools or passive knowledge transfer, the experiential learning approach used in this
study actively engaged learners in questioning, evaluating, and experimenting with AI-generated outputs.
This pedagogical contrast may explain the observed gains: rather than being shown how ChatGPT
functions, participants had to uncover its limitations through guided discovery and reflection. This
active involvement, supported by structured prompts and peer discussion, aligns with previous research
highlighting the superiority of experiential models in fostering deep learning and metacognitive skills
(Kolb & Kolb, 2009; Lin et al., 2025).
This study offers several contributions to the field of EFL education. First, it demonstrates the
potential for critical thinking gains to be achieved through a brief, targeted, and resource-efficient
intervention, making such initiatives highly promising and scalable for resource-constrained settings.
Second, by integrating covariates such as AI skill level and prior knowledge into ANCOVA analyses, the
study addresses methodological gaps noted in earlier AI-in-EFL studies (e.g., Hapsari & Wu, 2022; Teng,
2025). Third, by proposing a conceptual model that views AI literacy as encompassing cognitive, ethical,
and strategic dimensions, the study advances understanding of how to critically engage EFL learners
with emerging technologies.
The outcomes of this study resonate strongly with the objectives of the United Nations Sustainable
Development Goals, notably SDG 4: Quality Education and SDG 9: Industry, Innovation, and
Infrastructure (United Nations, n.d.). SDG 4 emphasizes the provision of inclusive and equitable quality
education and the promotion of lifelong learning opportunities for all. By equipping EFL learners with
critical thinking skills and AI literacy, the intervention contributes to the transformation of education
systems to address the emerging challenges and opportunities of the generative AI era. This aligns with
UNESCO’s advocacy for integrating AI competencies into education to foster human-centered and
ethical use of technology.
Simultaneously, the study contributes to SDG 9 by fostering innovation in educational practices
through the integration of AI tools like ChatGPT. By developing a structured framework—the Critical
AI Engagement Cycle—the research promotes sustainable industrialization and innovation within
the educational sector. This approach not only enhances the technological capabilities of learners
but also encourages the development of infrastructure that supports innovative teaching and learning
methodologies. Such initiatives are crucial for building resilient educational systems that can adapt to
technological advancements and prepare students for the demands of the modern workforce.

Online First View

14 International Journal of TESOL Studies

5.3 Implications for English Language Teaching, Learning, and Policy

The findings of this study offer significant implications for English language teaching, learning, and
educational policy. Beyond fostering general critical thinking, the Critical AI Engagement Cycle
suggests potential contributions to EFL education in several ways. First, it may support the development
of academic literacy skills fundamental to EFL success, including source evaluation, evidence-
based argumentation, and critical academic writing (Abdelhalim, 2024). Encouraging students to
verify information, identify bias, and critically assess AI-generated texts can strengthen their reading
comprehension and analytical writing abilities in English.
Second, the intervention may promote critical language awareness (Darwin et al., 2024), sensitizing
learners to discourse patterns, rhetorical inconsistencies, and persuasive techniques embedded in AI-
generated language. This metalinguistic sensitivity is particularly important for advanced EFL learners
navigating increasingly complex English academic texts. Third, by emphasizing strategic prompting and
critical verification, the workshop may foster greater learner autonomy (Xu & Liu, 2025), encouraging
students to engage with AI tools in a self-directed and critical manner.
Furthermore, by raising awareness of ethical considerations, the intervention encourages more
responsible language practices among students, addressing growing concerns about plagiarism and
academic integrity in AI-assisted writing (Shen & Tao, 2025). Developing ethical AI literacy alongside
critical thinking skills is essential for preparing EFL learners to participate responsibly in AI-mediated
academic environments.
In addition, this study contributes to the global call for evidence-informed and sustainable education
in the generative AI era (Baskara, 2023; Corbeil & Corbeil, 2025). Rather than relying on speculative
claims or anecdotal enthusiasm surrounding generative AI, the intervention was explicitly grounded
in theoretical frameworks and supported by empirical validation. This alignment between pedagogical
design, cognitive theory, and assessment underscores the importance of basing instructional decisions
on rigorous, context-sensitive research. As AI tools continue to be integrated into English language
education, adopting an evidence-informed approach can help educators avoid premature implementation,
ensure ethical standards, and tailor interventions to specific learner needs. The present study serves
as an example of how short-term, resource-efficient programs can be both theoretically grounded and
empirically tested, supporting scalable innovation in AI-assisted EFL contexts.
For educators, the Critical AI Engagement Cycle provides a practical framework for integrating
critical thinking and AI literacy into English language instruction without requiring major curricular
overhauls. It can be flexibly implemented through targeted classroom tasks that prompt students to
critically engage with AI outputs. For instance, learners might evaluate the consistency of AI-generated
explanations across different prompts, detect logical fallacies or unsupported claims in model responses,
or challenge ChatGPT’s suggestions using counterexamples. These activities cultivate habits of
questioning, verification, and evidence-based reasoning that are transferable across academic contexts.
Modular workshops, embedded within writing, reading, or research skills courses, can meaningfully
enhance students’ critical and communicative competencies while adapting flexibly to different
institutional contexts. Such integrations require minimal technological infrastructure and can be
facilitated using guided prompts, peer critique, and reflective tasks—tools already familiar to language
educators.
For policymakers and curriculum designers, these findings highlight the urgency of incorporating
AI literacy components into EFL programs. Equipping students with the cognitive, ethical, and strategic
skills needed to engage with AI-generated information is vital for fostering informed, ethical, and
autonomous English language users. Moreover, the success of a short, focused intervention suggests that
even minimal curriculum adaptations can yield substantial benefits. Resource-limited institutions may
particularly benefit from implementing scalable, experiential models like the Critical AI Engagement

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 15

Cycle to build 21st-century skills among EFL learners without extensive financial or technological
investments.

5.4 Limitations and Future Research Directions

Despite its contributions, the study has several limitations. First, the quasi-experimental design, with
participants drawn from intact classes rather than randomized groups, may introduce selection biases that
limit causal inference. However, baseline characteristics such as age, gender, AI familiarity, and pretest
scores were comparable across groups, and no significant pre-intervention differences were found.
To further minimize potential confounding effects, these variables were included as covariates in the
ANCOVA analysis. Although random assignment would provide stronger internal validity, these design
and statistical safeguards help mitigate the impact of selection bias. Second, the sample was restricted to
one Vietnamese university, limiting generalizability to broader EFL populations. Future research should
replicate this intervention across diverse educational contexts, including secondary and international EFL
settings.
Third, the study measured critical thinking immediately post-intervention; longitudinal studies
are needed to assess the durability of gains over time. Future research could also explore how critical
thinking skills transfer to academic writing, reading comprehension, and oral communication tasks in
English. Comparative studies examining single-session versus multi-session interventions, or experiential
versus expository AI literacy approaches, would deepen understanding of effective pedagogical models.
Finally, although the study used a validated critical thinking scale adapted for AI-assisted EFL
contexts, relying solely on self-reported measures may introduce biases such as social desirability or
overestimation of ability. Future research should consider supplementing survey data with additional
sources such as reflective journals, classroom observations, or performance-based assessments.
Triangulating these data sources would offer richer insights into how learners apply critical thinking
skills in real-world AI-mediated academic tasks and enhance the validity of findings.

6 Conclusion

This study provides compelling evidence that a short, structured experiential learning intervention can
significantly enhance EFL learners’ critical thinking skills when interacting with generative AI tools such
as ChatGPT. By operationalizing Kolb’s experiential learning cycle into the Critical AI Engagement
Cycle workshop, learners were guided through a systematic process of exploration, reflection,
conceptualization, and strategic experimentation. The intervention not only improved students’ analytical
reasoning, evidence evaluation, logical thinking, and open-mindedness but also emphasized ethical and
strategic engagement with AI outputs.
Importantly, the results underscore that critical thinking development in AI-enhanced environments
does not necessarily require long, resource-intensive programs. Even a 90-minute focused session,
when thoughtfully designed, can yield substantial cognitive benefits. These findings have significant
implications for AI literacy education, particularly in resource-constrained EFL contexts where time and
institutional capacity for curricular innovation may be limited.
Nevertheless, further research is necessary to explore the long-term sustainability of these gains, their
transferability to real-world academic tasks, and their adaptability across diverse cultural and linguistic
settings. Future work could also refine the experiential model, incorporating students’ reflections and
emotional responses to better scaffold critical engagement with emerging technologies.
Overall, the Critical AI Engagement Cycle offers a practical, scalable, and theoretically grounded
pathway for cultivating discerning, reflective learners capable of navigating the complexities of the AI
era responsibly and ethically.

Online First View

16 International Journal of TESOL Studies

Acknowledgements

The authors would like to express their sincere gratitude to all the students who participated in this study
for their time and engagement. Special thanks are extended to Ms. Nguyen Hoang Nhat Quyen for her
valuable assistance during the early stages of the project. We also appreciate the constructive feedback
provided by the editor and the anonymous reviewers, which greatly contributed to improving the clarity
and quality of this manuscript.

Funding Statement

This research is funded by the Foundation for Science and Technology Development of Dalat University.
Grant No: 1153/QĐ-ĐHDL.

Ethics Approval Statement

All procedures performed in the study were in accordance with the ethical standards of the Institutional
and/or National Research Committee and with the 1964 Helsinki Declaration and its later amendments
or comparable ethical standards.

Declaration of AI Use

The authors used generative AI technology (ChatGPT by OpenAI) to support the proofreading and
language refinement of this manuscript. The use of AI was conducted under human oversight, and
the final manuscript was thoroughly reviewed and edited to ensure accuracy and integrity. The author
remains fully responsible and accountable for the content and conclusions presented in this work.

Appendix A. Critical Thinking Instrument

The following items were developed based on the four critical thinking dimensions adapted from
Imjai et al. (2025): Analytical Skills, Logical Reasoning, Evidence Evaluation, and Open-Mindedness.
Participants rated each statement on a 5-point Likert scale ranging from 1 (Strongly Disagree) to 5
(Strongly Agree). Vietnamese translations are provided in parentheses for participant accessibility.

Analytical Skills

AS1. I can identify the main points and supporting ideas in AI-generated content.
(Tôi có thể xác định các luận điểm chính và các ý hỗ trợ trong nội dung do AI tạo ra.)
AS2. I am able to break down AI-generated information into smaller parts to better understand it.
(Tôi có thể phân tích nội dung do AI tạo ra thành các phần nhỏ hơn để hiểu rõ hơn.)
AS3. I can distinguish between factual content and opinions in AI-generated responses.
(Tôi có thể phân biệt giữa thông tin thực tế và ý kiến trong phản hồi do AI tạo ra.)

Logical Reasoning

LR1. I use logical reasoning when deciding how much to trust AI-generated suggestions.
(Tôi sử dụng lập luận logic khi quyết định mức độ tin tưởng vào các đề xuất do AI tạo ra.)

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 17

LR2. I make sure that my conclusions are not solely influenced by AI outputs.
(Tôi đảm bảo rằng quyết định của tôi không chỉ bị ảnh hưởng bởi các kết quả của AI.)
LR3. I can detect inconsistencies or contradictions in AI-generated content and adjust my conclusions
accordingly.
(Tôi có thể phát hiện sự không nhất quán hoặc mâu thuẫn trong nội dung do AI tạo ra và điều chỉnh kết
luận của mình cho phù hợp.)

Evidence Evaluation

EE1. I assess the credibility of AI-generated information by cross-referencing it with reliable sources.
(Tôi đánh giá độ tin cậy của thông tin do AI tạo ra bằng cách đối chiếu với các nguồn đáng tin cậy.)
EE2. I verify the accuracy and source of AI-generated content before applying it in my teaching or
learning tasks.
(Tôi kiểm tra độ chính xác và nguồn gốc của nội dung do AI tạo ra trước khi áp dụng vào công việc
giảng dạy hoặc học tập của mình.)
EE3. I critically evaluate whether AI-generated content is supported by sufficient evidence or data.
(Tôi đánh giá một cách khoa học liệu nội dung do AI tạo ra có được hỗ trợ bởi đủ bằng chứng hoặc dữ
liệu hay không.)

Open-Mindedness

OM1. I consider and respect perspectives that differ from AI-generated suggestions, integrating both AI
and non-AI insights into my decision-making.
(Tôi xem xét và tôn trọng các quan điểm khác với các đề xuất do AI đưa ra, kết hợp cả những hiểu biết
từ AI và không phải AI vào quá trình ra quyết định của mình.)
OM2. I am willing to experiment with new ways to critically assess and use AI-generated content in my
academic work.
(Tôi sẵn sàng thử nghiệm các cách thức mới để đánh giá một cách phản biện và sử dụng nội dung do AI
tạo ra trong công việc học tập của mình.)
OM3. I remain open to revising my ideas when AI-generated information presents new insights or
challenges my assumptions.
(Tôi luôn cởi mở trong việc điều chỉnh ý tưởng của mình khi thông tin do AI tạo ra đưa ra những hiểu
biết mới hoặc thách thức các giả định của tôi.)

References

Abdelhalim, S. (2024). Using ChatGPT to promote research competency: English as a Foreign Language
undergraduates’ perceptions and practices across varied metacognitive awareness levels. Journal of
Computer Assisted Learning, 40(3), 1261–1275. https://doi.org/10.1111/jcal.12948
Almazrou, S., Alanezi, F., Almutairi, S. A., AboAlsamh, H. M., Alsedrah, I. T., Arif, W. M., Alsadhan, A.
A., AlSanad, D. S., Alqahtani, N. S., AlShammary, M. H., Bakhshwain, A. M., Almuhanna, A. F.,
Almulhem, M., Alnaim, N., Albelali, S., & Attar, R. W. (2024). Enhancing medical students critical
thinking skills through ChatGPT: An empirical study with medical students. Nutrition and Health,
02601060241273627. https://doi.org/10.1177/02601060241273627
Alshammari, J. (2024). Revolutionizing EFL learning through ChatGPT: A qualitative study. Revista
Amazonia Investiga, 13(82), 208–221. https://doi.org/10.34069/AI/2024.82.10.17

Online First View

18 International Journal of TESOL Studies

Avsheniuk, N., Lutsenko, O., Svyrydiuk, T., & Seminikhyna, N. (2024). Empowering language learners’
critical thinking: Evaluating ChatGPT’s role in english course implementation. Arab World English
Journal, 210–224. https://doi.org/10.24093/awej/ChatGPT.14
Baskara, F. R. (2023). Integrating ChatGPT into EFL writing instruction: Benefits and challenges.
International Journal of Education and Learning, 5(1), 44–55. https://doi.org/10.31763/ijele.
v5i1.858
Chen, K., Tallant, A. C., & Selig, I. (2024). Exploring generative AI literacy in higher education: Student
adoption, interaction, evaluation and ethical perceptions. Information and Learning Sciences.
https://doi.org/10.1108/ILS-10-2023-0160
Cong-Lem, N., Soyoof, A., & Tsering, D. (2024). A systematic review of the limitations and associated
opportunities of ChatGPT. International Journal of Human–Computer Interaction, 1–16. https://
doi.org/10.1080/10447318.2024.2344142
Corbeil, J. R., & Corbeil, M. E. (2025). Teaching and learning in the age of generative AI: evidence-
based approaches to pedagogy, ethics, and beyond. Taylor & Francis.
Daradoumis, T., & Arguedas, M. (2020). Cultivating students’ reflective learning in metacognitive
activities through an affective pedagogical agent. Educational Technology & Society, 23(2), 19–31.
JSTOR.
Darwin, Rusdin, D., Mukminatien, N., Suryati, N., Laksmi, E. D., & Marzuki. (2024). Critical thinking
in the AI era: An exploration of EFL students’ perceptions, benefits, and limitations. Cogent
Education, 11(1), 2290342. https://doi.org/10.1080/2331186X.2023.2290342
De La Puente, M., Torres, J., Troncoso, A. L. B., Meza, Y. Y. H., & Carrascal, J. X. M. (2024).
Investigating the use of ChatGPT as a tool for enhancing critical thinking and argumentation skills
in international relations debates among undergraduate students. Smart Learning Environments,
11(1), 55. https://doi.org/10.1186/s40561-024-00347-0
Deng, R., Jiang, M., Yu, X., Lu, Y., & Liu, S. (2025). Does ChatGPT enhance student learning? A
systematic review and meta-analysis of experimental studies. Computers & Education, 227,
105224. https://doi.org/10.1016/j.compedu.2024.105224
Dizon, G., Gold, J., & Barnes, R. (2025). ChatGPT for self-regulated language learning: University
English as a foreign language students’ practices and perceptions. Digital Applied Linguistics, 3,
102510. https://doi.org/10.29140/dal.v3.102510
Facione, P. A. (2011). Critical thinking: What it is and why it counts. Insight Assessment, 1(1), 1–23.
Fullana, J., Pallisera, M., Colomer, J., Fernández Peña, R., & Pérez-Burriel, M. (2016). Reflective
learning in higher education: A qualitative study on students’ perceptions. Studies in Higher
Education, 41(6), 1008–1022. https://doi.org/10.1080/03075079.2014.950563
Furze, L., Perkins, M., Roe, J., & MacVaugh, J. (2024). The AI assessment scale (AIAS) in action:
A pilot implementation of GenAI-supported assessment. Australasian Journal of Educational
Technology, 40(4), 38–55. Scopus. https://doi.org/10.14742/ajet.9434
Hapsari, I. P., & Wu, T.-T. (2022). AI chatbots learning model in english speaking skill: Alleviating
speaking anxiety, boosting enjoyment, and fostering critical thinking. 13449 LNCS, 444–453.
Scopus. https://doi.org/10.1007/978-3-031-15273-3_49
Hu, Y. (2025). Generative AI, communication, and stereotypes: Learning critical AI literacy through
experience, analysis, and reflection. Communication Teacher, 39(1), 6–12.
Huang, J., & Teng, M. F. (2025). Peer feedback and ChatGPT-generated feedback on Japanese EFL
students’ engagement in a foreign language writing context. Digital Applied Linguistics, 2, 102469.
https://doi.org/10.29140/dal.v2.102469

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 19

Imjai, N., Yordudom, T., Yaacob, Z., Saad, N. H. M., & Aujirapongpan, S. (2025). Impact of AI literacy
and adaptability on financial analyst skills among prospective Thai accountants: The role of critical
thinking. Technological Forecasting and Social Change, 210, 123889. https://doi.org/10.1016/
j.techfore.2024.123889
Kolb, A. Y., & Kolb, D. A. (2009). The learning way: Meta-cognitive aspects of experiential learning.
Simulation & Gaming, 40(3), 297–327. https://doi.org/10.1177/1046878108325713
Kolb, D. (1984). Experiential learning: Experience as the source of learning and development. In Journal
of Business Ethics (Vol. 1).
Lee, H.-Y., Chen, P.-H., Wang, W.-S., Huang, Y.-M., & Wu, T.-T. (2024). Empowering ChatGPT with
guidance mechanism in blended learning: Effect of self-regulated learning, higher-order thinking
skills, and knowledge construction. International Journal of Educational Technology in Higher
Education, 21(1), 16. https://doi.org/10.1186/s41239-024-00447-4
Lewis, P. R., & Sarkadi, Ş. (2024). Reflective artificial intelligence. Minds and Machines, 34(2), 14.
https://doi.org/10.1007/s11023-024-09664-2
Liang, W., & Wu, Y. (2024). Exploring the use of ChatGPT to foster EFL learners’ critical thinking
skills from a post-humanist perspective. Thinking Skills and Creativity, 54, 101645. https://doi.
org/10.1016/j.tsc.2024.101645
Lin, C.-J., Lee, H.-Y., Wang, W.-S., Huang, Y.-M., & Wu, T.-T. (2025). Enhancing reflective thinking
in STEM education through experiential learning: The role of generative AI as a learning aid.
Education and Information Technologies, 30(5), 6315–6337. https://doi.org/10.1007/s10639-024-
13072-5
Liu, W., & Wang, Y. (2024). The effects of using AI tools on critical thinking in english literature classes
among EFL learners: An intervention study. European Journal of Education, 59(4), e12804. https://
doi.org/10.1111/ejed.12804
Schenck, A. (2024). Examining relationships between technology and critical thinking: A study of south
korean EFL learners. Education Sciences, 14(6), Article 6. https://doi.org/10.3390/educsci14060652
Schön, D. A. (2017). The reflective practitioner: How professionals think in action. Routledge.
Shen, X., & Tao, Y. (2025). Metacognitive strategies, AI-based writing self-efficacy and writing anxiety
in AI-assisted writing contexts: A structural equation modeling analysis. International Journal of
TESOL Studies, 7(1), 70–87. Scopus. https://doi.org/10.58304/ijts.20250105
Shen, X., & Teng, M. F. (2024). Three-wave cross-lagged model on the correlations between critical
thinking skills, self-directed learning competency and AI-assisted writing. Thinking Skills and
Creativity, 52, 101524. https://doi.org/10.1016/j.tsc.2024.101524
Strobl, C., Menke-Bazhutkina, I., Abel, N., & Michel, M. (2024). Adopting ChatGPT as a writing buddy
in the advanced L2 writing class. Technology in Language Teaching & Learning, 6(1), 1168. https://
doi.org/10.29140/tltl.v6n1.1168
Teng, M. F. (2023). Scientific writing, reviewing, and editing for open-access TESOL journals: The
role of ChatGPT. International Journal of TESOL Studies, 5(1), 87-91. https://doi.org/10.58304/
ijts.20230107
Teng, M. F. (2024). A systematic review of ChatGPT for English as a foreign language writing:
Opportunities, challenges, and recommendations. International Journal of TESOL Studies, 6(3),
36–57. https://doi.org/10.58304/ijts.20240304
Teng, M. F. (2025). Metacognitive awareness and EFL learners’ perceptions and experiences in utilising
ChatGPT for writing feedback. European Journal of Education, 60(1). https://doi.org/10.1111/
ejed.12811

Online First View

20 International Journal of TESOL Studies

Todd, R. W. (2025). Generative AI as a disrupter of language education. International Journal of TESOL

Studies, 1–9. https://doi.org/10.58304/ijts.250127
Ulla, M. B., & Teng, M. F. (2024). Generative artificial intelligence (AI) applications in TESOL:
Opportunities, issues, and perspectives. International Journal of TESOL Studies, 6(3), 1–3. https://
doi.org/10.58304/ijts.20240301
United Nations. (n.d.a). Goal 4: Ensure inclusive and equitable quality education and promote lifelong
learning opportunities for all. https://sdgs.un.org/goals/goal4?utm_source=chatgpt.com
United Nations. (n.d.b). Goal 9: Build resilient infrastructure, promote sustainable industrialization and
foster innovation. https://www.un.org/sustainabledevelopment/infrastructure-industrialization/
Uştuk, Ö., & De Costa, P. I. (2021). Reflection as meta-action: Lesson study and EFL teacher
professional development. TESOL Journal, 12(1), e00531. https://doi.org/10.1002/tesj.531
Ward, F., Cho, J., Jung, J. K., Tognocchi, C., Beadles, T., & Cheon, J. (2025). ChatGPT for the
intellectual soul: A Deweyan perspective on AI-based multicultural classroom praxis. Multicultural
Education Review, 17(1), 1–18. https://doi.org/10.1080/2005615X.2025.2467759
Wei, J., & Li, H. (2024). A systematic review of critical thinking development in information and
communication technology-supported english as a foreign language teaching from 2015 to 2024.
Forum for Linguistic Studies, 6(6), 990–1006. Scopus. https://doi.org/10.30564/fls.v6i6.7478
Xu, J., & Liu, Q. (2025). Uncurtaining windows of motivation, enjoyment, critical thinking, and
autonomy in AI-integrated education: Duolingo Vs. ChatGPT. Learning and Motivation, 89.
Scopus. https://doi.org/10.1016/j.lmot.2025.102100
Yang, S., Liu, Y., & Wu, T.-C. (2024a). ChatGPT, a new “Ghostwriter”: A teacher-and-students poetic
autoethnography from an EMI academic writing class. Digital Applied Linguistics, 1, 2244. https://
doi.org/10.29140/dal.v1.2244
Yang, S., Liu, Y., & Wu, T.-C. (2024b). ChatGPT, a new “Ghostwriter”: A teacher-and-students poetic
autoethnography from an EMI academic writing class. Digital Applied Linguistics, 1, 2244. https://
doi.org/10.29140/dal.v1.2244
Yusuf, A., Bello, S., Pervin, N., & Tukur, A. K. (2024). Implementing a proposed framework for
enhancing critical thinking skills in synthesizing AI-generated texts. Thinking Skills and Creativity,
53, 101619. https://doi.org/10.1016/j.tsc.2024.101619
Zou, X., Su, P., Li, L., & Fu, P. (2024). AI-generated content tools and students’ critical thinking: Insights
from a Chinese university. IFLA Journal, 50(2), 228–241. https://doi.org/10.1177/03400352231214963

Dr. Ngo Cong-Lem is a lecturer in the Faculty of Foreign Languages at Dalat University, Vietnam. He
holds a PhD from Monash University’s Faculty of Education, Australia, ranked 8th globally in Education
(2024–2025 U.S. News rankings). His research interests involve TESOL, the use of generative AI
in education, and influential psychological factors in learning and teaching such as critical thinking,
emotion, and agency. He has published in prestigious journals, such as Teaching and Teacher Education
and System (Elsevier), Language Learning & Technology, and the International Journal of Human-
Computer Interaction (Taylor & Francis). ORCID: https://orcid.org/0009-0005-3299-477X

Thang Tat Nguyen, Ph.D., is an Associate Professor at the Faculty of Foreign Languages, Dalat
University, Vietnam. His research interests involve cognitive linguistics, socio-linguistics and teaching
English as a second language. ORCID: https://orcid.org/0000-0001-8486-2385

Online First View

Ngo Cong-Lem, Thang Tat Nguyen and Khanh Nhat Hoang Nguyen 21

Khanh Nhat Hoang Nguyen, MA, is a Lecturer in English at the Center for Foreign Languages and
Human Resource Training at Dalat University. She teaches Academic Writing and TESOL methodology,
and conducts research in applied linguistics, language education, and teacher development in Vietnam.
Recently, she has investigated AI applications in English teaching, including adaptive learning platforms
and automated feedback systems. She also contributes to curriculum design for diverse learner
populations. Outside work, she reads professional literature and collaborates on innovative teaching
projects. ORCID: https://orcid.org/0000-0001-6074-9935

Online First View