Garuda - Garba Rujukan Digital

PERFORMANCE AND ACCURACY OF CHATGPT IN GENERATING MALAY ACADEMIC TEXTS: A COMPARATIVE STUDY WITH EXPERT CORRECTIONS Yaqin, Lalu Nurul; Hassan, Hasmidar; Yusof, Badriyah
LLT Journal: A Journal on Language and Language Teaching Vol 28, No 1 (2025): April 2025
Publisher : English Education Study Programme of Sanata Dharma University, Yogyakarta, Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24071/llt.v28i1.11698

The increasing use of artificial intelligence in academic writing has raised concerns about the accuracy and coherence of AI-generated texts, particularly in underrepresented languages like Malay. This study evaluates the performance of ChatGPT in generating Malay academic texts by comparing AI-generated outputs with expert corrected versions, focusing on grammatical errors, structural inconsistencies, and lexical inaccuracies. A comparative analysis was conducted on two datasets: Trained Dataset (TD), where prompts included detailed context, and Untrained Dataset (UTD). ChatGPT-generated texts were reviewed by Malay linguistics and translation experts, who identified and corrected grammatical errors. A quantitative and qualitative analysis assessed error frequency and categorized linguistic challenges. Findings reveal that UTD contained significantly more grammatical errors (87 errors) than TD (18 errors), demonstrating the role of structured prompts in enhancing text quality. Common errors in UTD included incorrect sentence structure (27.59%), omission of names (14.94%), and inappropriate word choices (11.49%). While TD showed improved grammatical accuracy, errors in phrase structure, conjunction usage, and affixation persisted. The study concludes that AI-generated Malay texts lack syntactic stability, requiring expert intervention and model refinement. These findings highlight the need for linguistic adaptation, expanded training datasets and the integration of expert to enhance AI-generated Malay academic writing. Ultimately, this study presents a case study that provides empirical evidence that context-aware prompt engineering and expert-in-the-loop approaches are essential for enhancing the quality of AI outputs, especially in non-English settings. It also advocates for the development of AI models that can capture nuances and linguistic diversity, vital for inclusive education for all.

Metaphorical Symbols in Qur'anic Discourse: A Cognitive-Linguistic Analysis Chaer, Hasanuddin; Rasyad, Abdul; Sukri, Sukri; Efendi, Mahmudi; Yaqin, Lalu Nurul
Suhuf: International Journal of Islamic Studies Vol. 37 No. 2 (2025): November
Publisher : Universitas Muhammadiyah Surakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.23917/suhuf.v37i2.12081

This article examines the relationship between language, thought, and religious experience in Islamic theology using George Lakoff’s cognitive linguistic framework. Focusing on theological metaphors in selected Qur’anic verses, Al-Baqarah 255 (Ayat al-Kursi), An-Nur 35, Ibrahim 24–25, Al-Hadid 13, and Ar-Ra’d 28, it applies conceptual metaphor theory to reveal how the Qur’an conveys theological and spiritual meanings through cognitive structures. The study proceeds in four stages: identifying and categorizing Qur’anic theological metaphors; analyzing their structure and function through cognitive linguistics; exploring how metaphor links language with religious thought; and interpreting the results to assess their theological significance. The findings show that Qur’anic metaphors transform abstract theological concepts into concrete, accessible ideas, deepening Muslims’ spiritual understanding. This research contributes to cognitive linguistics and religious studies by demonstrating how metaphor serves as a bridge between divine revelation and human cognition. Future studies could develop cognitive linguistic approaches for teaching theology, emphasizing metaphor as a key tool for conceptual and spiritual comprehension.

Charting the Future of Prompt Engineering: Critical Reflections on Methodology, Ethics, and Research Directions Verawati, Ni Nyoman Sri Putu; Yaqin, Lalu Nurul
International Journal of Essential Competencies in Education Vol. 4 No. 1 (2025): June
Publisher : Lembaga Penelitian dan Pemberdayaan Masyarakat (LITPAM)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36312/ijece.v4i1.2908

Prompt engineering has emerged as a transformative strategy for optimizing Large Language Models (LLMs), offering a cost-effective alternative to full model fine-tuning. In a recent bibliometric review, Fatawi et al. (2024) analyzed 437 Scopus-indexed publications from January 2022 to February 2024, using VOSviewer to identify key thematic clusters—including transformer architectures, deep learning innovations, and few-shot learning—and documenting a fivefold increase in related publications over the review period. Building on their macro-level mapping, this commentary extends the discussion by articulating the strategic and democratizing potential of prompt engineering while addressing critical gaps in methodology and ethical oversight. We critique the review’s reliance on a single English-language database, its exclusion of preprints and non-English sources, and its omission of qualitative insights into user practices and system impacts. In response, we offer concrete recommendations to guide future research: diversify data sources for bibliometric analysis, implement rigorous prompt audit frameworks, conduct longitudinal A/B testing in real-world environments, and adopt mixed-methods approaches to capture human-centered dynamics. We also explore emerging synergies—such as quantum-enhanced NLP and neuro-linguistic prompt design—as promising frontiers for advancing prompt optimization. By addressing these gaps, this commentary aims to ensure that prompt engineering evolves not only as a technical solution but as a responsible and inclusive foundation for next-generation AI development.

Transformasi Pendidikan Digital dan Reproduksi Ketimpangan Sosial: Studi Empiris di Indonesia Yaqin, Lalu Nurul; Nugraha, Rizky Aditya; Fadhilah, Nur Aini
PERSEPTIF: Jurnal Ilmu Sosial dan Humaniora Vol. 4 No. 1 (2026): PERSEPTIF: Jurnal Ilmu Sosial dan Humaniora
Publisher : Lembaga Penelitian dan Pendidikan (LPP) Kalibra

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70716/perseptif.v4i1.458

The digital transformation of education in Indonesia has accelerated alongside national policies promoting educational digitalization and technology integration. However, this transformation has not consistently led to equitable educational outcomes and, in many cases, has reproduced existing social inequalities. This study aims to analyze how digital education transformation contributes to the reproduction of social inequality in Indonesia, particularly across regional, socioeconomic, and institutional dimensions. Employing a qualitative descriptive approach, this research utilizes a systematic literature review and thematic analysis of 27 national and international scholarly articles. The findings reveal that disparities in technological access, digital literacy, infrastructure, and institutional capacity are key factors exacerbating educational inequality. Furthermore, uniform and context-insensitive digital education policies tend to marginalize vulnerable groups, especially those in rural and underdeveloped regions. This study concludes that Indonesia’s digital education transformation requires inclusive, context-sensitive, and adaptive policy frameworks to ensure that digitalization functions as a tool for educational equity rather than a mechanism for social reproduction of inequality.

Developing a Local Wisdom-Based “Patju” Religious Culture Model to Improve Students’ Worship Practices in Islamic Senior High Schools Sayuti, Lalu; Burhanudin, Nunu; Salmiwati; Yaqin, Lalu Nurul; Famuji, Untung
Al-Tadzkiyyah: Jurnal Pendidikan Islam Vol 17 No 2 (2026): Al-Tadzkiyyah: Jurnal Pendidikan Islam
Publisher : Universitas Islam Negeri Raden Intan Lampung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24042/atjpi.v17i2.30465

This study developed and evaluated a “Patju”-based religious culture model to improve students’ worship practices in Nahdlatul Wathan Islamic senior high schools in East Lombok, West Nusa Tenggara, Indonesia. The study was motivated by the need to transform school religious routines from formal compliance into internalised worship discipline grounded in local wisdom and Islamic character values. A research and development design using the ADDIE framework was employed, consisting of analysis, design, development, implementation, and evaluation stages. The study involved three Islamic senior high schools, with data collected through observation, interviews, documentation, questionnaires, expert validation, individual trials, and group trials. Qualitative data were analysed through data reduction, data display, and conclusion drawing, while quantitative data were analysed using descriptive statistics to determine the model’s validity, practicality, and effectiveness. The findings showed that the developed model and its supporting prayer module were valid, with an overall validation score of 4.46. The effectiveness test indicated an increase in students’ worship practice scores from approximately 2.41 before implementation to 4.21–4.22 after implementation. The practicality test also showed positive results, with post-implementation practicality scores ranging from 4.13 to 4.22. These findings indicate that the model improved students’ prayer discipline, religious responsibility, honesty in worship reporting, peer support, and motivation to enhance the quality of worship. This study contributes to Islamic education scholarship by offering a culturally responsive religious culture model that integrates Lombok’s local value of “Patju” with structured worship habituation and character formation in Islamic senior high schools.

Generative AI in Translation: Trust, Transformation, and Pedagogical Integration in the Post-LLM Era Yaqin, Lalu Nurul; Ismail, Isma Noornisa
International Journal of Linguistics and Indigenous Culture Vol. 4 No. 1 (2026): March
Publisher : LITPAM

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36312/ijlic.v4i1.4676

The emergence of large language models (LLMs) like Catgut has sparked both excitement and anxiety in the translation field. This literature review critically examines how generative AI is reshaping professional translation and translator education in the post-LLM era. Early studies showed that ChatGPT could produce translations between major languages with quality on par with neural machine translation, fueling hopes for productivity gains. However, trust in generative AI among translators remains limited. Surveys indicate that practitioners adopt tools like ChatGPT cautiously, primarily for ancillary tasks such as summarisation or drafting ideas. Key concerns, including data privacy, confidentiality, output accuracy, and the threat of job displacement, undermine translators’ trust in AI systems. We situate these concerns within technology acceptance models (TAM2) and AI trust frameworks, highlighting how perceived usefulness, reliability, and transparency shape adoption. At the same time, generative AI is transforming translation workflows: translators’ roles are evolving toward post-editing and quality control, raising questions of professional identity and autonomy. In translator education, recent experiments integrating ChatGPT show improved student performance and confidence, provided that pedagogical approaches emphasise critical AI literacy and post-editing skills. This review synthesises theoretical and empirical insights from 2023 to 2025 to present a balanced perspective on generative AI in translation. We argue that a human-centric integration of GenAI, one that builds translator trust, addresses ethical/legal risks, and nurtures new competencies, will be essential for harnessing these technologies’ benefits without undermining professional standards.

Trend Kajian Bahasa Reseptif dan Ekspresif Autism Spectrum Disorder: Analisis Bibliometrik Natasya, Nur Fazeerah; Yaqin, Lalu Nurul; Damit, Ashrol Rahimy
International Journal of Linguistics and Indigenous Culture Vol. 4 No. 1 (2026): March
Publisher : LITPAM

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36312/ijlic.v4i1.4796

Kajian ini bertujuan untuk menganalisis arah aliran ataupun trend berkenaan kajian terhadap bahasa reseptif dan ekspresif Autism Spectrum Disorder (seterusnya ASD) dalam tempoh satu dekad yang lalu (2014–2024). Bahasa reseptif merujuk kepada kemampuan individu untuk memahami komunikasi verbal, manakala bahasa ekspresif melibatkan keupayaan menyampaikan idea dan maklumat secara verbal. Individu dengan ASD sering menghadapi cabaran dalam kedua-dua aspek bahasa ini, yang memberi kesan kepada perkembangan komunikasi dan interaksi sosial mereka. Dengan memanfaatkan kaedah analisis bibliometrik khususnya aplikasi VOSviewer, kajian ini menganalisis data daripada pangkalan data Scopus untuk memahami tema yang dikaji, jenis dokumen, tahun penerbitan, jumlah penerbitan, senarai nama penulis dan jumlah penerbitan, tempat penerbitan serta rangkaian kolaborasi penyelidik dan institusi. Hasil kajian mendapati sebanyak 726 artikel telah ditemui dalam pangkalan data Scopus. Jumlah tersebut menunjukkan peningkatan penerbitan ‘Bahasa Reseptif dan Ekspresif ASD’ yang signifikan bermula dari tahun 2019 sehingga 2024, dan Boston University selaku penyumbang penerbitan terbanyak berbanding institusi tinggi yang lain bersama penulis-penulis utama, seperti Tager-Flusberg dan Kasari, memainkan peranan penting dalam memajukan wacana ilmiah dalam bidang ini. Kata kunci ‘Autism Spectrum Disorder’, ‘Receptive Language’ dan ‘Expressive Language’ pula tersenarai sebagai kata kunci utama dalam penerbitan berindeks Scopus ini. Kajian ini memberikan gambaran bukan sahaja tentang trend dan perkembangan penyelidikan mengenai bahasa reseptif dan ekspresif ASD, malah juga hubungannya dalam jaringan yang lebih luas sama ada dari segi tajuk perbincangan, bentuk penerbitan, tajuk penyelidikan dan senarai penerbit. Dari segi kepentingan, kajian ini dapat dijadikan sebagai titik rujukan untuk kajian lanjutan yang berkaitan dengan bahasa reseptif dan ekspresif ASD terutama sekali dalam memahami dan memantau perkembangannya dalam konteks kini. Trends in Receptive and Expressive Language Studies in Autism Spectrum Disorder: A Bibliometric Analysis Abstract This study aims to analyse the trends in research on receptive and expressive language in Autism Spectrum Disorder (hereafter ASD) over the past decade (2014–2024). Receptive language refers to an individual’s ability to comprehend verbal communication, whereas expressive language involves the ability to convey ideas and information verbally. Individuals with ASD often experience challenges in both aspects of language, which in turn affect their communication development and social interaction. Employing a bibliometric analysis approach, particularly through the use of the VOSviewer application, this study analyses data extracted from the Scopus database to identify key research themes, document types, publication years, publication output, authorship patterns, publication venues, as well as networks of collaboration among researchers and institutions. The findings reveal that a total of 726 articles were identified in the Scopus database. This number indicates a significant increase in publications on “Receptive and Expressive Language in ASD” beginning in 2019 and continuing through 2024. Boston University emerges as the leading contributor in terms of publication output compared to other higher education institutions. In addition, prominent scholars such as Tager-Flusberg and Kasari have played a pivotal role in advancing scholarly discourse in this field. The keywords “Autism Spectrum Disorder,” “Receptive Language,” and “Expressive Language” are identified as the most frequently occurring keywords in the Scopus-indexed publications. This study provides not only an overview of the trends and developments in research on receptive and expressive language in ASD, but also highlights their interconnections within a broader scholarly network, including thematic focus, types of publications, research topics, and publication outlets. In terms of its significance, this study serves as a valuable reference point for future research related to receptive and expressive language in ASD, particularly in understanding and monitoring its development in the contemporary context.

Syntactic Challenges in ChatGPT-5’s Translation of English News Texts into Standard Malay: A Generative Transformation Model Syafiee, Amieziezaitul Syazlien Ezzeq Ezrynah Amirul; Yaqin, Lalu Nurul
Journal of Language and Literature Studies Vol. 6 No. 2 (2026): June
Publisher : Lembaga Penelitian dan Pemberdayaan Masyarakat (LITPAM)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36312/jolls.v6i2.4590

The growing use of AI translation has increased the need to evaluate whether generative models can produce grammatically accurate translations for underrepresented languages such as Malay. English-to-Malay translation remains challenging because Malay syntax requires accurate verb affixation, modifier placement, phrase ordering, and noun phrase + verb phrase (FN + FK) alignment in Standard Malay. This study examines the syntactic and lexical accuracy of ChatGPT-5 in translating English news texts into Standard Malay, with a focus on FN + FK structures. Grounded in Generative Transformation Theory and Nik Safiah Karim’s Malay grammar framework, the research analyzes how underlying English sentence structures are transformed into Malay surface structures. Using a qualitative descriptive linguistic design supported by descriptive error analysis, authentic bilingual sentences from the Borneo Bulletin were analyzed across social, economic, technological, cultural, and sports domains. The samples were selected from news sentences containing FN + FK structures, translated using a standardized ChatGPT-5 prompt, coded according to syntactic error categories, and validated through review by two qualified linguists and Malay grammar specialists. Findings show that ChatGPT-5 generally preserves complex sentence structures, including verb affixation, modifier placement, and phrase order, while accurately translating culturally embedded expressions and technical terms, such as mock cheque, Turnaround, and king of fruits. Minor syntactic deviations were observed, particularly in morphological mapping, modifier sequencing, and lexical narrowing, but they did not significantly affect the meaning. The study demonstrates ChatGPT-5’s potential as a supportive AI tool for multilingual translation and highlights the value of syntactically informed evaluation. These insights inform both the theoretical understanding of Malay syntax in AI translation and practical applications for professional, academic, and media contexts.

Charting the Future of Prompt Engineering: Critical Reflections on Methodology, Ethics, and Research Directions Verawati, Ni Nyoman Sri Putu; Yaqin, Lalu Nurul
International Journal of Essential Competencies in Education Vol. 4 No. 1 (2025): June
Publisher : Lembaga Penelitian dan Pemberdayaan Masyarakat (LITPAM)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36312/ijece.v4i1.2908

Prompt engineering has emerged as a transformative strategy for optimizing Large Language Models (LLMs), offering a cost-effective alternative to full model fine-tuning. In a recent bibliometric review, Fatawi et al. (2024) analyzed 437 Scopus-indexed publications from January 2022 to February 2024, using VOSviewer to identify key thematic clusters—including transformer architectures, deep learning innovations, and few-shot learning—and documenting a fivefold increase in related publications over the review period. Building on their macro-level mapping, this commentary extends the discussion by articulating the strategic and democratizing potential of prompt engineering while addressing critical gaps in methodology and ethical oversight. We critique the review’s reliance on a single English-language database, its exclusion of preprints and non-English sources, and its omission of qualitative insights into user practices and system impacts. In response, we offer concrete recommendations to guide future research: diversify data sources for bibliometric analysis, implement rigorous prompt audit frameworks, conduct longitudinal A/B testing in real-world environments, and adopt mixed-methods approaches to capture human-centered dynamics. We also explore emerging synergies—such as quantum-enhanced NLP and neuro-linguistic prompt design—as promising frontiers for advancing prompt optimization. By addressing these gaps, this commentary aims to ensure that prompt engineering evolves not only as a technical solution but as a responsible and inclusive foundation for next-generation AI development.

Title

Found 19 Documents
Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title Search

Found 19 Documents Search

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Abstract

Title

Found 19 Documents
Search