Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 426
Percepciones estudiantiles sobre herramientas de IA
para el aprendizaje de inglés: oportunidades y
desafíos
University Students' Perceptions of AI-Powered Tools for English
Language Learning: Opportunities and Challenges
Montoya-Espinoza, Luisa María
1
Coloma-Chong, Johanna Katherine
2
https://orcid.org/0009-0009-6097-9621
https://orcid.org/0009-0000-5985-9803
lmontoya@uagraria.edu.ec
jcoloma@uagraria.edu.ec
Universidad Agraria del Ecuador, Ecuador,
Guayaquil.
Universidad Agraria del Ecuador, Ecuador,
Guayaquil.
Autor de correspondencia
1
DOI / URL: https://doi.org/10.55813/gaea/rcym/v4/n1/161
Resumen: La integración de herramientas de inteligencia
artificial en la educación superior transforma fundamentalmente
los procesos de enseñanza-aprendizaje, particularmente en la
adquisición de lenguas extranjeras. Este estudio investigó las
percepciones de estudiantes universitarios sobre herramientas
de inteligencia artificial para el aprendizaje del inglés,
centrándose en las oportunidades y desafíos desde las
perspectivas de los aprendices. Se empleó un diseño de métodos
mixtos descriptivo con sesenta estudiantes de nivel intermedio-
alto matriculados en dos secciones de Inglés 7 en la Universidad
Agraria del Ecuador. Los datos se recolectaron mediante un
cuestionario estructurado de treinta y dos ítems y doce
entrevistas semiestructuradas. Los resultados revelaron
percepciones generalmente positivas sobre las oportunidades
que ofrecen las herramientas de inteligencia artificial, con
puntuación media de cuatro puntos cero uno, destacando
disponibilidad continua, retroalimentación inmediata y apoyo
personalizado. Sin embargo, los participantes también
identificaron desafíos significativos, incluyendo preocupaciones
sobre dependencia excesiva, incertidumbre respecto al uso
apropiado y cuestiones de integridad académica. Los hallazgos
demuestran que los estudiantes poseen conciencia
metacognitiva sofisticada sobre las implicaciones del uso de
inteligencia artificial, reconociendo simultáneamente beneficios
genuinos y riesgos potenciales. El estudio contribuye evidencia
empírica que centra las voces estudiantiles en conversaciones
sobre integración de inteligencia artificial en educación
lingüística, sugiriendo la necesidad urgente de marcos
pedagógicos explícitos, directrices institucionales claras y
desarrollo de alfabetización crítica en inteligencia artificial.
Palabras clave: inteligencia artificial, aprendizaje de inglés,
percepciones estudiantiles, educación superior, tecnología
educativa.
Artículo Científico
Received: 13/Ene/2026
Accepted: 04/Feb/2026
Published: 26/Feb/2026
Cita: Montoya-Espinoza, L. M., & Coloma-
Chong, J. K. (2026). Percepciones
estudiantiles sobre herramientas de IA para el
aprendizaje de inglés: oportunidades y
desafíos. Revista Científica Ciencia Y
Método, 4(1), 426-
442. https://doi.org/10.55813/gaea/rcym/v4/n1
/161
Revista Científica Ciencia y Método (RCyM)
https://revistacym.com
revistacym@editorialgrupo-aea.com
info@editoriagrupo-aea.com
© 2026. Este artículo es un documento de
acceso abierto distribuido bajo los términos y
condiciones de la Licencia Creative
Commons, Atribución-NoComercial 4.0
Internacional.
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 427
Artículo Científico
Enero – Marzo 2026
Abstract:
The integration of artificial intelligence-powered tools in higher education
fundamentally transforms teaching and learning processes, particularly in foreign
language acquisition. This study investigated university students' perceptions of
artificial intelligence-powered tools for English language learning, focusing on
opportunities and challenges from learners' perspectives. A descriptive mixed-
methods design was employed with sixty upper-intermediate level students enrolled in
two sections of English 7 at Universidad Agraria del Ecuador. Data were collected
through a thirty-two item structured questionnaire and twelve semi-structured
interviews. Results revealed generally positive perceptions of opportunities provided
by artificial intelligence tools, with a mean score of four point zero one, highlighting
continuous availability, immediate feedback, and personalized support. However,
participants also identified significant challenges, including concerns about over-
reliance, uncertainty regarding appropriate use, and academic integrity issues.
Findings demonstrate that students possess sophisticated metacognitive awareness
about implications of artificial intelligence use, simultaneously recognizing genuine
benefits and potential risks. The study contributes empirical evidence centering student
voices in conversations about artificial intelligence integration in language education,
suggesting urgent need for explicit pedagogical frameworks, clear institutional
guidelines, and development of critical artificial intelligence literacy.
Keywords: artificial intelligence, English learning, student perceptions, higher
education, educational technology.
1. Introducción
The integration of artificial intelligence (AI) in higher education represents one of the
most transformative developments in contemporary pedagogy, fundamentally
reshaping how students’ access, process, and apply knowledge across disciplines
(Almuhanna, 2025; Cai et al., 2025). This technological revolution becomes particularly
significant in language education, where AI-powered tools offer unprecedented
opportunities to address longstanding pedagogical challenges, including limited
access to authentic language practice, personalized feedback, and individualized
learning pathways (Ali et al., 2025; Baba Khouya & Ismaili Alaoui, 2025). As
educational institutions worldwide increasingly adopt AI technologies, understanding
how students perceive and interact with these tools emerges as a critical factor in
determining their effective integration and educational impact (Sok et al., 2025).
Within the domain of English as a Foreign Language (EFL) instruction, AI-powered
applications demonstrate considerable potential to enhance multiple dimensions of
language acquisition. Recent research indicates that AI tools support writing
development through automated feedback and content generation (Muslimin et al.,
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 428
Artículo Científico
Enero – Marzo 2026
2024; Zhao, 2025), facilitate speaking practice via conversational agents and
pronunciation assessment systems (Nguyen Huu, 2025; Zou et al., 2025), and improve
reading comprehension through adaptive learning platforms (Jose, 2025). Beyond
skill-specific applications, AI technologies influence broader aspects of language
learning, including students' willingness to communicate (Zou et al., 2025), their
understanding of feedback processes (Y. et al., 2025), and their approaches to
collaborative writing (Nguyen et al., 2024). The diversity of available AI tools—from
general-purpose platforms like ChatGPT (Toosi, 2025; Yiğit et al., 2025) to specialized
applications such as Cami AI for writing instruction (Muslimin et al., 2024) and Microsoft
Teams Reading Progress for literacy development (Jose, 2025)—creates a complex
ecosystem that requires careful examination of user experiences and perceptions.
However, the rapid proliferation of AI technologies in EFL contexts outpaces empirical
understanding of how students experience, evaluate, and utilize these tools in their
language learning processes. While technological capabilities continue to advance,
student perceptions ultimately mediate the actual adoption, sustained engagement,
and learning outcomes associated with AI-powered applications (Ali et al., 2025;
Jamshed et al., 2024). Research across diverse educational contexts reveals varying
patterns of student attitudes, ranging from enthusiastic adoption and perceived
benefits to concerns about over-reliance, academic integrity, and diminished critical
thinking skills (Nwagbara, 2025; Yiğit et al., 2025). These divergent perspectives exist
even among students within the same institutions and programs (Babanoğlu et al.,
2025), suggesting that individual differences, prior experiences, and contextual factors
significantly shape how learners perceive AI's role in their education. Understanding
these perceptions becomes essential for educators seeking to leverage AI tools
effectively while addressing legitimate student concerns and optimizing learning
experiences (Waluyo & Rouaghe, 2025).
The current body of research on AI in EFL education demonstrates several notable
patterns and gaps. Studies examining teacher perspectives on AI integration reveal
cautious optimism tempered by concerns about implementation challenges and
pedagogical appropriateness (Ali et al., 2025; Almuhanna, 2025). Research on specific
AI applications provides evidence of effectiveness for particular skills or contexts (Jose,
2025; Muslimin et al., 2024; Nguyen Huu, 2025), yet comprehensive understanding of
student perceptions across multiple AI tools and learning contexts remains limited.
Furthermore, much existing research focuses on either technological capabilities or
learning outcomes, with insufficient attention to the student experience itself—how
learners perceive the opportunities AI tools create the challenges they encounter, and
the implications for their learning processes and academic development. As Isotalus
et al. (2025) note in their study of AI as a feedback provider, understanding user
perceptions proves crucial for successful technology integration, particularly when
tools assume roles traditionally filled by human instructors.
This study addresses this research gap by investigating university students'
perceptions of AI-powered tools for English language learning, with particular attention
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 429
Artículo Científico
Enero – Marzo 2026
to both the opportunities these technologies present and the challenges they pose from
learners' perspectives. The research seeks to provide comprehensive insight into how
students experience AI integration in their language learning, what benefits they
perceive, what concerns they harbour, and how these perceptions might inform more
effective pedagogical approaches to AI implementation in EFL contexts. By centering
student voices in the conversation about AI in language education, this study aims to
contribute empirical evidence that can guide educators, administrators, and technology
developers in creating more learner-responsive AI-enhanced language learning
environments.
2. Materiales y métodos
Research Design
This study employed a descriptive mixed-methods research design (Creswell & Creswell,
2017) to investigate university students' perceptions of AI-powered tools for English language
learning. The mixed-methods approach enabled comprehensive exploration of both the
breadth and depth of student experiences, combining quantitative data on perception patterns
with qualitative insights into the reasoning, contexts, and nuances underlying these
perceptions (Leavy, 2022). The research was conducted during the second academic
semester of 2024-2025 at Universidad Agraria del Ecuador, Balzar and Palestina campus.
Participants and Context
The study population consisted of 60 undergraduate students enrolled in two sections of
English 7, an upper-intermediate level course (B1+ according to the Common European
Framework of Reference) within the institutional English program at Universidad Agraria del
Ecuador. English 7 represents the seventh semester of required English instruction, with
students typically having completed six prior semesters of language study. Participants ranged
in age from 19 to 24 years (M = 21.3, SD = 1.4) and represented diverse academic programs
across agricultural sciences, engineering, economics, and related fields offered by the
institution.
Inclusion criteria required participants to be: (a) currently enrolled in English 7 during the study
period; (b) having completed at least five previous semesters of English instruction; (c) having
prior exposure to or experience with at least one AI-powered tool for language learning
purposes; and (d) willing to provide informed consent for participation. Exclusion criteria
eliminated students who: (a) had not used any AI tools for English learning purposes; (b) were
repeating the course for a third or subsequent time; or (c) withdrew from the course during the
data collection period.
Of the 62 students initially enrolled across both sections, 60 met the inclusion criteria and
agreed to participate, yielding a participation rate of 96.8%. Two students were excluded: one
had never utilized AI tools for language learning, and another withdrew from the course for
personal reasons before data collection commenced.
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 430
Artículo Científico
Enero – Marzo 2026
Data Collection Instruments
The study utilized two complementary data collection instruments designed to capture both
quantitative and qualitative dimensions of student perceptions.
Perception Questionnaire: A structured questionnaire consisting of 32 items organized into
four domains: (1) Awareness and Usage Patterns (8 items addressing frequency, types of AI
tools used, and purposes of use); (2) Perceived Opportunities (10 items examining perceived
benefits for language skills, learning processes, and academic outcomes); (3) Perceived
Challenges (10 items exploring concerns, difficulties, and limitations encountered); and (4)
Future Integration Preferences (4 items assessing willingness for continued use and
suggestions for improvement). Items employed a 5-point Likert scale (1 = Strongly Disagree
to 5 = Strongly Agree) for perception statements, along with multiple-choice and open-ended
questions for usage patterns and experiences.
The questionnaire was developed based on existing literature on AI in EFL education (Ali et
al., 2025; Jamshed et al., 2024; Sok et al., 2025) and piloted with 15 students from English 6
courses to ensure clarity, appropriate length, and content validity. Minor revisions to wording
and item order were made following pilot feedback. Internal consistency reliability for the final
instrument achieved Cronbach's alpha values of .87 for the Perceived Opportunities subscale
and .84 for the Perceived Challenges subscale, indicating good reliability.
Semi-structured Interviews: Following questionnaire administration, semi-structured interviews
were conducted with a purposively selected subsample of 12 participants (20% of total sample,
6 from each course section) to explore their perceptions in greater depth. Selection criteria
ensured diversity in usage patterns (high vs. low frequency users), perceived proficiency
levels, and initial questionnaire responses (positive, neutral, and critical perspectives).
Interview protocol consisted of eight open-ended questions addressing: specific AI tools used
and contexts of use; detailed descriptions of perceived benefits and learning advantages;
challenges, concerns, or negative experiences encountered; comparison with traditional
learning methods; impact on motivation and engagement; ethical considerations and academic
integrity awareness; and recommendations for effective AI integration in EFL instruction.
Interviews lasted 20-30 minutes each and were conducted in Spanish to ensure participant
comfort and expression depth, then transcribed verbatim.
Data Collection Procedures
Data collection followed a sequential process over six weeks during the academic semester.
In week one, the research team introduced the study to both English 7 sections, explained its
purpose and procedures, addressed questions, and obtained informed consent from willing
participants. Students received assurance that participation was voluntary, responses would
remain confidential, and their decisions would not affect course grades or academic standing.
During weeks two through three, participants completed the perception questionnaire during
regularly scheduled class time, requiring approximately 25-30 minutes. The questionnaire was
administered via Google Forms to facilitate data collection and organization. Participants
completed questionnaires individually without consultation, and the research team remained
present to address clarification questions while ensuring no influence on responses.
Weeks four through six involved conducting the 12 semi-structured interviews with selected
participants. Interviews took place in private, quiet locations on campus at times convenient
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 431
Artículo Científico
Enero – Marzo 2026
for participants. All interviews were audio-recorded with explicit permission and subsequently
transcribed for analysis. Participants were assigned pseudonyms to protect anonymity, and all
identifying information was removed from transcripts.
Data Analysis
Quantitative data from the questionnaire underwent analysis using SPSS version 28.
Descriptive statistics (frequencies, percentages, means, standard deviations) characterized
participant demographics, usage patterns, and perception distributions across items and
subscales. Independent samples t-tests compared perceptions between the two course
sections, and correlation analysis examined relationships between usage frequency and
perception variables. Statistical significance was set at p < .05.
Qualitative data from interview transcripts were analyzed using thematic analysis following the
iterative process of familiarization, initial coding, theme development, theme review, and theme
definition. Two researchers independently coded three randomly selected transcripts,
discussed discrepancies, and refined the coding framework to establish inter-rater reliability
(Cohen's kappa = .82). The remaining transcripts were then systematically coded using the
established framework. Themes emerging from qualitative analysis were used to
contextualize, explain, and enrich quantitative findings, providing deeper understanding of the
patterns observed in questionnaire data.
Ethical Considerations
The study protocol received approval from the Research Ethics Committee of Universidad
Agraria del Ecuador prior to participant recruitment or data collection (Approval Code: UAE-
CEI-2024-089). All participants provided written informed consent after receiving detailed
information about the study's purpose, procedures, voluntary nature, confidentiality
protections, and their right to withdraw at any time without penalty. No personally identifiable
information was collected beyond that necessary for research purposes, and all data were
stored securely with access limited to the research team. Participants received no
compensation for participation, ensuring voluntary involvement free from undue influence. The
study adhered to all institutional policies and national regulations governing research with
human participants.
3. Resultados
3.1. Demographic Characteristics and AI Tool Usage Patterns
The final sample comprised 60 undergraduate students from Universidad Agraria del
Ecuador, distributed across two English 7 course sections (Section A: n=30; Section
B: n=30). The sample included 34 female students (56.7%) and 26 male students
(43.3%), with ages ranging from 19 to 24 years (M=21.3, SD=1.4). Academic program
representation included Agricultural Engineering (n=18, 30%), Economics (n=15,
25%), Food Engineering (n=12, 20%), Environmental Engineering (n=9, 15%), and
Agroindustrial Engineering (n=6, 10%).
Regarding AI tool familiarity and usage, all 60 participants (100%) reported having
used at least one AI-powered tool for English learning purposes within the previous six
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 432
Artículo Científico
Enero – Marzo 2026
months. Table 1 presents the distribution of AI tools used by participants and their
reported frequency of use.
Table 1
AI Tools Used for English Language Learning and Frequency of Use (N=60)
Users
(n)
Percentage
Daily
Use
Weekly
Use
Monthly
Use
Occasional
58
96.7%
24
21
8
5
52
86.7%
31
15
4
2
38
63.3%
12
18
6
2
27
45.0%
8
11
5
3
19
31.7%
3
9
4
3
16
26.7%
2
7
4
3
14
23.3%
6
5
2
1
11
18.3%
4
4
2
1
8
13.3%
1
3
2
2
Note: Participants could report using multiple tools. Frequency categories are mutually exclusive for
each tool (Autors, 2026).
The most frequently utilized AI tool was ChatGPT (96.7%), followed by Google
Translate with AI features (86.7%) and Grammarly (63.3%). Participants reported
using AI tools primarily for writing assistance (n=56, 93.3%), translation support (n=54,
90%), grammar checking (n=49, 81.7%), vocabulary learning (n=43, 71.7%), and
speaking practice (n=28, 46.7%). The mean number of different AI tools used per
participant was 3.8 (SD=1.6, range=1-7).
3.2. Perceived Opportunities of AI-Powered Tools
Participants demonstrated generally positive perceptions regarding the opportunities
AI tools provide for English language learning. Table 2 summarizes descriptive
statistics for items assessing perceived opportunities across different dimensions.
Table 2
Descriptive Statistics for Perceived Opportunities (N=60)
Item
M
SD
Agreement
AI tools help me write more accurately in English
4.12
0.78
86.7%
AI tools provide immediate feedback on my language errors
4.23
0.71
90.0%
AI tools are available 24/7 when I need language support
4.45
0.68
93.3%
AI tools help me learn new vocabulary more effectively
3.98
0.84
80.0%
AI tools increase my confidence in using English
3.87
0.91
75.0%
AI tools help me understand grammar rules better
3.76
0.88
71.7%
AI tools save time when completing English assignments
4.18
0.75
85.0%
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 433
Artículo Científico
Enero – Marzo 2026
AI tools provide personalized learning support
3.65
0.96
68.3%
AI tools help me practice English outside of class
4.31
0.73
91.7%
AI tools make learning English more engaging
3.54
1.02
65.0%
Overall Perceived Opportunities Subscale
4.01
0.61
80.7%
Note: Scale: 1=Strongly Disagree to 5=Strongly Agree. *Agreement = percentage of participants
selecting 4 or 5 (Autors, 2026).
The highest-rated opportunity was 24/7 availability (M=4.45, SD=0.68), followed by
opportunities for practice outside class (M=4.31, SD=0.73) and immediate feedback
provision (M=4.23, SD=0.71). The overall perceived opportunities subscale achieved
a mean score of 4.01 (SD=0.61), indicating strong agreement with positive affordances
of AI tools. No statistically significant differences emerged between the two course
sections in overall perceived opportunities (t(58)=1.34, p=.185).
Participants identified specific learning domains where AI tools proved most beneficial.
Writing assistance emerged as the most valued application, with 51 participants (85%)
reporting significant improvements in writing quality, organization, and error reduction.
Vocabulary expansion represented another major benefit, with 47 participants (78.3%)
indicating that AI tools facilitated discovering contextually appropriate words and
expressions. Additionally, 44 participants (73.3%) valued AI tools for providing
explanations of grammar concepts and language rules in accessible formats.
3.3. Perceived Challenges and Concerns
Despite recognizing opportunities, participants also identified substantial challenges
and concerns regarding AI tool use for language learning. Table 3 presents descriptive
statistics for perceived challenges.
Table 3
Descriptive Statistics for Perceived Challenges (N=60)
Item
M
SD
Agreement*
I sometimes rely too much on AI tools instead of learning
independently
3.94
0.89
78.3%
AI tools sometimes provide incorrect or inappropriate
suggestions
3.76
0.85
71.7%
I worry about academic integrity when using AI tools
3.68
1.04
68.3%
AI tools may reduce my critical thinking skills
3.52
1.08
61.7%
I am unsure when it is appropriate to use AI tools for
coursework
3.47
1.02
58.3%
AI tools sometimes cannot understand the context of my
writing
3.41
0.94
56.7%
Using AI tools feels like cheating on assignments
2.98
1.15
41.7%
AI tools are too complex or difficult to use effectively
2.67
1.08
30.0%
I do not trust the information AI tools provide
2.54
1.01
25.0%
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 434
Artículo Científico
Enero – Marzo 2026
AI tools distract me from actual learning
2.43
1.06
21.7%
Overall Perceived Challenges Subscale
3.14
0.68
52.4%
Note: Scale: 1=Strongly Disagree to 5=Strongly Agree. *Agreement = percentage of participants
selecting 4 or 5 (Autors, 2026).
The most prominent challenge was over-reliance on AI tools (M=3.94, SD=0.89), with
78.3% of participants acknowledging this concern. Accuracy and appropriateness of
AI suggestions also emerged as significant issues (M=3.76, SD=0.85), followed by
academic integrity concerns (M=3.68, SD=1.04). The overall perceived challenges
subscale achieved a mean score of 3.14 (SD=0.68), indicating moderate agreement
with various concerns. No statistically significant differences existed between course
sections for overall perceived challenges (t(58)=0.87, p=.388).
Correlation analysis revealed a significant negative relationship between frequency of
AI tool use and concerns about difficulty of use (r=-.34, p=.008), suggesting that more
frequent users developed greater comfort and competence. However, frequency of use
showed a significant positive correlation with over-reliance concerns (r=.41, p=.001),
indicating that heavier users were more aware of dependency risks.
3.4. Preferences for Future AI Integration
Regarding future integration of AI tools in English language instruction, 54 participants
(90%) expressed willingness to continue using AI tools for language learning, while 49
participants (81.7%) supported formal integration of AI tools into course curriculum.
Participants identified several conditions they considered important for effective AI
integration: clear guidelines on appropriate use (n=57, 95%), teacher training on AI
tool capabilities and limitations (n=53, 88.3%), balanced approaches combining AI
tools with traditional methods (n=51, 85%), and explicit instruction on academic
integrity and ethical AI use (n=48, 80%).
When asked to rate their preference for different AI integration models, participants
favored a supplementary approach (n=42, 70%) where AI tools complement teacher
instruction rather than replacing it, compared to AI as the primary learning method
(n=3, 5%) or no AI integration (n=6, 10%). Nine participants (15%) expressed no strong
preference.
3.5. Qualitative Findings from Semi-Structured Interviews
Thematic analysis of the 12 semi-structured interviews revealed five major themes that
contextualized and elaborated upon quantitative findings.
3.5.1. AI Tools as Learning Facilitators and Confidence Builders
Participants consistently described AI tools as reducing anxiety associated with
language production, particularly in writing. One participant explained: "When I write
an essay, I feel more confident because I can check my grammar and vocabulary with
ChatGPT before submitting. It's like having a tutor available anytime" (Participant 3).
Another noted: "I used to be afraid of making mistakes, but now I can practice writing
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 435
Artículo Científico
Enero – Marzo 2026
without judgment from the AI tool" (Participant 7). This theme aligned with quantitative
findings showing that 75% of participants agreed AI tools increased their confidence.
3.5.2. Tension Between Assistance and Over-Dependence
Interview participants acknowledged a difficult balance between beneficial assistance
and problematic over-reliance. Participant 5 articulated this tension: "Sometimes I
know I should try to write something myself first, but it's so easy to just ask ChatGPT...
I worry I'm not really learning." Similarly, Participant 11 stated: "The problem is that AI
tools make everything so easy that I don't challenge myself anymore. I need to be more
disciplined." This theme corresponded with the quantitative finding that 78.3% of
participants recognized over-reliance as a concern.
3.5.3. Uncertainty About Ethical Boundaries
Multiple interview participants expressed confusion about what constitutes appropriate
versus inappropriate AI tool use for coursework. Participant 2 explained: "I'm never
sure how much help from AI is okay. Can I use it to improve my grammar? To get
ideas? To write a whole paragraph? Nobody has explained this clearly." Participant 8
echoed: "Different teachers have different rules about AI, so I don't know what's
allowed." This uncertainty manifested in the quantitative finding that 58.3% of
participants were unsure when AI tool use was appropriate for coursework.
3.5.4. Quality and Reliability Concerns
Participants reported experiences with incorrect, contextually inappropriate, or
culturally insensitive AI suggestions. Participant 4 described: "Sometimes ChatGPT
gives me formal vocabulary when I need informal or suggests expressions that sound
unnatural. I've learned I need to verify everything." Participant 9 noted: "I once used a
phrase the AI suggested, and my teacher said it was technically correct, but no native
speaker would say it that way." These experiences validated quantitative results
showing 71.7% of participants agreed that AI tools sometimes provide incorrect
suggestions.
3.5.5. Desire for Structured Integration with Pedagogical Guidance
Interview participants emphasized wanting explicit instruction on effective and ethical
AI use rather than informal, self-directed application. Participant 1 stated: "Teachers
should show us the best ways to use these tools for learning, not just tell us not to use
them." Participant 12 suggested: "It would be helpful if we had lessons about how to
use AI tools properly—like when to use them, when not to, and how to learn from them
instead of just copying." This theme supported quantitative findings showing 95% of
participants wanted clear guidelines on appropriate AI use and 88.3% desired teacher
training on AI tools.
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 436
Artículo Científico
Enero – Marzo 2026
3.5.6. Integration of Quantitative and Qualitative Results
The integration of quantitative and qualitative data revealed a complex, nuanced
picture of student perceptions. While quantitative data demonstrated strong agreement
with AI tools' benefits (M=4.01) and moderate concern about challenges (M=3.14),
qualitative data illuminated the contextual factors, emotional dimensions, and practical
experiences underlying these numerical patterns. The complementary findings
suggested that students recognize genuine learning opportunities in AI tools while
simultaneously grappling with uncertainty about appropriate use, concerns about over-
dependence, and desires for clearer pedagogical frameworks to guide their AI-
enhanced language learning.
4. Discusión
The findings of this study reveal a complex landscape of student perceptions regarding
AI-powered tools for English language learning, characterized by simultaneous
recognition of substantial opportunities and legitimate concerns. The high overall
perceived opportunities score (M=4.01) indicates that university students at
Universidad Agraria del Ecuador view AI tools as valuable resources for language
development, consistent with recent research documenting positive student attitudes
toward AI in EFL contexts (Ali et al., 2025; Sok et al., 2025). However, the moderate
perceived challenges score (M=3.14) suggests that enthusiasm for AI tools coexists
with critical awareness of potential drawbacks, reflecting the nuanced perspectives
documented in other studies examining student experiences with educational AI
(Jamshed et al., 2024; Yiğit et al., 2025).
Students' strong appreciation for 24/7 availability and immediate feedback aligns with
research demonstrating that accessibility and responsiveness constitute key
advantages of AI-enhanced language learning (Nguyen Huu, 2025; Y. et al., 2025).
The finding that 96.7% of participants utilize ChatGPT reflects broader patterns of
general-purpose AI adoption in educational settings (Toosi, 2025; Yiğit et al., 2025),
while substantial use of specialized tools like Grammarly (63.3%) and emerging
platforms such as Cami AI indicates diversification in students' AI tool repertoires. This
diversity supports Muslimin et al.'s (2024) observation that specialized AI applications
offer targeted support for specific language skills, though our findings suggest that
general-purpose tools like ChatGPT maintain dominant positions in students' actual
practice patterns.
The prominence of over-reliance concerns (78.3% agreement) represents a significant
finding that extends beyond previous research focused primarily on effectiveness or
satisfaction measures. This self-awareness regarding dependency risks mirrors
patterns observed in studies of AI use across diverse educational contexts (Nwagbara,
2025) and suggests that students possess metacognitive awareness of how AI tools
might influence their learning processes. The positive correlation between usage
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 437
Artículo Científico
Enero – Marzo 2026
frequency and over-reliance concerns (r=.41, p=.001) indicates that concerns emerge
through experience rather than speculation, supporting calls for explicit pedagogical
frameworks to guide appropriate AI integration (Waluyo & Rouaghe, 2025).
Students' uncertainty about appropriate AI use (58.3% experiencing confusion)
highlights a critical gap between technology availability and pedagogical guidance.
This finding resonates with Babanoğlu et al.'s (2025) documentation of diverse and
sometimes contradictory perspectives among prospective EFL teachers regarding AI
integration, suggesting that ambiguity exists at multiple levels of the educational
system. The qualitative finding that different instructors maintain different AI policies
without clear institutional guidelines indicates a need for coherent, well-articulated
frameworks that balance innovation with academic integrity, consistent with
recommendations emerging from research on AI-enhanced feedback processes (Y. et
al., 2025).
The study's mixed-methods approach reveals that quantitative patterns of perception
gain substantial explanatory depth from qualitative contextualization. For instance,
while 75% of participants agreed that AI tools increase confidence, interview data
illuminated the mechanisms through which this occurs—reducing performance
anxiety, enabling private practice without judgment, and facilitating iterative
improvement processes. Similarly, the 71.7% who identified accuracy concerns
provided specific examples of contextually inappropriate suggestions, culturally
insensitive outputs, and stylistically unnatural language that quantitative measures
alone would not capture. These complementary insights support the value of
integrating multiple data sources when examining educational technology adoption
(Creswell & Creswell, 2017; Leavy, 2022).
Students' overwhelming preference for supplementary rather than primary AI
integration (70% vs. 5%) aligns with pedagogical research emphasizing that
technology should augment rather than replace human instruction (Almuhanna, 2025;
Isotalus et al., 2025). This preference, combined with strong desire for explicit
guidelines (95%) and teacher training on AI capabilities (88.3%), suggests that
students recognize the need for informed mediation of AI-enhanced learning rather
than autonomous, unguided tool adoption. These findings parallel research
documenting the importance of teacher perspectives and competencies in successful
AI integration (Ali et al., 2025; Cai et al., 2025).
The study's context at Universidad Agraria del Ecuador, where English 7 students
possess upper-intermediate proficiency and substantial prior language learning
experience, suggests that the observed patterns characterize relatively advanced
learners who have developed metacognitive awareness of their learning processes.
Whether similar perception patterns exist among beginning-level students or learners
in different institutional contexts remains an empirical question requiring further
investigation. Additionally, the study's focus on two course sections, while yielding rich
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 438
Artículo Científico
Enero – Marzo 2026
data, limits generalizability across the broader spectrum of Ecuadorian higher
education or international EFL contexts.
These findings carry important implications for EFL pedagogy in AI-enhanced learning
environments. Educators should acknowledge that students already use AI tools
extensively and independently, making prohibition impractical and potentially
counterproductive. Instead, instruction should focus on developing critical AI literacy—
helping students understand tool capabilities and limitations, evaluate output quality,
recognize appropriate contexts for use, and integrate AI assistance into rather than
substituting for their own learning efforts. Institutional policies should provide clear,
consistent guidelines while allowing pedagogical flexibility, and professional
development should prepare instructors to model effective AI use and guide students
in developing sophisticated, ethically grounded approaches to AI-enhanced language
learning
5. Conclusiones
This study successfully achieved its objective of investigating university students'
perceptions of AI-powered tools for English language learning, documenting both the
opportunities these technologies present and the challenges they pose from learners'
perspectives. The research demonstrates that students at Universidad Agraria del
Ecuador possess sophisticated, nuanced views of AI tools that simultaneously
recognize their value for language development while maintaining critical awareness
of potential limitations and risks. This dual recognition—enthusiastic adoption coupled
with thoughtful caution—represents a more complex reality than simplistic narratives
of either technological determinism or reflexive resistance would suggest.
The findings reveal that AI tools have become integral to students' language learning
ecosystems, with universal adoption and diverse application across writing, translation,
grammar checking, and vocabulary development. Students value these tools primarily
for their accessibility, immediacy, and capacity to provide personalized support outside
traditional instructional contexts. However, this widespread adoption occurs largely
through informal, self-directed exploration rather than through pedagogically-guided
integration, creating a significant disconnect between students' actual AI use and
institutional acknowledgment or support for such practices.
A critical contribution of this research lies in documenting students' metacognitive
awareness of AI-related challenges, particularly over-reliance and uncertainty about
appropriate use. Rather than uncritically embracing AI tools, students demonstrate
reflexive understanding of how these technologies might influence their learning
processes, autonomy, and skill development. This awareness, however, does not
translate into clear action frameworks in the absence of explicit pedagogical guidance,
leaving students to navigate complex ethical and practical decisions about AI use
without adequate institutional support or consistent instructor direction.
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 439
Artículo Científico
Enero – Marzo 2026
The study contributes to the growing body of research on AI in language education by
centering student voices and experiences rather than focusing exclusively on
technological capabilities or learning outcomes. By employing a mixed-methods
approach, the research captures both the breadth of perception patterns across a
student cohort and the depth of individual experiences, revealing how quantitative
trends manifest in actual learning practices and decision-making processes. This
methodological approach proves particularly valuable for understanding emerging
technologies where lived experiences provide essential insights that complement
effectiveness studies.
The research findings carry significant implications for EFL pedagogy and institutional
policy in AI-enhanced learning environments. Educational institutions can no longer
treat AI tools as future possibilities or optional supplements; students already use these
technologies extensively, making the relevant question not whether to integrate AI but
how to do so thoughtfully and effectively. This reality necessitates explicit curricula for
developing critical AI literacy, clear institutional policies that balance innovation with
academic integrity, and professional development that prepares instructors to guide
rather than prohibit student AI use.
The strong student preference for supplementary rather than primary AI integration
provides important guidance for pedagogical design. Rather than replacing human
instruction or traditional learning methods, AI tools should complement and enhance
existing approaches, with clear delineation of when, how, and why particular tools
serve specific learning objectives. This balanced approach respects both the genuine
affordances AI technologies offer and the irreplaceable dimensions of human
interaction, cultural understanding, and authentic communication that remain central
to language education.
The study also highlights the urgency of developing coherent, consistent frameworks
for AI use in academic contexts. The confusion students experience regarding
appropriate boundaries reflects broader institutional uncertainty about how to respond
to rapidly evolving technologies. Clear guidelines—developed through inclusive
dialogue among students, instructors, administrators, and educational technology
specialists—can transform this uncertainty into productive engagement, enabling
students to leverage AI tools for learning enhancement while maintaining academic
integrity and developing independent competencies.
Finally, this research underscores that effective AI integration in language education
requires ongoing investigation and responsive adaptation rather than one-time
implementation. As AI technologies continue evolving and as students develop
increasingly sophisticated usage patterns, educational institutions must maintain
commitment to understanding learner perspectives, evaluating pedagogical
approaches, and refining policies based on empirical evidence. The present study
provides a foundational understanding of student perceptions at a specific moment in
AI's educational trajectory, establishing a baseline for future research examining how
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 440
Artículo Científico
Enero – Marzo 2026
these perceptions and practices evolve as both technologies and pedagogical
frameworks mature. Future investigations should explore how AI literacy instruction
influences student practices, how different instructional contexts shape AI integration
patterns, and how students' relationships with AI tools develop across their language
learning trajectories from beginning to advanced proficiency levels.
CONFLICTO DE INTERESES
“Los autores declaran no tener ningún conflicto de intereses”.
Referencias Bibliográficas
Ali, M. M., Fazal, S., Ubaid, U. U., Saeed, M. A., & Shah, S. R. (2025). Integrating AI-
Powered Digital English Textbook at Grade12: Pakistani Students and
Teachers' Perceptions. International Journal of Adult Education and
Technology (IJAET), 16(1), 1-20. https://doi.org/10.4018/IJAET.376488
Ali, Muhammad Mooneeb, et al., editors. AI Applications for English Language
Learning. IGI Global Scientific Publishing, 2025. https://doi.org/10.4018/979-8-
3693-9077-1
Almuhanna, M.A. Teachers' perspectives of integrating AI-powered technologies in K-
12 education for creating customized learning materials and resources. Educ
Inf Technol 30, 10343–10371 (2025). https://doi.org/10.1007/s10639-024-
13257-y
Baba Khouya, Y. & Ismaili Alaoui, A. (Eds.). (2025). Application of AI in the Teaching
and Learning of English as a Foreign Language (EFL). IGI Global Scientific
Publishing. https://doi.org/10.4018/979-8-3373-0969-9
Babanoğlu, M. P., Karataş, T. Ö., & Dündar, E. (2025). Envisioning the Future of AI-
Assisted EFL Teaching and Learning: Conceptual Representations of
Prospective Teachers. Sage Open, 15(2).
https://doi.org/10.1177/21582440251341590 (Original work published 2025)
Cai, H., Lu, L., Han, B. et al. Exploring pre-service teachers’ reflection mediated by an
AI-powered teacher dashboard in video-based professional learning: a pilot
study. Education Tech Research Dev 73, 1129–1154 (2025).
https://doi.org/10.1007/s11423-024-10442-1
Creswell, J. W. & Creswell, J. D. 2017. Research design: Qualitative, quantitative, and
mixed methods approaches.
https://books.google.com/books?hl=en&lr=&id=335ZDwAAQBAJ&oi=fnd&pg=
PT16&dq=research+design+qualitative,+quantitative,+and+mixed+methods+a
pproaches&ots=YExRHPtqtF&sig=lwe8HUK3_TjZn1rz5u8WNbWJyjo
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 441
Artículo Científico
Enero – Marzo 2026
Isotalus, P., Eklund, M., & Karppinen, K. (2025). Artificial intelligence as a feedback
provider in practicing public speaking. Communication Teacher, 39(1), 78–85.
https://doi.org/10.1080/17404622.2024.2407910
Jamshed, M., Alam, I., Sultan, S. A., & Banu, S. (2024). Using artificial intelligence for
English language learning: Saudi EFL learners’ opinions, attitudes and
challenges. Journal of Education and E-Learning Research, 11(1), 135–141.
https://doi.org/10.20448/jeelr.v11i1.5397
Jose, J. The impact of integrating Microsoft Teams – Reading Progress as an Artificial
Intelligence (AI) platform for promoting learners’ reading aloud skills. Educ Inf
Technol 30, 7077–7115 (2025). https://doi.org/10.1007/s10639-024-13074-3
Leavy, P. 2022. Research design: Quantitative, qualitative, mixed methods, arts-
based, and community-based participatory research approaches.
https://books.google.com/books?hl=en&lr=&id=qUiKEAAAQBAJ&oi=fnd&pg=P
P1&dq=research+design+qualitative,+quantitative,+and+mixed+methods+app
roaches&ots=RxOcDqCqbJ&sig=QUeoVUAX4udYREYA30_Anct8cIA
Muslimin, A. I., Nur, M., & Francisca Maria, I. (2024). Evaluating Cami AI Across SAMR
Stages: Students’ Achievement and Perceptions in EFL Writing Instruction:
Cami AI-SAMR in EFL Writing Instruction. Online Learning, 28(2).
https://doi.org/10.24059/olj.v28i2.4246
Nguyen Huu, H. (2025). AI for English speaking practice: A study of effectiveness and
engagement among Vietnamese university learners. Australasian Journal of
Educational Technology, 41(3), 45–61. https://doi.org/10.14742/ajet.10050
Nguyen, A., Hong, Y., Dang, B., & Huang, X. (2024). Human-AI collaboration patterns
in AI-assisted academic writing. Studies in Higher Education, 49(5), 847–864.
https://doi.org/10.1080/03075079.2024.2323593
Nwagbara, U.U. From Curiosity to Dependency: Nigerian Students’ Perspectives on
AI Integration in Academic Research. J Acad Ethics 23, 2051–2068 (2025).
https://doi.org/10.1007/s10805-025-09641-z
Sok, S., Heng, K., & Pum, M. (2025). Investigating High School Students’ Attitudes
Toward the Use of AI in Education: Evidence from Cambodia. Sage Open,
15(3). https://doi.org/10.1177/21582440251353575
Toosi, M. (2025). Using ChatGPT as a Brainstorming Tool in Writing. i-manager’s
Journal on English Language Teaching, 15(4), 1-7.
Waluyo, B., & Rouaghe, F. (2025). Beyond Teacher-Led Approaches: Student-Initiated
Translanguaging With Artificial Intelligence Tools in Foreign Language
Acquisition. Sage
Open, 15(3). https://doi.org/10.1177/21582440251362998 (Original work
published 2025)
Y., Gasevic, D., Chen, G., Charwat, N., Chan, P. W. K., Martinez-Maldonado, R.,
Gašević, D., & Tsai, Y.-S. (2025). Students’ Perceptions of Generative AI-
Revista Científica Ciencia y Método | Vol.04 | Núm.01 | Ene–Mar | 2026 | www.revistacym.com pág. 442
Artículo Científico
Enero – Marzo 2026
Powered Learning Analytics in the Feedback Process: A Feedback Literacy
Perspective. Journal of Learning Analytics, 12(1), 152-168.
https://doi.org/10.18608/jla.2025.8609
Yiğit, S., Berşe, S., Dirgar, E., & Gülhan Güner, S. (2025). Views of health sciences
undergraduates on ChatGPT, an artificial intelligence-powered language model:
A qualitative study. Innovations in Education and Teaching International, 62(4),
1258–1272. https://doi.org/10.1080/14703297.2024.2391044
Zhao, D. The impact of AI-enhanced natural language processing tools on writing
proficiency: an analysis of language precision, content summarization, and
creative writing facilitation. Educ Inf Technol 30, 8055–8086 (2025).
https://doi.org/10.1007/s10639-024-13145-5
Zou, Bin, et al. "Students' Willingness to Communicate (WTC) in Using Artificial
Intelligence (AI) Technology in English-Speaking Practice." IJICTE vol.21, no.1
2025: pp.1-18. https://doi.org/10.4018/IJICTE.375387