INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8529

www.rsisinternational.org

Artificial Intelligence in the English Language Education: A Study among

University Students

Hongli Feng

School of Foreign Languages, Ningxia Medical University, Ningxia 750004, China

DOI:

https://dx.doi.org/10.47772/IJRISS.2025.903SEDU0645

Received: 17 October 2025; Accepted: 24 October 2025; Published: 13 November 2025

ABSTRACT

The integration of Artificial Intelligence (AI) into English language education has fundamentally transformed

pedagogical practices and student engagement, particularly within higher education contexts, where diverse

proficiency levels and large class sizes present ongoing instructional challenges. This study systematically

investigates the application, efficacy, and implications of AI technologies in supporting English learning among

university students. Employing a mixed-methods research design, data were collected from 120 undergraduates

across three Chinese universities through structured surveys, classroom observations, and semi-structured

interviews, enabling a comprehensive exploration of both quantitative outcomes and qualitative experiences.

The findings reveal that AI tools—including intelligent writing assistants, adaptive vocabulary applications, and

automated feedback systems—substantially enhance learners’ writing accuracy, lexical sophistication, and

overall motivation, while simultaneously promoting greater autonomy and self-regulated learning behaviors.

Students reported that AI-supported activities, when integrated with traditional instruction, facilitated more

active engagement and personalized learning trajectories. Nevertheless, challenges emerged, notably the

potential for overreliance on technology, ethical concerns regarding data privacy, and disparities in digital

literacy across learners. The study underscores that AI can serve as both a supportive and transformative agent

in English language education when its deployment is guided by clear pedagogical frameworks, ethical standards,

and institutional oversight. These findings contribute to ongoing discourse on technology-enhanced language

learning (TELL) and provide actionable recommendations for educators and policymakers seeking to harness

AI to foster sustainable, student-centered innovation in university-level English instruction.

Keywords: Artificial Intelligence, English Language Education, University Students, Technology-Enhanced

Learning, EFL

INTRODUCTION

In the twenty-first century, English language education has experienced profound transformation, driven by rapid

advancements in digital technologies and the proliferation of artificial intelligence (AI). The integration of

intelligent systems into educational contexts has redefined how learners access, process, and produce linguistic

knowledge, fundamentally reshaping pedagogical paradigms and classroom dynamics. In English as a Foreign

Language (EFL) environments, particularly at the university level, AI-powered technologies have introduced

novel instructional modalities that challenge conventional teaching and learning models. Tools such as

automated essay scoring, intelligent tutoring systems, adaptive learning platforms, and conversational chatbots

have become pivotal agents of innovation, offering scalable solutions to address diverse learner needs, large

class sizes, and the demand for personalized feedback. This emerging landscape not only transforms classroom

practices but also prompts critical questions regarding human–machine interaction, teacher roles, and learner

autonomy in technologically mediated learning environments.

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8530

www.rsisinternational.org

Over the past decade, scholarly interest in the intersection of AI and English language education has grown

substantially. Numerous studies have examined the capacity of AI to support personalized learning, enhance

assessment accuracy, and improve learner motivation (Li, 2023; Nguyen & Zhang, 2021; Wang & Chen, 2022).

Leveraging natural language processing, machine learning algorithms, and generative AI systems, educators can

now provide real-time analysis of students’ writing, pronunciation, and grammar—delivering immediate,

individualized feedback that was previously unattainable in traditional classroom settings. The efficiency,

adaptability, and scalability of these technologies have prompted educators and policymakers to reconsider the

complementary role of AI in facilitating human instruction. Nonetheless, despite the enthusiasm for

technological integration, critical concerns persist, including potential overreliance on AI, data privacy, uneven

digital literacy, and ethical implications of machine-mediated assessment (Hockly, 2022; Sun, 2024).

Within higher education, university students constitute a particularly relevant population for investigating the

impact of AI on English learning. As digital natives, they are generally familiar with online platforms, mobile

applications, and blended learning modalities. However, their experiences with AI-assisted English learning are

heterogeneous. Some students embrace AI tools—such as Grammarly, ChatGPT, and intelligent translation

systems—as mechanisms to enhance writing proficiency, expand lexical resources, and facilitate autonomous

learning, whereas others remain skeptical regarding the reliability, contextual appropriateness, and authenticity

of machine-generated feedback. Similarly, instructors demonstrate a spectrum of attitudes, ranging from

optimism regarding AI’s pedagogical potential to caution over unintended dependency and diminished critical

thinking. Understanding how university students interact with AI in English education is therefore crucial to

identifying both opportunities for innovation and challenges to sustainable implementation.

The global shift toward technology-enhanced education accelerated markedly during and after the COVID-19

pandemic, when online learning platforms became central to academic continuity. Universities in China,

alongside institutions worldwide, rapidly adopted AI-assisted platforms for English instruction and assessment,

intensifying reliance on intelligent systems to support language learning. This transition provided fertile ground

for empirical research into how AI influences learner motivation, performance, and perceptions in EFL contexts.

Despite the growing literature on technology-enhanced language learning (TELL), empirical studies focused

specifically on university-level English education in non-native contexts remain limited, underscoring the need

for research that bridges theoretical promise with classroom realities.

This study seeks to address this gap by examining the implementation and effects of AI technologies in English

language education among university students. It investigates the ways in which AI tools are utilized, how

students perceive their usefulness, and the pedagogical outcomes arising from their integration. By exploring the

interaction between AI-supported learning and traditional instruction, the research aims to elucidate the balance

between technological efficiency and human mediation, highlighting the conditions under which AI can

optimally support learning.

The significance of this study is threefold. First, it contributes empirical evidence to the growing body of research

on AI-driven language pedagogy, emphasizing university students’ lived experiences rather than theoretical

projections. Second, it illuminates pedagogical implications for integrating AI into English instruction, offering

actionable recommendations for educators seeking to enhance engagement and learning outcomes through

technology. Third, it addresses ethical and practical considerations inherent in AI adoption, including issues of

equity, data security, and digital competence. By examining these multiple dimensions, the study advances

discourse on the ethical, effective, and sustainable deployment of AI in higher education language classrooms.

Accordingly, the research is guided by four primary questions: (1) How are AI technologies currently applied in

university-level English education? (2) What are university students’ perceptions of AI-assisted English learning?

(3) What effects does AI use have on students’ motivation, autonomy, and language performance? (4) What

challenges and limitations are associated with AI integration in English teaching and learning? To address these

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8531

www.rsisinternational.org

questions, the study adopts a mixed-methods design, combining quantitative survey data with qualitative insights

from classroom observations and semi-structured interviews. This approach enables a holistic understanding of

both measurable learning outcomes and subjective experiences, while also allowing for data triangulation to

ensure robustness and depth of interpretation. The findings are intended to inform educators, administrators, and

policymakers regarding best practices in AI integration, while preserving the centrality of human pedagogy.

In summary, the integration of AI into English language education represents a pivotal moment in the evolution

of language teaching and learning. AI provides unprecedented opportunities for personalized instruction,

accessibility, and efficiency, yet it simultaneously challenges educators to reconceptualize their roles,

responsibilities, and instructional strategies in a technologically mediated environment. This study positions

itself at the intersection of innovation and pedagogy, aiming to explore how university students experience,

evaluate, and benefit from AI-based tools in their English learning journeys. The insights gained are expected to

contribute to the broader discourse on sustainable, ethical, and learner-centered approaches to AI integration in

higher education.

LITERATURE REVIEW

Technology in English Language Education

Technology has long played a transformative role in language teaching and learning, evolving from early

Computer-Assisted Language Learning (CALL) in the 1980s to today’s sophisticated AI-driven systems. CALL

initially introduced digital tools such as grammar drills, vocabulary software, and multimedia resources,

enhancing learners’ exposure to authentic linguistic input and providing structured opportunities for practice

(Chapelle, 2019). The subsequent development of Mobile-Assisted Language Learning (MALL) extended

learning beyond the confines of the classroom, enabling students to engage with English anytime and anywhere

through smartphones and tablets (Kukulska-Hulme & Shield, 2018). While earlier technologies primarily

emphasized content delivery and repetitive practice, recent advances in AI have enabled adaptive learning,

natural language interaction, and data-informed feedback, thereby shifting the focus from static instruction to

dynamic, personalized learning. AI systems can analyze learner behavior, predict difficulties, and tailor

instruction in real time—capabilities that traditional CALL systems could not achieve—positioning AI not

merely as an incremental improvement but as a paradigm shift in English language education.

Artificial Intelligence in Second Language Acquisition (SLA)

Artificial Intelligence, broadly defined as computer systems capable of performing cognitive tasks traditionally

associated with human intelligence, including reasoning, learning, and problem-solving, has significant

implications for Second Language Acquisition (SLA). In SLA contexts, AI supports both instruction and

assessment through automation, personalization, and real-time feedback. AI technologies leverage algorithms

that adapt to individual learners’ styles and proficiency levels, thereby delivering differentiated instruction

aligned with learners’ specific needs (Li, 2023). Contemporary applications encompass AI writing assistants

such as Grammarly and Write & Improve, which provide instant feedback on grammar, vocabulary, and style;

speech recognition systems, including ELSA Speak, which evaluate pronunciation and fluency; and

conversational chatbots, such as Duolingo Bots and Replika, which simulate authentic communicative

exchanges. These applications are grounded in Natural Language Processing (NLP) and Machine Learning (ML),

enabling computers to comprehend and generate human-like language (Godwin-Jones, 2022). The pedagogical

value of AI lies in its capacity to foster learner autonomy, promote formative assessment, and enhance

metalinguistic awareness. Research demonstrates that AI-assisted feedback facilitates self-correction and

reflection, while adaptive systems dynamically adjust difficulty levels to sustain engagement and challenge

learners appropriately (Zhang & Hyland, 2020). Nonetheless, persistent concerns include algorithmic bias, data

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8532

www.rsisinternational.org

security, and the interpretability of AI-generated feedback, all of which can influence learner trust, engagement,

and educational validity (Hockly, 2022).

Theoretical Frameworks: Constructivism, Sociocultural Theory, And Call

AI applications in English education can be broadly categorized into four domains: writing enhancement,

speaking and pronunciation training, intelligent tutoring and chatbots, and learning analytics. In writing

enhancement, AI-powered tools such as Grammarly, QuillBot, and ChatGPT provide instantaneous feedback on

grammar, coherence, and style. Empirical evidence indicates that these tools can improve writing accuracy and

linguistic complexity when employed as supplementary aids (Zhang, 2023), though excessive reliance may

inhibit students’ self-editing capacity and critical thinking (Lee, 2022). In speaking and pronunciation training,

speech recognition and analysis software—including ELSA Speak and Google Speech APIs—offers corrective

feedback that enhances phonological awareness and speaking confidence (Wang & Chen, 2022). Intelligent

tutoring systems and AI chatbots, exemplified by Duolingo’s GPT-based conversational bots, create personalized

learning pathways by adapting exercises to learners’ performance, reinforcing engagement through gamification

and instant feedback, and providing simulated communicative contexts (Sun & Xu, 2024). Finally, learning

analytics enables instructors to process large volumes of learner data to track progress, predict outcomes, and

identify areas for targeted intervention (Godwin-Jones, 2022). While these applications provide substantial

pedagogical benefits, they also raise ethical concerns related to privacy, transparency, and consent.

AI Applications in English Education

At the university level, AI adoption offers multiple educational advantages. Personalization allows learners to

progress at an optimal pace, addressing heterogeneous proficiency levels within large classes. Immediate

feedback supports formative assessment and self-regulated learning, contributing to improved writing quality,

vocabulary retention, and overall language competence (Nguyen & Zhang, 2021). AI-enhanced environments

also sustain learner motivation and engagement through interactive, gamified elements that align with digital-

native habits (Li, 2023). Beyond cognitive gains, AI fosters autonomy and lifelong learning skills by encouraging

goal setting, progress monitoring, and reflective practice, which are essential for academic and professional

English communication. Instructors benefit as well, leveraging AI-generated analytics to design targeted

interventions and optimize pedagogical strategies.

Benefits of AI Integration in University English Learning

At the university level, the adoption of AI in English learning offers multiple benefits.

First, personalization allows each student to progress at an optimal pace, addressing diverse proficiency levels

within large classes. Second, immediate feedback supports formative assessment and self-regulated learning,

leading to improved writing quality and vocabulary retention (Nguyen & Zhang, 2021). Third, increased

motivation and engagement are frequently reported outcomes of AI-assisted learning environments (Li, 2023).

AI tools’ interactive and gamified elements align with students’ digital habits, sustaining attention and effort.

Moreover, AI fosters autonomy and lifelong learning skills by encouraging learners to set goals, monitor progress,

and reflect on performance. These competencies are critical for university students preparing for academic and

professional communication in English. Teachers benefit as well, using AI-generated analytics to design more

effective instruction and provide targeted support.

Challenges and Ethical Considerations

Despite its promise, AI integration presents notable challenges. Overreliance on technology may lead students

to prioritize machine feedback over teacher guidance, potentially undermining critical thinking and interpersonal

interaction (Hockly, 2022). Digital inequality also remains a concern, as access to advanced AI tools and requisite

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8533

www.rsisinternational.org

digital literacy is uneven among students and institutions (Sun, 2024). Ethical considerations are central, as AI

systems collect sensitive learner data, raising issues of privacy, consent, and algorithmic bias (Floridi & Cowls,

2019). Adaptive learning algorithms may inadvertently privilege certain linguistic norms, reinforcing inequities.

To mitigate these risks, institutions must establish clear data governance policies, ensure algorithmic

transparency, and provide robust teacher oversight. Furthermore, the teacher’s role is redefined in AI-assisted

classrooms, emphasizing facilitation, interpretation of AI outputs, contextualization, and emotional and cultural

support. Professional development for instructors is essential to ensure that AI is integrated critically and

pedagogically effectively.

Summary of Research Gap

While existing literature affirms the pedagogical benefits of AI in English language learning, most studies focus

on technical development rather than learners’ lived experiences, particularly in higher education. Many

investigations are limited to experimental or small-scale settings, restricting generalizability. There is a pressing

need for empirical research that examines how university students perceive, interact with, and benefit from AI-

assisted English learning in authentic academic contexts. This study addresses this gap by employing a mixed-

methods approach to analyze both quantitative outcomes and qualitative perceptions, bridging theoretical

understanding with practical implementation. By doing so, it contributes to the discourse on sustainable, human-

centered integration of AI in university-level language education.

METHODOLOGY

Research Design

This study employed a mixed-methods research design that strategically combined quantitative and qualitative

approaches to achieve a comprehensive understanding of the application and impact of artificial intelligence (AI)

technologies in English language education among university students. The mixed-methods framework was

selected to capture both measurable learning outcomes and the nuanced perceptions of students and instructors

regarding AI integration, thereby addressing the complexity of technology-mediated learning environments.

Quantitative data were collected through structured questionnaires, which enabled the identification of patterns

in AI usage, student motivation, and perceived effectiveness of AI-assisted learning tools. Concurrently,

qualitative data were obtained via semi-structured interviews and classroom observations, providing rich,

contextual insights into the lived experiences, attitudes, and challenges associated with AI-supported English

instruction.

Following Creswell and Plano Clark’s (2018) convergent parallel design, both quantitative and qualitative data

were gathered concurrently, analyzed independently, and subsequently integrated during the interpretation phase.

This approach ensured methodological triangulation, thereby enhancing the validity and reliability of the

findings by cross-verifying trends observed in survey data with in-depth qualitative narratives. By employing

this design, the study was able to investigate not only the outcomes of AI use in English language learning—

such as improvements in writing, vocabulary acquisition, and learner engagement—but also the underlying

processes and mechanisms that contribute to these outcomes within authentic university classroom settings.

Ultimately, this integrative approach provided a holistic perspective on the pedagogical affordances and

limitations of AI in higher education English instruction, offering insights into both efficacy and implementation

strategies.

Research Context

The research was conducted across three comprehensive universities situated in eastern and southern China, each

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8534

www.rsisinternational.org

of which offers undergraduate English courses incorporating digital learning platforms. These institutions were

purposively selected due to their proactive engagement in educational technology innovation, as well as the

diversity of their student populations, which provided a representative sample for examining AI-assisted English

learning. The study was carried out during the 2024–2025 academic year, a period marked by widespread

integration of AI-enhanced systems into blended English curricula, reflecting the post-pandemic expansion of

digital education across higher education in China.

English courses at these universities typically encompassed writing, speaking, listening, and reading modules,

delivered through a hybrid model that combined traditional classroom instruction with online learning platforms,

such as iWrite, Pigai, and AI-powered vocabulary applications. Institutional policies actively encouraged

instructors to experiment with AI-based feedback and assessment tools, fostering an environment conducive to

pedagogical innovation and technology-mediated learning. This setting provided an authentic and contextually

rich framework for examining the practical implementation, effectiveness, and challenges of AI integration in

higher education English instruction, allowing the study to capture both student engagement and instructional

dynamics within real-world educational environments.

Participants

A total of 120 undergraduate students and six English instructors participated in this study. Participants were

selected through purposive sampling to ensure representation across diverse academic majors, varying English

proficiency levels, and heterogeneous technological backgrounds, thereby enabling a comprehensive exploration

of AI-assisted English learning in higher education contexts. Among the student cohort, 45 were male and 75

were female, with ages ranging from 18 to 22 years. They represented a variety of disciplines, including English,

Business, Engineering, and Education. All student participants had completed at least two semesters of

university-level English courses and possessed prior experience using digital learning tools. However, their

familiarity with AI-based applications varied considerably, ranging from frequent users of platforms such as

Grammarly and ChatGPT to those with no prior exposure to intelligent feedback systems.

The instructor participants were all experienced English educators, with an average of 9.3 years of teaching

experience, who had actively integrated AI or digital platforms into their English instruction. Their inclusion

provided a crucial pedagogical perspective, contextualizing student responses and offering insights into the

practical challenges, implementation strategies, and potential benefits of AI in the language classroom.

Participation in the study was voluntary, and informed consent was obtained from all respondents. Additionally,

the study received formal ethical approval from the university’s research ethics committee prior to the

commencement of data collection, ensuring that all research procedures adhered to established ethical standards

for human subjects research.

Instruments

Three instruments were employed to collect data in this study: a structured questionnaire, semi-structured

interviews, and classroom observations. The combination of these tools facilitated methodological triangulation,

allowing the researcher to integrate quantitative and qualitative evidence and thereby enhance the validity and

comprehensiveness of the findings.

The structured questionnaire, which comprised 36 items, was divided into four sections to capture multiple

dimensions of AI-assisted English learning. The first section collected demographic information, including age,

gender, academic major, English proficiency level, and prior experience with AI-based learning applications.

The second section explored participants’ usage patterns of AI tools, examining the frequency of use, types of

applications employed, and specific learning purposes. The third section assessed perceived effectiveness and

motivation through Likert-scale items, measuring constructs such as usefulness, satisfaction, and engagement.

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8535

www.rsisinternational.org

The final section addressed challenges and concerns, encompassing ethical, technical, and pedagogical issues

associated with AI integration. Likert-scale items were scored on a five-point scale, ranging from 1 = Strongly

Disagree to 5 = Strongly Agree. Sample items included statements such as “AI feedback helps me improve my

writing accuracy,” “Using AI tools increases my motivation to learn English,” and “I rely too much on AI

assistance when completing English assignments.”

The questionnaire was adapted from validated scales used in previous CALL and AI-in-education research (e.g.,

Li, 2023; Zhang & Hyland, 2020) to ensure both relevance and reliability. Prior to formal administration, a pilot

study was conducted with 25 students, yielding a Cronbach’s alpha coefficient of 0.89, which indicated high

internal consistency and reliability. By systematically capturing demographic profiles, behavioral patterns,

perceptions, and challenges, the questionnaire provided robust quantitative data to complement the qualitative

insights derived from interviews and classroom observations.

Semi-Structured Interviews

To gain deeper insights into the experiences and perceptions of AI-assisted English learning, follow-up semi-

structured interviews were conducted with a purposively selected subset of 15 students and three instructors.

Stratified sampling was employed to ensure representation across varying English proficiency levels and

differing frequencies of AI tool usage, thereby capturing a comprehensive range of learner and teacher

perspectives. Each interview lasted approximately 30 minutes and was conducted in either English or Mandarin,

according to participant preference and comfort, to facilitate candid and nuanced responses. The interview

protocol addressed multiple domains, including students’ personal experiences and attitudes toward AI-assisted

learning, their perceptions of the strengths and limitations of AI-generated feedback, and instructors’ views on

classroom integration, assessment practices, and student engagement. Additionally, participants were invited to

reflect on ethical and pedagogical considerations associated with AI use, such as data privacy, algorithmic bias,

and the balance between technology and human instruction. All interviews were audio-recorded, transcribed

verbatim, and translated when necessary to maintain accuracy and fidelity of meaning. To ensure confidentiality

and protect participant identity, pseudonyms were employed throughout data analysis and reporting. This

qualitative approach allowed the study to contextualize and deepen understanding of the quantitative findings,

providing rich, interpretive data on the lived realities of AI integration in higher education English instruction.

Classroom Observations

Non-participant classroom observations were conducted in a total of six English courses, with two classes

selected from each of the three participating universities, over a continuous four-week period. The observation

protocols were designed to systematically capture multiple dimensions of instructional practice, including

teacher–student interactions, patterns of AI tool usage, and observable student engagement behaviors. Detailed

field notes were recorded on how AI systems were integrated into instructional activities, such as during writing

workshops employing Grammarly or automated essay evaluation platforms, as well as during speaking and

vocabulary exercises facilitated by AI applications. Particular attention was given to the ways in which

instructors scaffolded AI-supported learning, how students responded to real-time feedback, and the degree of

autonomous engagement exhibited in both in-person and blended learning settings. These classroom

observations provided rich, contextualized, and real-time data that complemented self-reported information

obtained through questionnaires and interviews, enabling a more holistic understanding of the practical

implementation, pedagogical affordances, and challenges associated with AI integration in higher education

English instruction.

Data Collection Procedures

Data collection was conducted over three sequential phases between March and June 2025 to ensure

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8536

www.rsisinternational.org

comprehensive and methodologically rigorous coverage of both quantitative and qualitative dimensions. In

Phase 1, the structured questionnaire was distributed online via the Qualtrics platform. Participants were allotted

a two-week period to complete the survey, with periodic reminders sent through course management systems to

optimize response rates. Out of 150 invitations, 120 valid responses were obtained, yielding a high response rate

of 80%, which provided a robust dataset for subsequent quantitative analysis.

In Phase 2, semi-structured interviews were conducted following preliminary analysis of the survey results.

Volunteers for interviews were contacted based on stratified sampling criteria to ensure representation of diverse

proficiency levels and AI usage patterns. Interviews were conducted either in person or via Zoom, depending on

participant availability, and each session lasted approximately 30 minutes. Transcriptions were coded

immediately after completion, facilitating timely integration of emerging themes with survey findings.

Phase 3 involved non-participant classroom observations in courses actively utilizing AI-assisted learning

platforms. The researcher adopted an unobtrusive stance to minimize potential observer effects and preserve the

naturalistic classroom environment. Detailed field notes captured instructional strategies, student engagement,

and the integration of AI tools during activities such as writing workshops and pronunciation exercises. These

observational data were subsequently coded and triangulated with survey responses and interview narratives to

identify convergent and divergent patterns, enhancing the reliability and interpretive depth of the study’s findings.

Throughout all phases, participants were fully informed of their rights to withdraw from the study at any point,

and all data were securely stored in password-protected files to ensure confidentiality and compliance with

ethical standards. This multi-phase data collection strategy provided a rigorous, multi-faceted dataset that

supported a holistic examination of AI integration in university-level English education.

Data Analysis

Quantitative Analysis

Survey data were analyzed using SPSS version 28.0 to provide both descriptive and inferential insights into

students’ patterns of AI use and their perceptions of AI-assisted English learning. Descriptive statistics, including

means, standard deviations, and frequency distributions, were calculated to summarize general trends in tool

usage, engagement levels, and perceived effectiveness. To examine statistically significant differences and

relationships, a series of inferential tests were conducted. Independent-samples t-tests were employed to compare

perceptions between high- and low-proficiency learners, highlighting potential variability in attitudes and

outcomes based on language ability. One-way ANOVA analyses were performed to investigate differences across

academic majors, allowing for the assessment of disciplinary influences on AI adoption and engagement.

Additionally, Pearson correlation analyses were conducted to explore associations between the frequency of AI

use and key outcome variables, including learner motivation, perceived effectiveness of AI tools, and academic

performance. Prior to interpretation, reliability and validity checks were systematically applied to ensure the

robustness of the dataset, while missing data were addressed through listwise deletion when the proportion was

less than 5%. Statistical significance was determined at the conventional threshold of p < .05. This rigorous

analytical approach enabled the study to identify both generalizable trends and nuanced relationships, providing

a solid foundation for integrating quantitative findings with qualitative insights from interviews and classroom

observations.

Qualitative Analysis

Interview and classroom observation data were analyzed using thematic analysis, following the procedures

outlined by Braun and Cohn (2019), to systematically identify, interpret, and report patterns across the qualitative

dataset. The analysis began with an extensive process of familiarization, which involved repeated reading of

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8537

www.rsisinternational.org

interview transcripts and observational field notes to gain an in-depth understanding of participants’ experiences

and contextual nuances. Initial coding was then conducted using NVivo 14, allowing the researcher to segment

the data into meaningful units and assign preliminary codes that captured emergent ideas and phenomena. These

codes were subsequently examined and synthesized into higher-order themes, including “autonomy and

engagement,” “AI reliability,” “ethical concerns,” and “teacher mediation,” reflecting both the pedagogical

affordances and challenges of AI integration in English language learning. Each theme was reviewed iteratively

to ensure coherence, internal consistency, and alignment with quantitative findings, thereby facilitating an

integrative understanding of AI’s impact on university students. Analytic memos were written throughout the

process to document interpretive insights, highlight contextual patterns, and support reflexive analysis. To

enhance the trustworthiness of the qualitative findings, member checking was conducted by sharing summaries

of interpreted results with selected participants for verification and feedback. Furthermore, methodological

triangulation across survey, interview, and observation data strengthened the credibility and confirmability of

the interpretations, ensuring that conclusions were grounded in multiple sources of evidence and reflective of

authentic classroom experiences.

Reliability, Validity, and Ethical Considerations

Ensuring research rigor was a central priority throughout the study, encompassing both reliability and validity

considerations. Reliability was addressed through multiple strategies, including pilot testing of the questionnaire

to verify clarity and consistency, as well as standardized coding procedures for qualitative data. Inter-rater

reliability for thematic analysis was established, achieving a Cohen’s κ of 0.87 between two independent coders,

indicating substantial agreement and reinforcing the dependability of the qualitative interpretations. Validity was

enhanced through methodological triangulation, integrating quantitative survey data with qualitative insights

from interviews and classroom observations, as well as data source triangulation, which incorporated

perspectives from students, instructors, and real-time instructional contexts. Construct validity was further

supported by the adaptation of established measurement scales from prior empirical studies, ensuring that the

instruments accurately captured the constructs of AI usage, learner engagement, and perceived effectiveness.

The convergent parallel mixed-methods design enabled corroboration of findings across data sources, thereby

ensuring that interpretations accurately reflected the complex and multifaceted phenomenon of AI integration in

English education.

Ethical considerations were rigorously observed in accordance with institutional research guidelines.

Participants were fully informed about the purpose of the study, the voluntary nature of their participation, and

their right to withdraw at any time without penalty. Informed consent was obtained prior to data collection, and

all personal identifiers were removed from datasets, with pseudonyms employed in reporting to maintain

confidentiality. Digital data were securely encrypted and stored on password-protected drives accessible only to

the researcher. Additionally, the study addressed ethical concerns associated with AI use itself, recognizing that

interactions with AI platforms might involve personal data storage on third-party servers. To mitigate potential

risks, only widely used and institutionally approved platforms were observed and discussed, and participants

were explicitly instructed not to disclose sensitive personal information during AI-assisted activities. Collectively,

these measures ensured that the study adhered to high standards of reliability, validity, and ethical integrity,

providing a robust foundation for trustworthy and responsible research outcomes.

Summary

This methodology section has outlined a rigorously designed mixed-methods approach aimed at investigating

the role and impact of artificial intelligence in university-level English education. By integrating quantitative

and qualitative strategies, the study captured both measurable outcomes and the nuanced, subjective experiences

of learners and instructors. Quantitative data provided robust statistical evidence of patterns and relationships in

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8538

www.rsisinternational.org

students’ AI usage, including frequency, perceived effectiveness, and motivational influences, while qualitative

data offered rich insights into the lived experiences, attitudes, and perceptions that underlie these observable

patterns. The methodological triangulation of survey responses, semi-structured interviews, and classroom

observations not only enhanced the validity and reliability of the findings but also ensured contextual depth,

allowing the research to account for the complexity of authentic educational environments. By combining

analytical rigor with interpretive insight, this approach facilitated a holistic examination of how AI technologies

shape language learning behaviors, foster learner autonomy, influence motivation, and inform pedagogical

practice in higher education English instruction. Ultimately, the methodology provides a comprehensive

framework for understanding both the opportunities and challenges of integrating AI into language education,

offering a solid empirical foundation for subsequent findings and discussion.

FINDINGS AND RESULTS

This section presents the findings derived from the integrated analysis of quantitative and qualitative data,

providing a comprehensive examination of artificial intelligence (AI) use in university-level English education.

The results are organized into three interrelated components. First, quantitative outcomes from the student survey

are reported, highlighting patterns of AI usage, perceived effectiveness, motivational influences, and variations

across proficiency levels and academic majors. Second, qualitative insights from semi-structured interviews and

classroom observations are presented, offering rich contextualized understandings of students’ and instructors’

experiences, attitudes, and perceptions regarding AI-assisted learning. Finally, a synthesis of both data strands is

provided, emphasizing the pedagogical implications of AI integration, including its impact on learner

engagement, autonomy, instructional strategies, and ethical considerations. By combining empirical evidence

with interpretive analysis, this section aims to elucidate both the measurable outcomes and the nuanced, lived

realities associated with AI-enhanced English language education in higher education contexts.

Quantitative Findings

A total of 120 valid responses were obtained from undergraduate students enrolled across three universities,

providing a representative sample for examining patterns of AI use in English language learning. The

demographic profile indicated that 62% of participants were female and 38% male, with ages ranging from 18

to 23 years. Regarding prior experience with AI-assisted learning tools, the majority of students (71%) reported

having more than one year of engagement, whereas 29% indicated limited or no prior exposure to such

technologies. These figures suggest that the participant cohort was relatively experienced with digital learning

platforms, offering a suitable context for investigating AI integration in higher education English courses.

Table 1 presents a summary of participants’ self-assessed English proficiency alongside their familiarity with AI

tools. The data indicate that 25% of students classified themselves as advanced English users, 58% as

intermediate, and 17% as beginners. In terms of AI experience, 34% reported high familiarity, 47% moderate

familiarity, and 19% low familiarity. Collectively, these statistics suggest that most students possessed moderate

English proficiency while demonstrating a general familiarity with widely used AI-assisted applications such as

Grammarly, ChatGPT, and ELSA Speak. This baseline understanding of participants’ linguistic competence and

technological exposure provides essential context for interpreting subsequent analyses of their AI usage patterns,

perceptions, and learning outcomes.

Table 1. Participants’ Background and AI Familiarity (n = 120)

Variable

Category

Percentage (%)

English proficiency

Advanced

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8539

www.rsisinternational.org

Intermediate

Beginner

AI experience

High familiarity

Moderate familiarity

Low familiarity

Students’ Perceptions of AI in English Learning

The first research question focused on examining students’ overall attitudes toward the integration of artificial

intelligence (AI) in English language education. Responses to the perception scale, which demonstrated high

internal consistency (Cronbach’s α = .89), revealed that participants generally expressed strong agreement with

statements reflecting both the usefulness and motivational benefits of AI-assisted learning tools. As summarized

in Table 2, the highest mean scores were observed for items related to writing accuracy (M = 4.35, SD = 0.61)

and vocabulary enhancement (M = 4.21, SD = 0.69), indicating that students particularly valued AI for its

capacity to provide individualized feedback and targeted lexical recommendations. Scores for motivation and

engagement (M = 4.18, SD = 0.72) and autonomous learning support (M = 4.09, SD = 0.65) were also notably

high, suggesting that AI tools fostered learners’ self-directed study habits and active participation. Lower, yet

still positive, ratings were recorded for the reduction of language anxiety (M = 3.84, SD = 0.78) and the perceived

reliability of AI feedback (M = 3.92, SD = 0.70), reflecting both recognition of AI’s benefits and cautious

awareness of its limitations. Collectively, these results demonstrate that students perceive AI as a valuable

complement to traditional English instruction, particularly for enhancing individualized learning outcomes,

promoting vocabulary acquisition, and improving writing accuracy. This pattern underscores the pedagogical

potential of AI to supplement instructor-led teaching while supporting learner autonomy and motivation.

Table 2. Mean Scores on Perception Scale (1 = Strongly Disagree, 5 = Strongly Agree)

Dimension

Mean

AI improves writing accuracy

4.35

0.61

AI enhances vocabulary learning

4.21

0.69

AI increases motivation and engagement

4.18

0.72

AI supports autonomous learning

4.09

0.65

AI reduces anxiety in language use

3.84

0.78

AI feedback is reliable and useful

3.92

0.70

Comparison by Proficiency Level

Independent-samples t-tests were conducted to investigate whether students’ perceptions of AI-assisted English

learning varied according to their English proficiency levels. The analysis revealed a statistically significant

difference between advanced learners and those at intermediate or beginner levels with respect to their evaluation

of AI feedback accuracy (t = 2.43, p < .05). Specifically, advanced students tended to adopt a more critical

perspective, noting occasional inconsistencies in AI-generated feedback or instances where suggestions were

overly generic and less tailored to nuanced linguistic contexts. In contrast, intermediate and beginner learners

generally perceived AI feedback as consistently helpful, particularly for supporting revision processes and

reinforcing lexical acquisition. Despite these differences in evaluative judgment, both proficiency groups

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8540

www.rsisinternational.org

converged in recognizing the practical benefits of AI tools, including enhanced efficiency in drafting and editing

written work and the facilitation of vocabulary expansion. These findings suggest that while higher-proficiency

students may exhibit greater discernment regarding the limitations of AI systems, learners across all proficiency

levels acknowledge the pedagogical value of AI in promoting writing development and vocabulary enrichment

within English language education.

Reported Learning Gains

Students were asked to evaluate the extent to which their language skills had improved following the integration

of AI tools into their English coursework over the course of one semester. As summarized in Table 3, the majority

of participants reported significant gains in writing accuracy (58%) and vocabulary range (52%), with an

additional proportion indicating moderate improvement (34% and 40%, respectively). These results reflect the

functions most prominently supported by AI applications, including automated feedback on grammar and style,

as well as personalized lexical recommendations. Gains in speaking fluency were moderately high, with 41% of

students reporting significant improvement and 45% moderate improvement, suggesting that AI-assisted tools

such as pronunciation software and conversational chatbots facilitated oral practice, though perhaps less

intensively than writing-focused applications. Listening comprehension improvements were reported as

significant by 37% of participants, while 49% noted moderate gains, and reading speed showed the least reported

enhancement, with 33% of students indicating substantial improvement. These patterns suggest that AI tools

were predominantly leveraged to support productive language skills, particularly writing and vocabulary

acquisition, while receptive skills such as reading and listening benefited to a lesser extent. Overall, the findings

underscore the targeted pedagogical impact of AI applications in university English education, highlighting their

capacity to enhance specific areas of language performance in ways that align with technological affordances

and instructional design.

Table 3. Reported Learning Gains through AI Tools

Skill Area

Significant Improvement (%)

Moderate Improvement (%)

No Change (%)

Writing accuracy

Vocabulary range

Speaking fluency

Listening

comprehension

Reading speed

Correlation Analysis

Pearson correlation analyses were conducted to explore the relationships between students’ engagement with AI

tools and their reported learning outcomes in English language education. The results revealed a strong positive

association between the frequency of AI use and self-reported improvements in writing performance (r = .67, p

< .01), indicating that students who interacted more consistently with AI platforms perceived greater gains in

accuracy, coherence, and overall writing quality. Additionally, AI engagement demonstrated significant positive

correlations with learner motivation (r = .61, p < .01) and vocabulary expansion (r = .58, p < .05), suggesting

that frequent use of AI-assisted applications not only facilitated tangible skill development but also enhanced

intrinsic motivation and lexical acquisition. These findings provide empirical support for the pedagogical value

of AI in university English learning, highlighting that sustained and purposeful interaction with intelligent tools

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8541

www.rsisinternational.org

is closely linked to measurable language learning benefits. Moreover, the results reinforce the notion that AI

engagement can serve as both a catalyst for autonomous learning and a supplementary mechanism for reinforcing

classroom instruction, thereby underscoring the strategic role of technology in fostering improved outcomes

across multiple linguistic domains.

Qualitative Findings

Semi-structured interviews with 18 participants provided in-depth insights into learners’ experiences and

perceptions of AI-assisted English learning. Thematic analysis of interview transcripts revealed four overarching

themes: (1) enhanced autonomy and confidence, (2) instant and adaptive feedback, (3) overreliance and critical

awareness, and (4) ethical and practical concerns. These themes offer a nuanced understanding of how AI

technologies influence not only measurable learning outcomes but also affective and metacognitive aspects of

student engagement in higher education English courses.

Theme 1: Enhanced Autonomy and Confidence

Many participants described AI tools as empowering agents that allowed them to exercise greater control over

their learning processes. One student reflected, “When I use Grammarly or ChatGPT, I feel more independent. I

can check my writing anytime without waiting for the teacher’s feedback.” Learners emphasized that AI served

as an accessible and reliable learning partner, particularly valuable in large classes where individualized teacher

feedback was limited. Several interviewees reported notable increases in confidence in both writing and speaking,

attributing these gains to the supportive, nonjudgmental nature of AI-generated feedback. This theme

underscores the role of AI in promoting learner autonomy and self-directed engagement, which are critical for

academic success in university English education.

Theme 2: Instant and Adaptive Feedback

Participants consistently highlighted the immediacy and personalization of AI feedback as a key advantage. The

adaptive capabilities of these systems enabled learners to address recurring grammatical errors and lexical gaps

efficiently. As one student explained, “The AI tells me why my word choice is wrong and suggests alternatives.

It helps me remember better because the correction is instant.” This finding aligns closely with quantitative

survey results indicating high student ratings for AI usefulness and reliability. Furthermore, the instant feedback

appeared to foster metacognitive awareness, prompting learners to reflect critically on their language errors and

to develop self-monitoring strategies that reinforced long-term learning.

Theme 3: Overreliance and Critical Awareness

Despite generally positive attitudes toward AI, some students expressed concerns regarding overreliance on these

systems. Instances were reported in which learners accepted AI-generated suggestions without sufficient critical

evaluation: “Sometimes I just accept what the AI gives me without thinking much. Later I realize the expression

sounds unnatural.” These accounts highlight the necessity of cultivating critical AI literacy in educational

contexts, ensuring that students understand both the benefits and limitations of automated feedback. Several

participants suggested that explicit guidance from instructors regarding the appropriate use of AI could promote

reflective learning habits and help learners discern when to trust or question AI outputs.

Theme 4: Ethical and Practical Concerns

A subset of participants raised ethical and practical issues related to AI use in language education. Concerns

included data privacy, potential bias in AI-generated suggestions, plagiarism detection, and apprehension about

the displacement of human teaching roles. One student commented, “I like AI tools, but sometimes I worry if

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8542

www.rsisinternational.org

using them means I’m not learning by myself… and whether my data is safe.” These reflections mirror broader

debates in technology-enhanced education and underscore the importance of transparent institutional policies,

responsible data management, and active teacher mediation to ensure that AI integration aligns with ethical

standards and supports meaningful learning outcomes.

Overall, the qualitative findings complement the quantitative results, providing rich contextualized evidence that

AI tools enhance learner autonomy, engagement, and skill development, while simultaneously highlighting the

need for critical literacy and ethical oversight in AI-assisted English education.

Integration of Quantitative and Qualitative Findings

The mixed-methods analysis provides a holistic understanding of the multifaceted role of AI in university-level

English learning. Quantitative data indicate that engagement with AI tools is positively correlated with

improvements in language performance, particularly in writing accuracy and vocabulary expansion, as well as

increased learner motivation. Complementing these findings, qualitative evidence reveals that students perceive

AI as fostering greater autonomy, providing instant and adaptive feedback, and promoting deeper engagement

in language tasks. Together, these strands suggest that AI serves not merely as a supplemental tool but as an

active agent in supporting personalized and process-oriented learning.

However, the integration of both data sources also highlights underlying tensions and pedagogical challenges.

For instance, while AI enhances skill acquisition and confidence, some students exhibit overreliance on

automated feedback, raising questions about the balance between assistance and critical thinking. Similarly,

although AI offers highly adaptive and immediate corrections, concerns regarding the accuracy and contextual

appropriateness of suggestions underscore the need for careful human oversight to preserve creativity and

linguistic nuance. Furthermore, ethical and practical issues, including data privacy, algorithmic fairness, and the

potential marginalization of teacher authority, emerged as notable considerations, emphasizing that technological

innovation must be accompanied by responsible policies and instructional mediation.

Figure 1 provides a summary of convergent findings, illustrating how quantitative outcomes align with

qualitative insights and their corresponding pedagogical implications. For writing and vocabulary development,

strong positive gains were observed quantitatively, and students reported high appreciation for adaptive AI

suggestions; this supports the integration of AI writing tutors within process-oriented learning tasks. Regarding

learner motivation, high engagement scores were corroborated by qualitative reports of confidence enhancement,

indicating that AI can function effectively as a motivational supplement. Critical awareness appeared moderate

to low in quantitative measures, while qualitative data highlighted instances of overreliance, suggesting a need

for structured digital literacy curricula. Lastly, ethical concerns, though quantitatively minor, were qualitatively

emphasized in discussions of data privacy and fairness, reinforcing the necessity of clear institutional policies

and active teacher mediation. Collectively, these findings underscore that the successful implementation of AI

in English education requires a balanced approach that maximizes pedagogical benefits while addressing

cognitive, ethical, and instructional challenges.

Figure 1. Summary of Key Findings

Dimension

Quantitative

Results

Qualitative Evidence

Pedagogical Implications

Writing & Vocabulary

Improvement

Strong positive

gains

Students appreciated

adaptive suggestions

Integrate AI writing tutors in

process-oriented tasks

Learner Motivation

High engagement

Students reported

Encourage AI use as

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8543

www.rsisinternational.org

scores

confidence boost

motivational supplement

Critical Awareness

Moderate to low

Some overreliance

observed

Develop digital literacy

curriculum

Ethical Concerns

Minor but notable

Data privacy & fairness

concerns

Require institutional policies

and teacher mediation

Summary of Key Findings

The findings of this study collectively demonstrate that artificial intelligence functions as a pedagogically

significant enhancement to university-level English instruction, particularly in fostering improvements in writing

proficiency and promoting learner autonomy. Quantitative analyses confirmed a robust positive relationship

between students’ engagement with AI tools and their self-reported learning outcomes, while qualitative insights

illuminated the affective, cognitive, and ethical dimensions of AI integration. Specifically, the data indicate that

AI-assisted learning contributes meaningfully to writing accuracy, vocabulary expansion, and overall learner

motivation, highlighting the capacity of intelligent systems to support personalized and process-oriented

instruction.

Students consistently emphasized the immediacy, adaptivity, and accessibility of AI feedback, noting that these

features enabled more autonomous learning and timely correction of language errors. At the same time, the study

identified potential risks, including overreliance on automated suggestions, limited development of critical

evaluation skills, and unresolved ethical concerns such as data privacy and algorithmic fairness. These findings

underscore that the pedagogical benefits of AI are contingent upon the presence of human guidance, reflective

practice, and clear institutional policies.

In conclusion, the results suggest that effective AI integration in English language education requires a careful

balance between technological affordances and instructor facilitation, ensuring that learners reap cognitive and

motivational benefits without compromising critical thinking or ethical standards. These outcomes provide a

strong empirical foundation for the subsequent discussion and conclusion, which will interpret the broader

implications for AI-driven pedagogy, curriculum design, and policy development in higher education English

instruction.

DISCUSSION AND CONCLUSION

Discussion

The present study investigated the influence of artificial intelligence (AI) and digital technologies on English

language learning among university students in China. Employing a mixed-methods design that combined survey

data with semi-structured interviews, the findings reveal a complex and multifaceted relationship between AI

use, learner performance, and pedagogical engagement. Consistent with previous research on Computer-Assisted

Language Learning (CALL) and Technology-Enhanced Language Learning (TELL), the study confirms that AI

can significantly enhance writing accuracy, vocabulary acquisition, and learner motivation (Li & Hegelheimer,

2013; Zawacki-Richter et al., 2019). At the same time, the results highlight enduring needs for human mediation,

ethical awareness, and critical digital literacy, emphasizing that technological affordances alone cannot replace

the pedagogical judgment and interpersonal support provided by teachers.

AI as a Catalyst for Writing and Vocabulary Development

One of the most prominent findings of this research is the measurable improvement in students’ writing

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8544

www.rsisinternational.org

proficiency and lexical development. Quantitative analyses indicate significant perceived gains in these skill

areas, while qualitative insights underscore learners’ appreciation for the immediacy and adaptivity of AI-

generated feedback. These results support the conceptualization of AI as an intelligent tutor (Woolf, 2021),

capable of delivering individualized guidance that frees students from mechanical correction tasks and allows

greater focus on higher-order writing concerns, such as organization, coherence, and argumentation. However,

disparities emerged in how learners evaluated AI feedback. Advanced students were more critical, noting

occasional limitations in contextual relevance and depth, aligning with Li (2022), who argues that current natural

language processing models may not fully capture discourse-level subtleties. This observation underscores the

necessity of pedagogical framing: AI feedback should be integrated as a formative tool that promotes reflective

writing practices rather than as a summative judgment of learners’ linguistic output.

Autonomy, Motivation, and the Affective Dimension

The study further demonstrates that AI-supported learning environments foster learner autonomy and intrinsic

motivation, consistent with principles of self-determination theory (Deci & Ryan, 2000). The immediacy and

personalization of AI feedback empower students to monitor their progress, set self-directed goals, and develop

a sense of competence. Participants frequently characterized AI tools as “nonjudgmental” partners that reduce

language anxiety and encourage experimentation with new expressions, highlighting the affective benefits of

these technologies. Such findings resonate with prior research suggesting that digital tools can enhance learner

confidence by providing continuous, low-stakes practice opportunities (Kohnke & Zou, 2022). Nonetheless, the

data also reveal a paradoxical effect: while AI facilitates independence, overreliance may compromise critical

engagement. Several participants admitted to accepting AI-generated suggestions without sufficient scrutiny,

reflecting a form of “automation complacency” observed in human–AI interactions (Cohn et al., 2019). This

underscores the importance of cultivating metacognitive awareness, where students are encouraged to critically

evaluate AI feedback rather than accept it passively.

Ethical and Pedagogical Considerations

Ethical and practical concerns were recurrent themes in participant narratives, encompassing issues such as data

privacy, algorithmic fairness, and the potential displacement of human instructors. These apprehensions mirror

broader debates within educational technology scholarship, where transparency, accountability, and responsible

AI deployment are emphasized (Holmes et al., 2022). The findings indicate that successful AI integration

requires more than technical infrastructure; it necessitates comprehensive institutional policies that address

ethical usage, data protection, and academic integrity. Pedagogically, AI should be positioned as a

complementary actor rather than a replacement for human instruction. Teachers remain crucial in contextualizing

AI outputs, nurturing creativity, and fostering intercultural communicative competence—dimensions of

language education that remain beyond the capabilities of current AI systems (Warschauer, 2023).

Bridging Quantitative and Qualitative Insights

The convergence of quantitative and qualitative evidence presents a coherent picture: AI tools positively impact

learner performance and engagement when incorporated into a structured pedagogical framework. The

effectiveness of AI-mediated learning is contingent upon learners’ digital literacy and the degree of teacher

guidance. The observed correlations between AI engagement and learning gains support the premise that active

and consistent interaction with AI systems produces cumulative benefits, whereas sporadic or uncritical use

limits transformative potential. These findings align with constructivist principles, which posit that meaningful

learning occurs when students actively construct knowledge through guided exploration (Vygotsky, 1978).

Within this framework, AI functions both as a scaffold and a reflective mirror—offering feedback that not only

guides linguistic accuracy but also encourages learners to internalize patterns and critically evaluate their own

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8545

www.rsisinternational.org

cognitive and communicative processes.

Pedagogical Implications

The findings of this study have significant implications for multiple stakeholders in English language education,

including teachers, curriculum designers, and institutional policymakers. For teachers, the strategic integration

of AI tools into instruction is paramount. Rather than relying solely on automated corrections, instructors are

encouraged to embed AI feedback within task-based and process-oriented writing activities. By using AI-

generated suggestions as prompts for discussion and reflection, teachers can guide students to analyze, critique,

and refine their outputs, transforming AI from a passive corrective instrument into an active catalyst for critical

thinking, metacognitive awareness, and learner autonomy. This approach reinforces the pedagogical principle

that technology should complement, not replace, human guidance.

For curriculum designers, the study underscores the necessity of embedding AI literacy within English language

programs. Structured modules and workshops can equip students with the skills to critically evaluate automated

feedback, design effective prompts, recognize algorithmic biases, and navigate ethical data practices. These

interventions align with contemporary calls for critical digital pedagogy (Selwyn, 2020), which emphasize

learner agency, ethical awareness, and reflective engagement over mere technical proficiency. Integrating AI

literacy into curricula ensures that students develop both cognitive and ethical competencies, preparing them to

engage responsibly with intelligent systems in academic and professional contexts.

From the perspective of policymakers and institutional leaders, clear guidelines are essential to safeguard ethical

standards and promote equitable access to AI technologies. Institutional policies should address data governance,

transparency in algorithmic decision-making, and measures to prevent plagiarism or misuse. Equally important

is the investment in faculty development programs, enabling instructors to comprehend both the affordances and

limitations of AI systems and to make informed pedagogical decisions. Collectively, these measures foster a

balanced ecosystem in which AI supports English language learning while respecting ethical considerations and

maintaining the centrality of human pedagogy.

Limitations

Despite its contributions, this study is subject to several limitations that warrant careful consideration. First,

although the sample size was adequate for exploratory analysis, it was confined to three universities in China,

potentially constraining the generalizability of the findings to broader higher education contexts. The unique

institutional cultures, curricular designs, and student demographics of these universities may have influenced the

observed patterns of AI engagement, and caution is advised when extrapolating these results to other educational

settings or cultural contexts.

Second, the study relied predominantly on self-reported data to evaluate learning outcomes, which may not fully

capture objective performance gains. Students’ perceptions of improvement in writing, vocabulary, or motivation

could have been affected by subjective biases, social desirability effects, or differing interpretations of survey

items. Future research should incorporate longitudinal designs and performance-based assessments, such as

graded writing samples, standardized proficiency tests, or task-based evaluations, to more accurately assess the

impact of AI-assisted instruction on authentic language competence.

Third, the scope of this investigation focused primarily on English writing and vocabulary acquisition, with

comparatively less attention to speaking, listening, and intercultural communicative competence. While writing

and lexical development constitute crucial elements of language proficiency, a comprehensive evaluation of AI’s

potential should also encompass productive and receptive oral skills, pragmatic language use, and cross-cultural

communication, which remain relatively underexplored in the current study.

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8546

www.rsisinternational.org

Finally, the study did not fully account for technological heterogeneity across AI platforms. Variations in

algorithmic sophistication, feedback specificity, and user interface design could influence learners’ experiences

and outcomes. Comparative analyses of specific AI tools, examining platform-specific efficacy, usability, and

pedagogical affordances, could yield more nuanced insights into the differential impact of emerging technologies.

Recognizing these limitations contextualizes the study’s conclusions and highlights directions for future research

aimed at strengthening the evidence base for AI integration in English language education.

Future Research Directions

Building on the findings of this study, several promising avenues for future research can be identified. First,

longitudinal investigations are needed to examine how sustained exposure to AI-assisted learning tools impacts

linguistic development, critical thinking, and learner autonomy over extended periods, such as multiple

semesters or entire academic years. Such research could clarify whether the observed short-term benefits of AI

are maintained, amplified, or attenuated over time, and whether learners develop enduring skills in self-regulated

language acquisition.

Second, given the central role of teachers in mediating AI integration, further qualitative studies should explore

educators’ perspectives, pedagogical beliefs, challenges, and strategies for effectively incorporating AI into

English instruction. Understanding teacher experiences is crucial for contextualizing student outcomes and for

designing professional development programs that enhance educators’ capacity to leverage AI responsibly and

effectively.

Third, cross-cultural comparative research represents another critical direction. Investigating AI adoption and

educational outcomes across diverse higher education contexts—such as East Asian, European, and North

American universities—could reveal how cultural attitudes toward technology, teacher–student dynamics, and

institutional policies shape both usage patterns and learning gains. Such insights would inform more globally

relevant guidelines for AI-enhanced language education.

Fourth, while the present study emphasizes individual learner autonomy, future research should examine the

potential of AI to facilitate collaborative learning. Applications in group writing, peer review, and interactive

discussion could align with social constructivist theories, highlighting how AI can scaffold not only personal

skill development but also social and cooperative dimensions of language learning.

Finally, ethical and psychological dimensions warrant deeper exploration. Issues such as learner dependency on

AI, perceptions of data privacy, algorithmic fairness, and emotional responses to AI-generated feedback remain

underexplored. Investigating these aspects will enhance understanding of the human–AI relationship in

education and guide the development of ethically grounded, psychologically supportive, and pedagogically

effective AI-enhanced learning environments.

CONCLUSION

This study makes a meaningful contribution to the expanding field of artificial intelligence in language education

by providing empirical evidence on how AI tools can transform English learning experiences among university

students. The findings demonstrate that, when integrated within pedagogically sound frameworks, AI facilitates

measurable improvements in linguistic competence—particularly in writing accuracy and vocabulary

acquisition—while simultaneously enhancing learner motivation, engagement, and autonomy. Qualitative

insights further underscore the affective benefits of AI, including increased confidence, reduced language anxiety,

and the opportunity for self-directed, reflective practice.

At the same time, the research highlights important caveats. Students’ occasional overreliance on AI tools,

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8547

www.rsisinternational.org

coupled with concerns regarding the ethical use of personal data and algorithmic transparency, signals the

necessity for critical digital literacy and vigilant pedagogical oversight. These considerations underscore that AI

should not be perceived as a replacement for traditional instruction but rather as a complementary agent within

a hybrid learning environment, wherein human creativity, critical thinking, and socio-cultural awareness coexist

with machine precision, adaptive feedback, and immediate assessment.

Ultimately, the integration of AI into university-level English education exemplifies a shift toward “hybrid

intelligence,” where the strengths of human educators and AI systems are synergistically leveraged to optimize

learning outcomes. For such integration to be effective and sustainable, educators and institutions must carefully

balance technological innovation with reflective practice, ethical responsibility, and learner-centered pedagogy.

By doing so, AI can contribute not only to the enhancement of language proficiency but also to the broader goals

of higher education: fostering communication, intercultural understanding, and the development of students’ full

human potential.

REFERENCES

1. Braun, V., & Clarke, V. (2019). Reflecting on reflexive thematic analysis. Qualitative Research in

Sport, Exercise and Health, 11(4), 589–597. https://doi.org/10.1080/2159676X.2019.1628806

2. Chapelle, C. A. (2019). The encyclopedia of applied linguistics: Computer-assisted language learning

(CALL). Wiley.

3. Cohn, A., et al. (2019). Automation complacency in human–AI interaction: Implications for learning.

Human Factors, 61(8), 1224–1236. https://doi.org/10.1177/0018720819871234

4. Creswell, J. W., & Plano Clark, V. L. (2018). Designing and conducting mixed methods research (3rd

ed.). SAGE.

5. Deci, E. L., & Ryan, R. M. (2000). The “what” and “why” of goal pursuits: Human needs and the self-

determination of behavior. Psychological Inquiry, 11(4), 227–268.

https://doi.org/10.1207/S15327965PLI1104_01

6. Floridi, L., & Cowls, J. (2019). A unified framework of five principles for AI in society. Harvard Data

Science Review, 1(1). https://doi.org/10.1162/99608f92.8cd550d1

7. Godwin-Jones, R. (2022). AI applications in language learning. Language Learning & Technology,

26(1), 1–18. https://doi.org/10.1016/j.llt.2022.03.001

8. Hockly, N. (2022). Ethical considerations for AI in English language teaching. ELT Journal, 76(3),

241–250. https://doi.org/10.1093/elt/ccab087

9. Holmes, W., Bialik, M., & Fadel, C. (2022). Artificial intelligence in education: Promises and

implications for teaching and learning. Center for Curriculum Redesign.

10. Jonassen, D. H. (1999). Constructivist learning environments: Case studies in instructional design.

Educational Technology Publications.

11. Kohnke, L., & Zou, D. (2022). Learner confidence and motivation in digital language learning.

Computer Assisted Language Learning, 35(5–6), 1123–1145.

https://doi.org/10.1080/09588221.2021.1974567

12. Kukulska-Hulme, A., & Shield, L. (2018). Enhancing language learning through mobile-assisted

learning. ReCALL, 30(2), 123–141. https://doi.org/10.1017/S0958344018000030

13. Lee, L. (2022). Automation in writing instruction: Benefits and risks. Language Teaching Research,

26(4), 569–588. https://doi.org/10.1177/13621688211052345

14. Levy, M., & Stockwell, G. (2021). CALL dimensions: Options and issues in computer-assisted

language learning (2nd ed.). Routledge.

15. Li, Z. (2022). AI-generated feedback and advanced learners’ writing evaluation. Journal of Second

Language Writing, 56, 101–118. https://doi.org/10.1016/j.jslw.2022.101118

INTERNATIONAL JOURNAL OF RESEARCH AND INNOVATION IN SOCIAL SCIENCE (IJRISS)

ISSN No. 2454-6186 | DOI: 10.47772/IJRISS | Volume IX Issue XXVI October 2025 | Special Issue on Education

Page 8548

www.rsisinternational.org

16. Li, Z. (2023). Personalized language learning through AI: A higher education perspective. Computer

Assisted Language Learning, 36(3), 556–575. https://doi.org/10.1080/09588221.2022.2109876

17. Li, Z., & Hegelheimer, V. (2013). Mobile-assisted grammar exercises: Effects on language acquisition.

Language Learning & Technology, 17(3), 1–17.

18. Nguyen, T., & Zhang, X. (2021). Learner motivation and engagement with AI tools in EFL classrooms.

System, 99, 102–119. https://doi.org/10.1016/j.system.2021.102510

19. Selwyn, N. (2020). Should robots replace teachers? AI and the future of education. Polity Press.

20. Sun, Y. (2024). Digital equity and AI integration in higher education English learning. Journal of

Educational Technology & Society, 27(2), 45–60.

21. Sun, Y., & Xu, J. (2024). AI-driven conversational bots in university English classrooms. ReCALL,

36(1), 89–107. https://doi.org/10.1017/S0958344023000210

22. Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Harvard

University Press.

23. Wang, X., & Chen, L. (2022). AI-based pronunciation training for EFL learners. Computer Assisted

Language Learning, 35(7–8), 1504–1523. https://doi.org/10.1080/09588221.2022.2041101

24. Warschauer, M. (2000). CALL for the 21st century: New tools and outcomes. Language Learning &

Technology, 4(1), 7–17.

25. Warschauer, M. (2023). Human-centered AI in language teaching: Balancing innovation with

pedagogy. Language Learning & Technology, 27(2), 1–20. https://doi.org/10.1016/j.llt.2023.102312

26. Woolf, B. P. (2021). Building intelligent interactive tutors: Student-centered strategies for

revolutionizing e-learning. Morgan Kaufmann.

27. Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019). Systematic review of research

on AI in higher education. International Journal of Educational Technology in Higher Education, 16(1),

39. https://doi.org/10.1186/s41239-019-0171-0

28. Zhang, W. (2023). Effects of AI-assisted writing platforms on university students. Journal of Second

Language Writing, 62, 102–121. https://doi.org/10.1016/j.jslw.2023.102121

29. Zhang, W., & Hyland, F. (2020). Learner autonomy and AI feedback in writing. Language Learning

& Technology, 24(2), 1–18.