Publication Search

This study aims to examine the effectiveness of implementing forgiveness therapy in enhancing the personal growth of inmates at the Tangerang Class II-A Women’s Correctional Facility. The method used was a pre-experimental study with a one-group pretest-posttest design. Nine subjects aged 20-40 years were selected from 50 inmates based on low to moderate scores on the forgiveness scale and the psychological well-being scale. The intervention consisted of six sessions adapted from the forgiveness model developed by Enright and Fitzgibbons. Data analysis utilized a paired-sample t-test and the N-Gain formula. The results showed a significant increase in personal growth scores from the pretest (M = 7.00; SD = 1.225) to the posttest (M = 9.44; SD = 1.424), with t(8) = −8.315 and p < 0.001. An N-Gain value of 0.610 indicates moderate effectiveness. The follow-up measurement (M = 9.67; SD = 1.118) did not differ significantly from the post-test (p = 0.347), indicating the sustainability of the changes achieved. Forgiveness scores also increased from 65.33 (pre-test) to 86.67 (post-test) and 87.33 (follow-up), confirming the role of forgiveness as a psychological mechanism for personal growth. These findings support the integration of structured forgiveness therapy into correctional programs for women’s prisons.

The rapid expansion of digital finance services, particularly Buy Now Pay Later (BNPL) platforms such as Shopee PayLater, has significantly altered consumption patterns among university students. The aim of this study is to examine the effect of Shopee PayLater usage and lifestyle on consumptive behavior, as to test the moderating role of financial literacy among Accounting Education students at Universitas Negeri Semarang (class of 2022). A quantitative explanatory approach was employed, with data collected via Likert-scale questionnaires distributed to 63 respondents selected through purposive sampling. Data were analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM) with SmartPLS 4. Results show that Shopee PayLater usage positively and significantly affects consumptive behavior (path coefficient = 0.440, p = 0.000), as does lifestyle (path coefficient = 0.408, p = 0.003). Financial literacy, however, does not directly influence consumptive behavior (p = 0.676). Notably, financial literacy significantly strengthens—rather than weakens—the effect of lifestyle on consumptive behavior (interaction coefficient = 0.253, p = 0.019), while it does not moderate the relationship between Shopee PayLater usage and consumptive behavior (p = 0.712). These findings contribute theoretically by extending the Theory of Planned Behavior, demonstrating that BNPL accessibility and lifestyle orientation are stronger predictors of consumptive behavior than financial knowledge alone. Practically, this study suggests that financial literacy interventions must be accompanied by self-regulation strengthening and consumer protection policies to effectively curb excessive consumption driven by digital credit services.

This study aims to investigate students' perceptions of Qur'an learning (Tahsin and Tajwid) in improving their Qur'an recitation skills at Madrasah Diniyah Nurul Huda Bangodua, Cirebon. This study employed a qualitative approach using a descriptive method. The research participants consisted of 10 fourth-grade students and one teacher responsible for Qur'an learning. Data were collected through observation, interviews, and documentation. The data were analyzed using the Miles, Huberman, and Saldaña model, which includes data reduction, data display, and conclusion drawing. The findings indicate that students have positive perceptions of Tahsin and Tajwid learning. The learning process helps students improve the pronunciation (makhraj) of Arabic letters, understand the rules of Tajwid, enhance the fluency of Qur'an recitation, and develop greater confidence in reading the Qur'an. Supporting factors contributing to the success of the learning process include the talaqqi method, direct guidance from the teacher, and a conducive learning environment. The challenges faced by students include limited instructional time, difficulty distinguishing certain Tajwid rules, and insufficient practice in reading the Qur'an at home. Overall, Tahsin and Tajwid instruction plays a significant role in improving students' Qur'an recitation skills at Madrasah Diniyah Nurul Huda Bangodua, Cirebon.

Wastewater Treatment Plant (WWTP) pada kawasan industri konvensional umumnya masih mengandalkan strategi pemeliharaan berbasis interval waktu yang tetap. Pendekatan tersebut berisiko menyebabkan penurunan performa pompa yang tidak teridentifikasi secara dini serta meningkatkan potensi pemborosan energi operasional. Penelitian ini mengembangkan Green Maintenance Framework berbasis machine learning untuk meningkatkan reliabilitas pompa sirkulasi pada sistem Moving Bed Biofilm Reactor (MBBR). Analisis dilakukan menggunakan dataset telemetri multi-sensor yang mencakup parameter getaran, temperatur, tekanan, debit aliran, dan rotasi per menit (RPM). Proses rekayasa fitur diterapkan melalui pembentukan System Efficiency Index untuk meningkatkan sensitivitas model terhadap indikator degradasi kinerja pompa. Model prediktif dibangun menggunakan algoritma Random Forest Classifier dengan skema pembagian data 80:20 secara stratified. Hasil pengujian menunjukkan bahwa model menghasilkan tingkat akurasi klasifikasi sebesar 100%, dengan variabel Vibration dan Temperature menjadi parameter yang paling dominan dalam proses prediksi. Analisis operasional memperlihatkan bahwa degradasi pompa menyebabkan penurunan flow rate meskipun nilai rotasi per menit (RPM) mengalami peningkatan, sehingga memicu kenaikan konsumsi energi dan meningkatkan risiko gangguan pada proses biologis Moving Bed Biofilm Reactor (MBBR). Dari aspek ekonomi, kondisi tersebut menyebabkan pemborosan energi sebesar 5.623 kWh atau setara Rp6.271.236, - per bulan untuk setiap unit pompa. Penelitian ini berkontribusi pada pengembangan sistem predictive maintenance berbasis kecerdasan buatan untuk mendukung efisiensi energi serta implementasi green manufacturing di kawasan industri.

The development of digital services by BPJS Ketenagakerjaan through the JMO (Jamsostek Mobile) application has triggered a surge in large-scale and unstructured user reviews on the Google Play Store, thereby complicating manual analysis and conventional sentiment analysis in accurately identifying specific issues. This research aims to implement the Aspect-Based Sentiment Analysis (ABSA) method to granularly evaluate JMO application reviews based on specific aspects, while simultaneously addressing class imbalance and computational efficiency issues. The proposed method combines the pretrained IndoBERT model as a contextual feature extractor, the SMOTE technique to balance the training data, and an artificial neural network (Neural Network) as the classification layer without performing full fine-tuning. The dataset used consists of 90,268 unique reviews categorized into five main aspects through keyword matching, namely General Satisfaction/Complaints, Performance & Stability, Service & Support, Feature Quality, and UI/UX, with initial lexicon-based labeling using the InSet Lexicon. The research results indicate that the proposed model successfully achieves highly optimal performance with an accuracy rate of 91.81% and a weighted F1-score of 92%. Furthermore, the implementation of SMOTE proved effective in enhancing model reliability on the minority class (negative sentiment), achieving an F1-score of 89%. The implications of this research contribute an accurate and efficient aspect-based sentiment analysis framework for developers, and serve as a strategic evaluation tool for BPJS Ketenagakerjaan in mapping specific user complaints to accelerate continuous improvements in the performance, stability, and service quality of the JMO application.

This study aims to identify the pragmatic failures of Large Language Models (LLMs) and the biases of Anglophone-based AI moderation algorithms in detecting Indonesian hate speech expressed through sarcasm, satire, euphemism, and local cultural metaphors. It also examines the extent to which AI systems understand and interpret the pragmatic meanings within the corpus. This study employs a qualitative descriptive approach with a comparative design. Data were collected through the documentation of hate speech expressions on social media containing elements of local cultural hatred. The data were analyzed using qualitative descriptive methods with pragmatic and thematic approaches. The findings show that all corpus data contain political satire and indirect hate expressed through irony, sarcasm, absurd metaphors, and popular culture wordplay. Testing with Claude AI showed that the system was capable of identifying the data as implicit criticism and recognizing the pragmatic functions of emoticons and contextual meanings in the utterances. However, the analysis also demonstrated limitations in understanding local sociocultural contexts, particularly the metaphors “daun nangka” and “daun sawit,” which were interpreted merely as absurd humor. These findings indicate that AI detection accuracy does not necessarily reflect a deep pragmatic and cultural understanding within the Indonesian context.

LAMPUKITA merupakan salah satu UMKM VCO di Kabupaten Fakfak. Hasil dari pembuatan VCO usaha tersebut,  menghasilkan blondo yang selanjutnya diolah menjadi minyak goreng.  Berdasarkan aspek pemasaran,VCO LAMPUKITA memiliki prospek yang baik. Namun, ketersediaan bahan baku kelapa yang memenuhi syarat pembuatan produk utama, jumlahnya terkadang tidak tersedia.  Penelitian sebelumnya fokus pada teknik ekstraksi VCO, masih sedikit membahas strategi bisnis spesifik untuk minyak goreng hasil blondo ditingkat UMKM dengan kendala bahan baku.  Tanpa strategi pengembangan yang tepat, LAMPUKITA akan sulit bertahan secara finansial meskipun produknya memiliki nilai gizi lebih unggul. Tujuan penelitian untuk mengidentifikasi kekuatan, kelemahan, peluang dan ancaman,  menentukan faktor internal, eksternal strategi terbaik dan model pengembangan bisnis baru yang dapat diterapkan LAMPUKITA. Penentuan lokasi secara purposive. Jenis data yaitu data primer dan data sekunder.  Teknik pengumpulan data yaitu observasi, wawancara dan dokumentasi.  Analisis data yang digunakan adalah deskriptif kualitatif (analisis SWOT dan BMC) dan kuantitatif (metode IFAS, EFAS dan QSPM).  Hasil analisis SWOT menunjukkan bahwa kekuatan utama usaha LAMPUKITA terletak pada keunggulan produk alami, pemanfaatan blondo VCO, serta cita rasa khas kelapa Fakfak. Kelemahan utamanya adalah keterbatasan ketersediaan kelapa yang memenuhi standar VCO, kapasitas produksi masih kecil, dan keterbatasan modal. Analisis IFAS dan EFAS menunjukkan LAMPUKITA berada pada kondisi cukup mendukung untuk dikembangkan secara internal dan eksternal. Keberlanjutan usaha ditentukan oleh manajemen rantai pasok  kelapa. Strategi paling tepat dikembangkan adalah strategi kemitraan dengan petani kelapa berbasis standar VCO.  LAMPUKITA dapat menerapkan berbagai model bisnis baru untuk menanggulangi keterbatasan bahan baku.

Penelitian ini mengkaji kesiapan transformasi digital pada usaha penjualan suku cadang otomotif menuju Autonomous Supply Chain (ASC) dengan menggunakan model MIISI. Pendekatan kuantitatif digunakan melalui survei terhadap 32 responden dan dianalisis menggunakan Partial Least Squares–Structural Equation Modeling (PLS-SEM). Hasil penelitian menunjukkan bahwa tingkat kesiapan berada pada kategori sedang hingga tinggi. Di antara dimensinya, Integration memiliki nilai tertinggi, sementara Instrumentation dan Standardization menunjukkan kesenjangan terbesar dibandingkan kondisi ideal. Tingkat kematangan digital berada pada Level 2–3, yang mengindikasikan bahwa sistem sudah terintegrasi sebagian namun belum sepenuhnya otonom. Selain itu, hasil SEM (R² = 0,164) menunjukkan bahwa tidak terdapat pengaruh langsung yang signifikan terhadap ASC, meskipun hubungan antar dimensi bersifat signifikan dan bertahap. Dimensi Integration menjadi faktor paling dominan, yang menegaskan bahwa kesiapan menuju ASC memerlukan proses transformasi digital yang berkelanjutan, terstruktur, dan terintegrasi.

Penelitian ini bertujuan untuk menganalisis pengaruh struktur modal terhadap nilai perusahaan, menganalisis pengaruh profitabilitas terhadap nilai perusahaan, serta menganalisis pengaruh likuiditas terhadap nilai perusahaan pada perusahaan manufaktur sektor makanan dan minuman yang terdaftar di Bursa Efek Indonesia periode 2022–2024. Jenis penelitian yang digunakan adalah penelitian kuantitatif dengan pendekatan asosiatif. Populasi penelitian mencakup seluruh iverusahaan manufaktur sektor makanan dan minuman yang terdaftar di Bursa Efek Indonesia periode 2022–2024 yang berjumlah 83 perusahaan. Sampel penelitian sebanyak 48 perusahaan yang dipilih melalui iverusa purposive sampling berdasarkan kriteria iv erusahaan yang terdaftar secara berturut-turut selama periode penelitian, mempublikasikan laporan keuangan secara lengkap, tidak mengalami kerugian selama periode penelitian, serta memiliki data yang sesuai dengan variabel penelitian. Jenis data yang digunakan adalah data sekunder yang bersumber dari laporan keuangan tahunan perusahaan yang dipublikasikan melalui situs resmi Bursa Efek Indonesia. Teknik pengumpulan data dilakukan melalui metode dokumentasi dengan mengakses dan mencatat data keuangan yang relevan. Analisis data menggunakan analisis regresi data panel dengan bantuan aplikasi Eviews melalui uji pemilihan model yang meliputi uji Chow, uji Hausman, dan uji Lagrange Multiplier untuk menentukan model yang paling sesuai. Hasil penelitian menunjukkan bahwa struktur modal yang diproksikan dengan Debt to Equity Ratio (DER) berpengaruh terhadap nilai perusahaan, yang menunjukkan bahwa komposisi penggunaan utang dan ekuitas dalam pendanaan perusahaan dapat memengaruhi persepsi investor terhadap kinerja dan prospek perusahaan. Profitabilitas yang diproksikan dengan Net Profit Margin (NPM) juga berpengaruh terhadap nilai perusahaan, yang mengindikasikan bahwa semakin tinggi kemampuan iverusahaan dalam menghasilkan laba maka semakin tinggi pula nilai perusahaan karena meningkatnya kepercayaan investor terhadap kinerja perusahaan. Sementara itu, likuiditas yang diproksikan dengan Current Ratio (CR) tidak berpengaruh signifikan terhadap nilai perusahaan, yang menunjukkan bahwa kemampuan perusahaan dalam memenuhi kewajiban jangka pendek belum tentu menjadi pertimbangan utama bagi investor dalam menilai perusahaan pada sektor makanan dan minuman.

Penelitian ini bertujuan untuk menganalisis praktik pembelajaran tafsir Al-Qur'an menggunakan bahasa Bali sebagai media instruksional di Desa Pegayaman, Buleleng. Di tengah arus globalisasi yang mengancam eksistensi bahasa daerah, masyarakat Pegayaman menunjukkan resiliensi budaya melalui integrasi nilai-nilai Islam dengan kearifan lokal Bali. Melalui studi kasus pada kegiatan pengabdian masyarakat di Masjid Safinatussalam pada 04 April 2026, penelitian ini mengeksplorasi bagaimana Ustadz Zihni Hasan mentransformasikan pemahaman Surah Al-Baqarah kepada jamaah remaja putri dan ibu-ibu melalui medium bahasa Bali. Metode penelitian yang digunakan adalah kualitatif deskriptif dengan pendekatan etnopedagogi. Hasil penelitian menunjukkan bahwa penggunaan bahasa daerah bukan sekadar alat komunikasi, melainkan strategi "pribumisasi" Islam yang memperkuat identitas "Nyama Selam"( penduduk islam yang telah lama menetap dan mengadopsi budaya bali dalam kehidupan sehari-hari). Penggunaan bahasa ibu terbukti efektif menurunkan hambatan kognitif dan emosional dalam memahami teks suci, sekaligus berfungsi sebagai sarana konservasi bahasa daerah. Temuan ini menegaskan pentingnya pendidikan agama Islam berbasis budaya sebagai model moderasi beragama di Indonesia.

The following study aims to analyze the influence of buzzer marketing, influencer marketing, and brand trust on purchase intention of Wardah products on TikTok Shop. The background of the study is based on the increasing use of TikTok Shop as a social commerce platform and the importance of digital marketing strategies in influencing consumer purchasing intention, especially in the cosmetics industry that demands brand trust. The research method uses a quantitative approach with a questionnaire distributed to 100 respondents. Data were analyzed through validity and reliability tests, classical assumption tests, and multiple linear regression. The results indicate that all instruments are valid and reliable and the model meets classical assumptions. Partially, buzzer marketing, influencer marketing, and brand trust have a significant effect on purchase intention, with coefficients of 0.494, 0.362, and 0.264, respectively. Simultaneously, all three variables also have a significant effect. The coefficient of determination value of 0.535 indicates that the variation in purchase intention is explained by these three variables.

This study aims to examine the role of teachers in developing emotional intelligence among adolescent students at SMP Kesatrian Surakarta. Emotional intelligence is an essential aspect that supports academic achievement, social relationships, and character development. This research employed a qualitative approach using a case study method. Data were collected through in-depth interviews with 15 teachers teaching grades VII to IX, participant observations in classroom settings, and analysis of relevant school documents. The findings reveal that teachers play a strategic role in fostering students’ emotional intelligence through three primary approaches: emotional modeling by demonstrating appropriate emotional management (45%), facilitating emotion-based learning activities (30%), and evaluating students’ emotional intelligence development (25%). These roles significantly contribute to improving students’ self-awareness, self-regulation, empathy, and social skills. Supporting factors include emotional intelligence training programs provided by the local education authority and the school’s commitment to creating a supportive learning environment. However, major challenges consist of high administrative workloads among teachers (70%) and limited counseling facilities (50%). This study recommends continuous teacher training, the integration of emotional intelligence modules into lesson plans, and collaboration with school psychologists to ensure long-term monitoring and development of students’ emotional competencies. The findings contribute to the enrichment of emotional intelligence literature within the context of Indonesian secondary education.

Stunting is a multifactorial public health problem influenced by various risk factors that may emerge during the prenatal period. Early identification of stunting risk during pregnancy is important to support preventive interventions. This study aimed to develop a stunting risk prediction model based on maternal prenatal factors using the Random Forest algorithm. Secondary data from 172 pregnant women, consisting of 83 stunting cases and 89 non-stunting cases, were analyzed. The predictor variables included maternal age during pregnancy, height, hemoglobin level, mid-upper arm circumference (MUAC), smoking history, hypertension, asthma, and diabetes mellitus. The research stages consisted of data preprocessing, model training using Stratified 5-Fold Cross Validation, performance evaluation, external testing, and feature importance analysis. Internal evaluation results showed an accuracy of 60%, precision of 60.6%, recall of 57.3%, F1-score of 58.9%, and AUC of 0.6688. External testing yielded an accuracy of 70% and an AUC of 0.6167. Feature importance analysis identified maternal age during pregnancy as the most influential variable in the prediction process. The findings indicate that maternal prenatal factors have potential for early stunting risk identification, although the predictive performance remains moderate. This approach may serve as a foundation for developing early screening tools to support targeted interventions among high-risk pregnancies.

Correctional institutions impose sustained psychological burdens on inmates, including social isolation, fractured family bonds, and disrupted identity development, which collectively heighten the risk of mental health deterioration. Unlike preventive psychoeducational approaches, group counseling operates as a curative, therapeutically structured service that enables participants to process shared difficulties through interpersonal dynamics. This community service study implemented group counseling within a Participatory Action Research (PAR) framework involving 50 early adult inmates at Lapas Kelas III Rangkasbitung. PAR was operationalized as an implementation framework in which each counseling cluster cycle followed the Plan-Act-Observe-Reflect sequence, enabling iterative refinement of therapeutic content. Pre-service needs assessment using the Alat Ungkap Masalah (AUM) Umum Masyarakat instrument identified dominant problems in family relations (13.33%), personal self (12.97%), and social relations (10.76%). Group counseling was conducted across three thematic clusters. Program evaluation revealed improvements in social interaction, self-acceptance, and emotional regulation among participants, with the PAR reflective cycle enabling context responsive adaptation of each session. These findings affirm that PAR-grounded group counseling constitutes a contextually adaptive model for psychosocial rehabilitation in correctional settings.

This study analyzes the implementation of the Central Papua One-Stop Service Information System (Sipersateng) at the Central Papua Provincial Secretariat as a model for administrative innovation in the New Autonomous Region (DOB). The main focus of the study is to explore how a region with extreme infrastructure limitations can leapfrog or institutionalize towards an agile and transparent digital bureaucracy. Using a qualitative descriptive-analytical method with a case study approach, this research involved in-depth interviews, participant observation, and documentation studies to capture the dynamics of the transition from manual to digital systems. The results show that Sipersateng successfully reduced administrative process time by up to 60%. This success was driven by three main dimensions: institutional legitimacy through strong leadership commitment, technical resilience through offline-syncing features to address internet signal fluctuations, and significant changes in work culture. Adaptation strategies such as peer-mentoring methods for senior officials and mandatory policies through gubernatorial instructions proved effective in breaking down organizational cultural resistance. This study concludes that Sipersateng is not merely a technical tool, but rather a manifestation of Frontier Digital Bureaucracy that is able to mitigate the risk of maladministration and build public accountability in the newly formed province. This innovation demonstrates that geographic limitations are not a major barrier to achieving clean governance if supported by inclusive and adaptive system design. These findings provide a theoretical contribution to the public management literature on digitalization in regions with high structural challenges.

Collaboration between classroom teachers and Islamic Religious Education (PAI) teachers constitutes an essential strategy for instilling religious moderation values in elementary schools. This study aims to: (1) describe the forms of collaboration developed between classroom teachers and PAI teachers; (2) assess the effectiveness of such collaboration in inculcating religious moderation values; (3) identify changes in students' attitudes as tangible outcomes of collaboration; and (4) analyze supporting and inhibiting factors of collaboration at SDN 3 Wanamekar. A descriptive qualitative approach was applied using field study methods. Data were gathered through in-depth interviews, participatory observation, and document review. Data analysis followed the interactive model of Miles, Huberman, and Saldana, encompassing data condensation, data display, and conclusion drawing. Data credibility was ensured through source triangulation and technique triangulation. Findings indicate that collaboration occurs in two primary forms: reactive collaboration as a spontaneous response to students' intolerant behavior, and planned collaboration involving curriculum coordination, partial co-teaching, and joint evaluation. Collaboration was assessed as sufficiently effective yet suboptimal due to time constraints and excessive administrative burdens. Positive changes in students' tolerance and inclusive attitudes were identified as concrete outcomes. Supporting factors include favorable interpersonal communication, principal support, and moderation value integration in the Merdeka Curriculum. Main inhibiting factors are time limitations, administrative burdens, shallow conceptual understanding of religious moderation, and the absence of a standardized collaboration model. The study recommends establishing structured collaboration schedules and religious moderation training programs for all teachers.

Income inequality is a structural development problem that can trigger social conflict, affect long-term economic growth, and hinder the achievement of SDGs. The purpose of this study is to analyze the factors that contribute to income inequality in East Java Province by combining demographic, social, and economic factors. The data used includes panel data from 38 districts and cities in East Java Province between 2017 and 2023 period. The analysis method used is panel data regression with a fixed effect model. To measure the effect of variables on income inequality, the testing was conducted partially and simultaneosly. The individual test results show that elderly dependency and poverty have a positive and significant relationship, while the human development index and economic growth are proportional but not significantly, and the open unemployment rate is not significant negative. Simultaneously, all variables have a significant effect on income inequality of 55.67 percent, with the remaining 44.37 percent influenced by variables outside the scope of this study.

This research aims to develop audio-visual learning media based on the Discovery Learning model to improve the ability to write news texts in grade XI students of SMK Mutiara Qolbu Cianjur. The Research and Development (R&D) method is used by referring to the development model which includes the stages of identification of potentials and problems, data collection, product development, product validation, limited scale trials, product revisions, large-scale trials, final revisions, and final products. The research subjects consisted of 103 students in grade XI of SMK Mutiara Qolbu Cianjur which were divided into three classes. Data collection techniques were carried out through observation, interviews, questionnaires, expert validation, and news text writing tests. Data were analyzed using quantitative and qualitative descriptive techniques. The results of the study show that audio-visual learning media based on the Discovery Learning model obtained a very good feasibility level based on the results of the validation of material experts with a percentage of 91% and the validation of media experts of 89%. The results of the trial showed that the learning media received a positive response from students with an average percentage of 87.25%. In addition, the use of Discovery Learning-based audio-visual learning media has proven to be effective in improving students' ability to write news texts. This is shown by the increase in the average score of students from 64.21 in the pre-test to 84.37 in the post-test. The use of audio-visual media also improves students' learning activities, such as active questioning, discussion participation, and the ability to identify news elements.

The development of Artificial Intelligence (AI), particularly Large Language Models such as ChatGPT, has changed the paradigm of creative writing learning in secondary schools. However, empirical evidence regarding the relationship between the intensity of this technology utilization and the quality of students' literary works is still limited. This study aims to examine the relationship between AI utilization and the poetry writing skills of grade XI students at Karya Budi Putussibau Senior High School. The method used was a quantitative correlational method with a sample of N = 27 students selected through a random sampling technique. Data were collected using an AI utilization questionnaire (30 valid and reliable statements, (α= 0.975) and a poetry writing test assessed with an analytical rubric. The results of descriptive statistical analysis showed that the level of AI utilization was in the high category (M =    3.70) and students' poetry writing skills were in the good category (M =  78.41). The Pearson Product Moment correlation test showed a very strong and significant positive relationship between the two variables (r = 0.978; p < 0.01). These findings confirm that proper AI integration acts as a creative catalyst that helps students explore diction and imagery structure in the poetry writing process.

This research is motivated by the low poetry writing skills of grade VIII B students at MTs Mujahidin Pontianak, evidenced by a pre-cycle average score of 64% and a classical completeness rate of only 34%. The primary issues include students' difficulties in developing imagination, word choice (diction), and the limited use of innovative learning media. This study aims to improve poetry writing skills through the implementation of the Project Based Learning (PjBL) model assisted by Canva digital media. This study employed Classroom Action Research (CAR) conducted in two cycles, comprising planning, action, observation, and reflection phases. The research subjects consisted of 23 students. The results indicated a significant improvement in each cycle. The students' average score increased to 80% in cycle I and reached 87% in cycle II. Classical completeness also rose from 65% in cycle I to 86% in cycle II. The use of Canva was proven to stimulate students' creativity and motivation in expressing ideas into aesthetic poetry. The conclusion of this study is that the integration of the PjBL model and Canva media is effective in improving the quality of both the process and the learning outcomes of poetry writing. It is suggested that teachers adopt this model as an innovative alternative for literature learning in the digital era.