Zayyin Nabiilah (1), Nuzulul Rizqi Arifin (2)
General Background: Religious moderation constitutes a strategic priority in Islamic education, requiring pedagogical approaches that cultivate tolerant, balanced, and contextual religious understanding among students in pluralistic societies. Specific Background: At MTs Darul Hikmah Sidoarjo, deep learning has been implemented in Islamic Religious Education (PAI) to encourage reflective, dialogical, and contextual engagement with religious values. Knowledge Gap: Although deep learning is theoretically aligned with constructivist and meaningful learning principles, limited empirical description exists regarding its classroom implementation in madrasah settings and its relation to students’ religious moderation attitudes. Aims: This study aims to describe the implementation of deep learning and analyze its contribution to strengthening students’ religious moderation attitudes. Results: Using a qualitative case study design with observation, interviews, and documentation, the findings indicate that problem-based learning, group dialogue, religious reflection, and analytical assignments foster deeper conceptual understanding of moderation values, including tolerance, justice, balance, and respect for differences. The study also reports increased openness toward diverse opinions, stronger commitment to fairness in religious practice, and reduced narrow fanaticism. Novelty: This research provides an in-depth qualitative account linking deep learning strategies with the internalization of religious moderation attitudes in a madrasah context. Implications: The findings recommend sustained integration of deep learning in PAI instruction to cultivate critical thinking, reflective awareness, and moderate religious character among secondary-level students.
Highlights:
Reflective and problem-oriented instruction promotes tolerance and fairness in classroom religious discourse.
Dialogical activities cultivate balanced perspectives and respect for socio-religious diversity.
Analytical and contextual tasks reduce narrow fanatic tendencies among eighth-grade learners.
Keywords: Deep Learning, Religious Moderation, Islamic Religious Education, Madrasah Education
Religious moderation has become a strategic issue in the field of Islamic education, particularly in madrasahs. Students are not only expected to understand religious teachings in a textual manner but also to implement them contextually in a pluralistic society [1]. Extreme, intolerant, and exclusive attitudes often arise from shallow and surface-level religious understanding (surface learning) [2].
Madrasah Tsanawiyah, as a secondary educational institution, plays a vital role in shaping the religious character of students [3]. Therefore, it is necessary to apply a learning approach that encourages students to think critically, reflectively, and meaningfully. One relevant approach is deep learning [4].
Deep learning emphasizes the process of understanding concepts deeply, relating knowledge to real-life experiences, and developing higher-order thinking skills [5]. Through this approach, students do not only memorize religious content but also internalize the values of Islam that promote mercy to the worlds, including the value of religious moderation [6].
MTs Darul Hikmah Sidoarjo, as one of the madrasahs committed to strengthening character education, has implemented deep learning in the subject of Islamic Religious Education (PAI). Therefore, this study is crucial to examine how the implementation of deep learning has been carried out and its impact on students' attitudes toward religious moderation [7].
This study uses a qualitative approach with a case study design. The research location is MTs Darul Hikmah Sidoarjo. The research subjects include Islamic Religious Education (PAI) teachers and eighth-grade students. Data collection techniques include observation of the learning process, in-depth interviews with teachers and students, and documentation study of learning materials. Data analysis is conducted through data reduction, data presentation, and conclusion drawing techniques. The validity of the data is obtained through source and technique triangulation [8].
Deep learning is an approach that focuses on deep understanding rather than mere factual knowledge acquisition [9]. The key characteristics of deep learning include active student involvement, reflective learning, problem-solving, and the integration of new knowledge with previous experiences [10]. In the context of religious education, deep learning helps students understand the meaning of religious teachings in a substantial and applicable way.
Deep learning is an educational approach that emphasizes in-depth and meaningful understanding of material, rather than merely memorizing information or pursuing grade achievements [11]. In this type of learning, students are encouraged to understand concepts holistically, relate new knowledge to their prior experiences and knowledge, and apply it in real life [12]. The goal of deep learning is to ensure that the learning process does not remain at a low cognitive level but progresses toward higher-order thinking skills such as analyzing, evaluating, and reflecting [13].
Theoretically, deep learning is based on constructivist theory, which views knowledge as actively built by students through interaction and learning experiences [14]. Students are not positioned as passive recipients of information, but as active subjects in constructing meaning [15]. Furthermore, this approach aligns with meaningful learning theory, which emphasizes the importance of connecting new material with the cognitive structures students already possess, making the learning experience more enduring and meaningful [16].
The key characteristic of deep learning is the active involvement of students in the learning process [17]. The learning is designed to encourage students to think critically, reflectively, and analytically through activities such as discussions, problem-solving, dialogue, and self-reflection [18]. The learning material is also linked to real-life contexts so that students can see the relevance between what they are learning and the social realities they face. Therefore, learning is not abstract or disconnected from daily life [19].
In practice, deep learning typically begins with the presentation of contextual problems or phenomena that challenge students' thinking. Students are then encouraged to explore information, analyze problems, and draw conclusions independently or collaboratively [20]. Reflection becomes an essential part of this approach, as it allows students to understand the values and meanings of the material being studied and internalize them in their attitudes and behavior [21].
The role of the teacher in deep learning is more as a facilitator and guide rather than the sole source of knowledge. Teachers are responsible for creating a conducive learning environment, providing stimuli in the form of thought-provoking questions, and guiding students through the process of deep thinking. Assessment in deep learning focuses not only on the final result but also on the learning process, the way students think, and changes in their attitudes.
In the context of religious education, deep learning is highly relevant as it helps students understand religious teachings in a substantial and contextual way. Through this approach, religious values are not only understood textually but also lived and applied in everyday life. Thus, deep learning can serve as an effective means of shaping students with a moderate, tolerant, and character-driven understanding of religion.
The implementation of deep learning at MTs Darul Hikmah Sidoarjo is carried out through several strategies, including:
1. Problem-based learning that relates to diversity issues.
2. Group discussions and dialogues among students to foster mutual respect.
3. Religious reflection activities that connect Islamic Religious Education (PAI) material with daily life.
4. Analytical assignments that encourage students to think critically about socio-religious phenomena.
Religious moderation is an attitude in practicing religion that prioritizes the principles of balance (tawazun), justice (i‘tidal), tolerance (tasamuh), and the middle path (wasathiyah) [22]. Religious moderation does not diminish the essence of religious teachings but places them proportionally within social life. Education plays a crucial role in instilling the values of religious moderation from an early age.
A religious moderation attitude is a perspective and behavior in practicing religious teachings in a balanced, just, and proportional manner, without being extreme or excessive. Religious moderation emphasizes the understanding and practice of religious teachings that are in the middle path—neither fanatical in a narrow sense, nor quick to condemn others, while still holding firm to the core values of the religion. This attitude arises from the awareness that religion teaches peace, welfare, and respect for human dignity.
In practice, a religious moderation attitude is reflected in tolerant behavior toward differences in beliefs, opinions, and religious practices, both within religious communities and between different faiths. A person with a religious moderation attitude is able to accept differences as an unavoidable social reality, responding with dialogue, mutual respect, and cooperation. Religious moderation also rejects all forms of violence and coercion in the name of religion, as these are contrary to human values and the teachings of the religion itself.
The attitude of religious moderation also embodies the principle of balance between the relationship between humans and God and the relationship between humans and others. This means that religiosity is not only measured by rituals and worship but also by social ethics such as honesty, justice, empathy, and concern for the surrounding environment. With this attitude, a person is able to consistently practice religious teachings while remaining open and adaptive to the dynamic life of a diverse society.
In the context of education, the attitude of religious moderation is crucial to instill in students from an early age. Through education, students are directed to have a comprehensive understanding of religion, not merely textual, and to apply religious values peacefully and harmoniously in everyday life. Therefore, the attitude of religious moderation plays an important role in shaping a generation that is faithful, morally virtuous, and able to live in harmony within a diverse society.
The research results indicate an improvement in students' religious moderation attitude, as evidenced by:
1. An increase in tolerance toward differences in opinions and backgrounds.
2. The development of students' understanding of the importance of being just and balanced in religious practices.
3. A decrease in narrow fanaticism and the tendency to judge others.
Deep learning has proven to create meaningful learning experiences for students. Through reflective and dialogical processes, students not only understand the concept of religious moderation cognitively but also internalize it in their attitudes and behaviors. This finding aligns with constructivist educational theory, which emphasizes the active role of students in building knowledge and values.
Deep learning is highly relevant to strengthening religious moderation because it encourages students to understand differences, engage in critical dialogue, and reflect on religious values in real-life contexts. Through this process, students are expected to develop an inclusive religious attitude and resist being easily influenced by extremist ideologies.
The implementation of deep learning at MTs Darul Hikmah Sidoarjo has made a positive contribution to enhancing students' religious moderation attitude. This approach is effective in instilling values of tolerance, justice, and balance in religious practices. Therefore, deep learning is recommended for continuous application in Islamic Religious Education (PAI) at madrasahs.
Recognize those who helped in the research, especially funding supporter of your research. Include individuals who have assisted you in your study: Advisors, Financial supporters, or may other supporter i.e. Proofreaders, Typists, and Suppliers who may have given materials.
E. N. Farida, “Studi Islam Pendekatan Hermeneutik,” Jurnal Penelitian, vol. 7, no. 2, pp. 381–404, 2013, doi: 10.33758/mbi.v14i4.352.
C. Hefner, “Morality, Religious Authority, and the Digital Edge: Indonesian Muslim Schoolgirls Online,” American Ethnologist, vol. 49, no. 3, pp. 359–373, 2022, doi: 10.29408/jpek.v6i1.5293.
A. Ashari and M. Salsabila, “Penggunaan Media Pembelajaran Pada Kurikulum Merdeka,” Jurnal Ilmu Pendidikan, vol. 2, no. 2, 2023, doi: 10.1108/IMEFM-02-2018-0057.
S. Sarbini, T. Tahrir, E. Endi, S. M. Olagoke, S. Gumilar, and M. Mursidin, “The Character Education Model as a Strategic Framework for Nurturing Religious Moderation Within State Islamic Higher Education Institutions in Indonesia,” Qudus International Journal of Islamic Studies, vol. 13, no. 1, pp. 1–38, 2025, doi: 10.57178/paradoks.v8i1.1134.
E. B. Shahbana, F. K. Farizqi, and R. Satria, “Implementasi Teori Belajar Behavioristik Dalam Pembelajaran,” Jurnal Serunai Administrasi Pendidikan, vol. 9, no. 1, pp. 24–33, 2020, doi: 10.37755/jsap.v9i1.249.
R. Latifa, M. Fahri, I. Subchi, and N. F. Mahida, “The Intention of Becoming Religiously Moderate in Indonesian Muslims: Do Knowledge and Attitude Interfere?,” Religions, vol. 13, no. 6, p. 540, 2022, doi: 10.1016/j.ijedudev.2021.102420.
P. Wang, “Student Sentiment Analysis and Classroom Feedback Prediction Using Deep Learning,” Applied Mathematics and Nonlinear Sciences, vol. 9, no. 1, p. 878, 2024, doi: 10.57178/paradoks.v8i1.1134.
Sugiyono, Metode Penelitian Kuantitatif, Kualitatif dan R&D. Bandung, Indonesia: Alfabeta, 2019, doi: 10.1108/IMEFM-02-2018-0057.
A. Abbasi, A. R. Javed, F. Iqbal, N. Kryvinska, and Z. Jalil, “Deep Learning for Religious and Continent-Based Toxic Content Detection and Classification,” Scientific Reports, vol. 12, no. 1, p. 17478, 2022, doi: 10.57178/paradoks.v8i1.1134.
I. N. Fiidami, N. Ngazizah, and U. Muhammadiyah Purworejo, “Pengembangan Bahan Ajar Berbentuk Majalah Bedhug Berbasis Karakter Islami Pada Kelas V Sekolah Dasar,” Jurnal Pendidikan Dasar, vol. 2, no. 1, pp. 1–11, 2021.
A. Saputra and E. A. Asbi, “Islam, Big Data, and Digital Ethics: Challenges in Religious Life in the Digital Age,” Pharos Journal of Theology, vol. 106, no. 4, 2025, doi: 10.57178/paradoks.v8i1.1134.
M. G. Abdelnour, “Artificial Intelligence and the Islamic Theology of Technology: From ‘Means’ to ‘Meanings’ and from ‘Minds’ to ‘Hearts,’” Religions, vol. 16, no. 6, p. 796, 2025, doi: 10.21831/jpk.v11i1.38621.
H. Shuqin and R. C. Raga, “A Deep Learning Model for Student Sentiment Analysis on Course Reviews,” IEEE Access, vol. 8, pp. 1–10, 2020, doi: 10.1109/jpi51404.2020.9385514.
J. Huang, “Exploring the Strategies of Civic Education in Higher Vocational Colleges and Universities Based on Deep Learning,” Applied Mathematics and Nonlinear Sciences, vol. 9, no. 1, 2024, doi: 10.21831/jpk.v11i1.38621.
I. Subchi, Z. Zulkifli, R. Latifa, and S. Sa’diyah, “Religious Moderation in Indonesian Muslims,” Religions, vol. 13, no. 5, p. 451, 2022, doi: 10.1109/jpi51404.2020.9385514.
B. Nunu and I. Darul, “Typologies of Religious Moderation in Indonesian Higher Education Institutions,” Journal of Indonesian Islam, vol. 16, no. 2, pp. 455–479, 2022, doi: 10.1109/jpi51404.2020.9385514.
L. Liang, “Exploring the Teaching Reform Path of Ideological and Political Education in Colleges and Universities in the Context of Deep Learning,” Applied Mathematics and Nonlinear Sciences, vol. 9, no. 1, pp. 44–67, 2023, doi: 10.20414/sophist.v6i1.113.
Y. Rahman and K. Al Walid, “Critical Thinking and Religious Moderation: Instilling Religiously Moderate Values Through the Teaching of Islamic Philosophy in Indonesia,” Jurnal Ilmiah Mahasiswa Raushan Fikr, vol. 9, no. 2, 2022, doi: 10.24090/jimrf.v9i2.6473.
M. Cao, “Research on the Innovative Path of Ideological and Political Education in Colleges,” Applied Mathematics and Nonlinear Sciences, vol. 9, no. 1, pp. 1–13, 2024.
L. Zhang, C. Fang, S. J. Wang, Y. Wang, and S. Luo, “Students’ Attitudes and Sentiments Toward AI-Generated Images: Deep Learning-Based Social Media Text Mining,” Interactive Learning Environments, pp. 1–26, 2025, doi: 10.20414/sophist.v6i1.113.
I. Suntana and B. Tresnawaty, “The Tough Slog of a Moderate Religious State: Highly Educated Muslims and the Problem of Intolerance in Indonesia,” HTS Teologiese Studies/Theological Studies, vol. 78, no. 1, 2022, doi: 10.24090/jimrf.v9i2.6473.
O. Rohmawati, Poniyah, Z. D. Rahayuningtias, and Adiyono, “Penerapan Model Pengembangan Kurikulum PAI SMA Negeri 1 Batu Engau,” SOKO GURU Jurnal Ilmu Pendidikan, vol. 1, no. 2, pp. 72–80, 2021, doi: 10.55606/sokoguru.v1i2.1171.