This case study investigates the adoption patterns and ethical implications of AI-powered tools among universitystudents for academic writing, aiming to bridge the gap between technological integration and pedagogical responsibility.We conduct a qualitative survey-based analysis involving undergraduate, postgraduate, and research scholars across disciplines at NGM College, focusing on their familiarity, usage frequency, and attitudes toward AI tools such as ChatGPT, Gemini, and Perplexity.The findings reveal a high level of familiarity with AI tools, with 48.6% of participants being "very familiar" and 47.2% "somewhat familiar," while ChatGPT emerges as the dominant tool (83.3%).Students primarily employ AI for writing assistance (81.9%), idea generation (75.0%), and research (72.2%), yet ethical concerns persist, as only 2.7% directly accept AI-generated content without modification.The study identifies a tension between efficiency gains and risks to academic integrity, with 51.4% of respondents using AI suggestions as inspiration but rewriting content independently.Moreover, the data highlights disciplinary variations in tool preferences and task-specific applications, underscoring the need for tailored pedagogical strategies.The research contributes to the growing discourse on AI in education by providing empirical evidence of student behaviors and proposing actionable recommendations for educators, such as redesigning assessments to emphasize critical thinking and integrating transparency mechanisms.These insights are particularly significant given the rapid proliferation of AI tools in academia, where balancing technological assistance with the preservation of original thought remains a pressing challenge.The study ultimately calls for a nuanced approach to AI integration, one that fosters responsible use while maintaining academic rigor and ethical standards.

The present study examines the use of nominal groups, focusing on their types and functions in academic writing, by analyzing a corpus of Pakistani English Research Articles (PERAC). The corpus has been compiled by taking 52 English research articles written by Pakistani scholars in the field of linguistics. After categorizing the nominal groups into simple, pre-modified, and post-modified, the corpus was annotated with the help of TagAnt (2.1.1), and further analysis was carried out through AntConc (3.5.9). The findings of the study show that pre-modified nominal groups (52.79 %) are more frequent than simple nominal groups (39.38 %) and post-modified nominal groups (7.83 %). The analysis of the examples has given the following types of premodification: adjectival (44.5 %), noun (25 %), adverbial (17 %), and participial (13.5 %). While prepositional phrases (51 %), clausal (20 %), infinitival (17.5 %), and adjectival (11.5 %) post-modification were also used in PERAC to provide additional richness and accuracy to the texts. This distribution follows the developmental pattern of structures advanced in nominal group complexity, as postulated by Biber et al. (2011). The findings of the study offer future researchers and academic writers the opportunities to work with implications and suggestions to help students use premodifiers and postmodifiers in nominal groups to enhance their academic writing skill. The present investigation underscores the imperative for additional research comparing the use and role of nominal groups across different disciplines and genres.

Large philosophy classes make it difficult to realize Confucius’ ideal of teaching in accordance with students’ aptitude. Conventional assessments poorly capture growth in higher-order writing skills. This article reports a 16-week reform of an undergraduate philosophical academic writing course that integrated an AI “learning companion,” structured “arguing-with-AI” activities, and peer-supported clinical tutoring. Drawing on work in intelligent tutoring, stealth assessment, and emerging research on generative AI in philosophy teaching, we designed an 8-week conceptual module plus an 8-week clinic-style supervision module. Using a quasi-experimental historical cohort design (N = 79), we compared rubric-based writing scores, process data from AI interaction logs, and student self-reports across cohorts. The AI-supported design was associated with greater increases in argumentative coherence, evidence use, and originality in this sample than a traditional design, and it expanded individualized feedback coverage. We discuss how generative AI can extend—but not replace—instructors’ capacity to enact individualized instruction at scale in a philosophy context, and outline implications for AI governance and academic integrity.

What happens when students write more fluently but engage less deeply in revision? This mixed-methods study examined how structured ChatGPT integration was associated with academic writing development among 320 Bangladeshi undergraduates in English for Academic Purposes courses. AI-assisted writing was associated with improved surface-level fluency and reduced writing anxiety, but also with lower authorial agency, weaker genre performance, and less recursive revision. To assess these shifts, the study introduces two analytic measures: revision entropy, capturing the variety and depth of revision moves, and the epistemic ownership index, estimating observable author control across writing decisions. Quantitative findings showed that AI-assisted drafts demonstrated polished language yet scored lower on coherence, thesis development, and source integration than drafts produced within teacher-scaffolded revision cycles. Revision entropy and epistemic ownership were significantly lower in the AI-mediated group. Qualitative data revealed four recurring tensions: compositional passivity, authorship uncertainty, genre detachment, and feedback fatigue. Conducted in intact (pre-existing, non-randomly assigned) classroom groups, the study reflects associations rather than causal effects. The findings underscore the need to structure AI use so it supports revision and rhetorical decision-making rather than replacing them.

Purpose This study examines how three instructional conditions – Conventional Tutorial, Flipped Classroom Design Thinking (FCDT), and an AI-supported FCDT-AI model using ChatGPT – shape undergraduate students' Digital Literacy within an open and distance learning (ODL) environment at Universitas Terbuka, Indonesia. It responds to the growing need for scalable pedagogical models that integrate flipped learning, design thinking, and generative AI across Asian open universities. Design/methodology/approach A within-subjects repeated-measures design was employed with 26 undergraduate students enrolled in an Academic Writing Techniques course. All participants experienced the three conditions in counterbalanced order via TUWEB, the institutional learning management system. Digital Literacy was measured after each condition using a multidimensional performance-based questionnaire. Quantitative analysis used Huynh–Feldt-adjusted repeated-measures ANOVA with Holm-adjusted post-hoc tests, while qualitative reflection logs were examined using reflexive thematic analysis to elucidate mechanisms underlying observed differences. Findings A significant and substantial main effect of instructional condition was identified, demonstrating a clear performance gradient: Conventional < FCDT < FCDT-AI. The AI-supported condition yielded the highest Digital Literacy scores and the broadest distribution of advanced practices. Qualitative themes further revealed progressive development from basic access and retrieval (Conventional), to structured evaluation and emerging digital production (FCDT), and to multimodal, reflective, and AI-mediated digital engagement (FCDT-AI). Research limitations/implications This study has several limitations. The small sample from a single programme at one ODL institution restricts generalisability, suggesting the need for replication across disciplines, universities, and learner profiles. The reliance on self-reported reflections may introduce subjectivity; integrating learning analytics or artefact analysis would strengthen triangulation. The AI scaffolding was intentionally limited for ethical reasons, meaning future studies could examine varying intensities or types of AI support. Despite these constraints, the findings offer empirically grounded implications for designing scalable, AI-supported flipped learning models in ODL environments. Practical implications The findings provide actionable guidance for ODL institutions seeking to strengthen Digital Literacy at scale. Tutors should integrate structured flipped-learning cycles supported by design thinking to guide learners from basic access toward evaluative and creative digital practices. Incorporating generative AI as guided scaffolding – rather than as an autonomous problem-solver – can expand students' idea generation, support multimodal production, and reduce cognitive load. Curriculum designers can embed FCDT-AI workflows into tutorial manuals, LKMs, and online learning activities to promote consistent digital engagement. Institutions may also develop training programmes to enhance tutors' digital pedagogy and ethical AI facilitation. Social implications Enhancing Digital Literacy through structured flipped and AI-supported models can help narrow digital inequities among geographically dispersed ODL learners. The FCDT-AI framework supports more inclusive participation by providing scaffolding that benefits students with lower digital readiness, thereby promoting equitable access to 21st-century competencies. As generative AI becomes more widespread in education and work, developing students' evaluative, ethical, and creative digital practices contributes to a more informed and responsible digital citizenry. The model also supports lifelong learning, empowering working adults to engage confidently in digitally mediated environments and strengthening broader community digital resilience. Originality/value The study offers one of the first empirically tested pedagogical models that systematically integrates flipped learning, design thinking, and generative AI to strengthen Digital Literacy in ODL environments. It provides a theoretically grounded and scalable framework (FCDT-AI) that can support Asian open universities in implementing ethical and effective AI-enhanced digital learning.

This article explores the integration of artificial intelligence (AI) in an English course for Academic and Professional Communication, situated within the broader context of digital transformation in higher education and the evolving demands of academic integrity. Rather than restricting AI use, the study adopts a critical digital pedagogy approach to foster students' awareness of ethical and effective engagement with AI tools. The activities were designed to encourage reflective practices, promote critical analysis of AI-generated content, and establish clear guidelines for responsible use. These activities include setting explicit expectations for AI-assisted work, integrating reflective assessments to evaluate AI outputs, and designing tasks that strengthen students’ critical thinking and academic integrity. Additionally, the approach aims to improve students’ retention of the course content by encouraging them to use AI not as a shortcut, but as a means to actively reflect, process, and engage with the material explained in class. By interacting with AI to rephrase, question, edit, or elaborate on key concepts, students become more conscious of the content, reinforcing understanding and promoting deeper learning. The results suggest that the students developed a more nuanced understanding of the capabilities and limitations, enabling them to engage with it as a support tool rather than a substitute for the original work. The study concludes that structured integration of AI can enhance students' digital literacy, content retention, and critical thinking skills in the academic context.

This research presents an analysis of challenges and strategies in translating journalistic content, which demands accuracy, speed, and sensitivity to cultural context. Because of these characteristics, translating journalistic texts require different skills and approaches compared to other types of texts. Unlike translator-journalists who find journalistic texts as part of their daily work, freelance translators, who are generally more familiar with literary, legal, advertising and marketing, medical, and business texts often faces unique difficulties due to the concise and clear nature of journalistic writing. These challenges require careful attention to semantic, syntactic, and pragmatic elements to ensure that meaning is effectively conveyed to the target audience. The main objective of this study is to identify and analyze the specific challenges faced by freelance translator when translating journalistic content for the first time and to explore the strategies used to address them. The study also aims to provide insights that can contribute to a broader understanding of translation practices in a journalistic context. A single case study design is used by involving a single freelance translator with two years of experience in translating academic writings, literary works, and promotional texts. Data were collected through in-depth interviews and analyzed using thematic analysis by Braun and Clarke (2006). The analysis revealed four key challenges: accuracy and readability, cultural nuance, bias and framing, and deadline pressures. To overcome these challenges, four main strategies are used: adaptation, localization, transcreation, and selective omission and addition. The findings reveal that freelance translators need a combination of linguistic skills, cultural sensitivity, and effective time management to produce credible, reliable, and accurate translations.

ABSTRACT This study examined the level of awareness, frequency of use, perceived benefits, and challenges associated with artificial intelligence tools for academic writing among academic librarians in universities in Nigeria, using the Technology Acceptance Model as its theoretical anchor. A cross-sectional survey design was adopted. Data were collected from 93 academic librarians through a structured questionnaire administered online. The study was guided by the Technology Acceptance Model. Descriptive statistics were used to analyse the research questions, while chi-square analysis tested the relationship between selected demographic variables and awareness of AI tools. The findings show a moderate level of awareness of widely known AI tools such as ChatGPT and Grammarly, while awareness of other platforms remained comparatively low. Actual use of AI tools for academic writing was generally low and mainly limited to basic tasks such as proofreading and idea generation. Perceived benefits were rated low, indicating weak perceived usefulness of AI tools for academic writing. In contrast, perceived challenges were rated high, reflecting concerns about output inaccuracy, data privacy risks, high cost of access, limited technical skills, inadequate institutional support, and poor internet connectivity. No significant relationship was found between awareness of AI tools and demographic variables such as gender, academic qualification, and work experience. The study concludes that although academic librarians in Nigerian universities are moderately aware of artificial intelligence tools for academic writing, negative perceptions regarding usefulness, ease of use, and trust significantly limit AI tool adoption. Addressing these concerns requires clear institutional policies, targeted capacity-building programmes, and improved infrastructural support to enable ethical and effective integration of AI tools into academic writing practices.

Although the current generative AI academic writing assistant has deeply penetrated the literature retrieval and data analysis of business academic research, whether it can effectively improve the efficiency of business academic research remains to be verified. In this paper, the generative AI academic writing assistant based on business academic research is divided into two modules, information extraction and writing output, to form a business academic writing assistant model. The model proposes a Bert-based extractive summarization method in the extraction of key information of academic text, adopts BertSum to extract the feature vectors of academic text, uses BiGRU to capture the contextual relationship between sentences, integrates GRTU encoder and attention mechanism to accurately extract the relevant information, and utilizes the classification layer to judge whether the sentence stays or goes. In terms of academic text writing output, a selector is utilized to filter out important academic text arguments, and a rewriter is used to generate the corresponding complete academic research content. Logistic regression model was chosen as the research analysis tool, research samples were selected, business academic research efficiency was set as the dependent variable, and the prediction model was constructed based on the results of regression analysis parameter estimation of the seven independent variables. The prediction model of business academic research efficiency predicted 236 students with more than 80.00% accuracy for all three academic research efficiencies.

This study investigates student perceptions regarding the implementation of electronic portfolios (e-portfolios) in tertiary-level academic writing courses. Utilizing an experimental research design, the study compares the effectiveness and student experiences of e-portfolios against traditional paper-based portfolios within the L2 writing classroom. The research was conducted on two homogeneous groups enrolled in the "English Composition and Communication Skills" course, where variables were kept constant to ensure comparative accuracy. While one group utilized paper-based portfolios, the other group integrated e-portfolios as a core component of their formative assessment. A mixed-methods approach was employed for data collection and analysis. Quantitative data were gathered through a five-point Likert scale questionnaire administered at the end of the semester, while qualitative insights were obtained through focus group discussions (FGDs) using semi-structured frameworks. Quantitative results were processed using Stata statistical software, whereas FGD data underwent a rigorous process of transcription, translation, thematic coding, and categorization. The findings indicate that students using e-portfolios demonstrated superior organizational skills and were better prepared for examinations, despite encountering initial technological hurdles. Ultimately, the research highlights the positive impact of digital tools on student preparation and learning trajectories, offering significant implications for the future of formative assessment in higher education.

The Role of a Hybrid Model Combining Artificial Intelligence Feedback and Teacher Feedback in German Academic Writing on the Improvement of Learners' Writing Skills

This study examines the challenges and ethical dilemmas faced by English Language Education students at Muhammadiyah University of Luwuk in using Artificial Intelligence (AI) for academic writing. It identifies patterns of AI usage in the writing process. Data were collected through questionnaires and interviews. The findings show that although AI improves writing quality, increases time efficiency, and helps generate ideas, its use raises concerns. The main challenges include overreliance on AI, reduced critical thinking, potential plagiarism, difficulty creating effective prompts, overly general or irrelevant outputs, and unverifiable references. Interview results reveal linguistic and stylistic features of AI-generated texts. Students acknowledge the risks of misuse and propose solutions, including ethics training, university guidelines, mentoring, and access to reliable tools. This study recommends developing clear institutional policies, improving AI literacy, and implementing strict oversight to ensure AI is used ethically and effectively as a support tool rather than a substitute for students’ abilities.

The rapid development of generative Artificial Intelligence (AI) has significantly influenced academic practices, particularly in students’ academic writing. This study aims to examine the ethical use of generative AI in students’ academic writing practices in public communication contexts. The research employs a qualitative approach to explore how students utilize AI technologies and how ethical considerations shape their writing behavior. Data were collected through in-depth interviews with 10 key informants and questionnaires distributed to 30 respondents, complemented by direct observations of students who actively use AI tools in their academic writing. The findings reveal that students widely use generative AI to support various stages of the writing process, including idea generation, outlining, language improvement, and content organization. Students perceive AI as a helpful tool that enhances efficiency and productivity in completing academic assignments. However, the study also identifies several ethical concerns, such as the potential overreliance on AI, risks to academic integrity, and the possibility of reduced critical thinking if AI is used without responsible awareness. Furthermore, the results highlight that students with stronger digital literacy and understanding of academic ethics tend to use AI more responsibly, treating it as a supporting tool rather than a replacement for intellectual work. The study emphasizes the importance of establishing ethical guidelines and improving digital literacy in higher education institutions. These efforts are essential to ensure that AI technologies support academic communication while maintaining transparency, originality, and academic integrity in scholarly writing.

Academic writing in science and technology demands structured expression, evidence-based reasoning, and the ability to manage complex ideas and large volumes of information. However, these processes are often time-consuming and challenging. This study examined the integration of Artificial Intelligence (AI) technologies into academic writing and research workflows in science and technology as a productivity-enhancing approach. Specifically, it explored the awareness of AI tools among scholars and their perceived usefulness in improving research output. Guided by the Technology Acceptance Model (TAM), a sample of 364 respondents was selected using the Taro Yamane formula. Data were collected through structured questionnaires and analyzed using descriptive statistics, mean scores, and standard deviations. Findings revealed high awareness and frequent use of AI tools such as ChatGPT, Grammarly, Grammarly Plagiarism Checker, EndNote, Turnitin, Excel Plugins, Python Libraries, and SciBERT/Semantic Scholar. Respondents reported that AI tools significantly improve the quality of academic writing, reduce grammar and referencing errors, accelerate task completion, and allow greater focus on creative and experimental work. The results also indicated that AI tools are generally easy to learn and integrate into existing research workflows. The study concludes that AI has become a valuable ally in enhancing the efficiency, accuracy, and creativity of academic writing and research in science and technology. It recommends targeted training, institutional investment in AI resources, and the establishment of clear ethical guidelines to ensure responsible adoption.

Abstract Generative artificial intelligence is increasingly being used for academic writing. However, this is potentially a cause of concern if this is having an impact on how authors express themselves. In this study, I generate two corpora from the abstracts of academic articles in the ‘languages and linguistics’ field, one spanning 2018–2019 and another 2023–2024. To analyse these, I take a descriptive non-inferential corpus design to compare ngram frequency in the pre- and post-AI corpora to identify shifts in words and phrases related to authorial stance. I conclude by arguing that there is a possibility that generative AI is already influencing how abstracts are written, with a small shift away from authorial stance.

Topic closers refer to strategies for constructing paragraph-final sentences in academic English writing. Their functional relationship to topic sentences is comparable to that between discussion and introduction sections in research articles. This study evaluates a model of ten topic-closer types (Labels) by examining whether English L2 student writers can learn to judge their functional appropriateness more consistently. Using a pretest–intervention–posttest design, participants rated 100 topic-closer sentences for suitability as paragraph endings. Learning was assessed using two complementary measures: appropriateness ratings and rating inconsistency. Mixed-effects analyses revealed a reliable overall increase in appropriateness ratings following instruction, with no corresponding time × label interaction, indicating that improvement was general rather than category-specific. In contrast, rating inconsistency showed a significant and uniform decrease across labels, suggesting that participants converged on more stable evaluative criteria for paragraph-final function. These findings indicate that brief instruction can strengthen discourse-level judgment even before fine-grained categorical distinctions emerge. Therefore, the model shows promise as both a pedagogical framework for teaching paragraph endings and a foundation for computational applications, including automated feedback systems for L2 academic writing.

This article examines the theoretical and practical mechanisms for developing argumentation and critical thinking skills in the process of academic writing among higher education students. The study employs a mixed-methods approach to analyze students’ logical reasoning, quality of argumentation, and analytical thinking in written discourse. The findings indicate that interactive, argumentation-based instructional methods, such as the Claim–Evidence–Reasoning (CER) model, peer review, and problem-based writing tasks – significantly improve both the quality of academic writing and the level of critical thinking. The results highlight the importance of integrating modern pedagogical strategies into academic writing instruction.

International graduate students often use university writing centers for support with advanced academic writing. Yet writing center philosophies do not always align with multilingual writers’ expectations, especially when students seek language-focused feedback alongside rhetorical and process-oriented support. This qualitative study examines how international graduate students perceive writing center visits as contributing to their academic writing development and what features of those visits shape these perceptions. Guided by a social constructionist perspective, the study draws on semi-structured interviews with seven international graduate students at a U.S. university who reported regular writing center use. Data were analyzed thematically through an iterative codebook approach. Participants viewed writing center visits as contributing to rhetorical awareness, process-oriented development, greater strategic awareness of writing, increased confidence, and reduced anxiety. At the same time, they described recurring tension between the writing center’s no-proofreading philosophy and their need for support with grammar, vocabulary, and sentence-level clarity. They also emphasized that the perceived value of a session depended not only on tutor feedback but also on continuity, appointment access, session structure, and the center’s non-evaluative environment. The study argues that writing center effectiveness for international graduate students is best understood holistically, as shaped by pedagogical, affective, and institutional factors rather than by tutor performance alone. Implications are offered for writing center policy communication, tutor education, and multilingual graduate writing support.

Academic writing is a crucial skill for Indonesian university students, significantly influencing their EFL learning performance. This systematic literature review aims to examine the diverse learning strategies employed by these students to enhance their academic writing skills and reveal the predominant strategy employed by the students. Combining systematic literature review and thematic analysis, this study reveals five categories for academic writing strategies, those are metacognitive, cognitive, social, affective, and compensation strategies. This study highlights a predominant use of metacognitive strategies, comprising 36% of the strategies identified. This emphasis underscores students' commitment to higher-order thinking during the writing process and highlights the importance of metacognition in effective language learning. The distribution of cognitive, social, affective, and compensation strategies illustrates a multifaceted approach to academic writing. The findings emphasize the adaptability of EFL university students in Indonesia in addressing academic writing challenges through various strategies, thereby promoting a comprehensive learning environment. This study recommends further investigation into cultural and contextual factors influencing strategy selection, providing educators with insights for tailored instructional practices.

Students' Perceptions of the Flipped Academic Writing Classroom Learning Activities