<![CDATA[Abstrak Kecamatan Doko, Kabupaten Blitar, merupakan wilayah dengan keragaman budaya dan agama yang tinggi. Kondisi ini menuntut guru sekolah dasar untuk mampu merancang pembelajaran yang adaptif dan inklusif. Namun, hasil identifikasi kebutuhan menunjukkan 75% guru belum pernah mengikuti pelatihan pembelajaran berdiferensiasi dan multikultural berbasis deep learning, serta 80% masih menggunakan modul ajar yang kurang mempertimbangkan keberagaman siswa. Program pengabdian ini bertujuan meningkatkan kompetensi guru dalam menyusun perangkat pembelajaran berdiferensiasi-multikultural dengan dukungan teknologi kecerdasan artifisial dengan mitra guru di Gugus 1 Kec. Doko Kab Blitar sebanyak 50 guru. Metode yang digunakan yakni pendekatan partisipatif-kolaboratif dengan siklus plan-do-see selama enam bulan, meliputi perencanaan, pelatihan, implementasi, serta refleksi. Hasil kegiatan menunjukkan peningkatan signifikan kompetensi guru dengan kenaikan rata-rata skor pretest 70 menjadi posttest 90. Selain itu, guru berhasil menghasilkan produk inovatif berupa modul tematik multikultural, e-modul, video interaktif, serta LMS sederhana berbasis Google Sites. Program juga menghasilkan terbentuknya komunitas belajar digital sebagai wadah keberlanjutan. Dengan demikian, pengabdian ini terbukti efektif mendorong guru untuk merancang pembelajaran yang lebih adaptif, multikultural, dan berbasis teknologi. Kata kunci: pembelajaran berdiferensiasi; multikultural; deep learning; sekolah dasar. Abstract Doko Subdistrict, Blitar Regency, is an area characterized by high cultural and religious diversity, requiring elementary school teachers to design learning experiences that are adaptive and inclusive. However, the needs assessment revealed that 75% of teachers had never participated in training on differentiated and multicultural learning based on deep learning, and 80% were still using teaching modules that insufficiently considered student diversity. This community service program aimed to improve teachers’ competence in developing differentiated-multicultural learning tools supported by artificial intelligence technology in collaboration with 50 teachers from Gugus 1, Doko Subdistrict, Blitar Regency. The program applied a participatory-collaborative approach using a plan-do-see cycle over six months, encompassing planning, training, implementation, and reflection. The results showed a significant improvement in teachers’ competence, as indicated by an increase in the average pretest score from 70 to a posttest score of 90. In addition, the teachers successfully produced innovative learning products, including multicultural thematic modules, e-modules, interactive videos, and a simple LMS based on Google Sites. The program also resulted in the formation of a digital learning community as a sustainable platform for continued collaboration. These findings indicate that the community service program was effective in encouraging teachers to design learning that is more adaptive, multicultural, and technology-based. Keywords: differentiated learning; multicultural; deep learning; elementary school.]]>
