<![CDATA[ABSTRAK Kemajuan Akal Imitasi (AI) telah meningkatkan kualitas terjemahan melalui berbagai sistem terjemahan berbantuan AI yang menghasilkan terjemahan lebih cepat dan lebih akurat secara linguistik dibandingkan sebelumnya (Koehn, 2020; Toral & Way, 2018). Namun, AI masih menghadapi tantangan dalam mempertahankan makna budaya, khususnya pada Indigenous Cultural Knowledge Units (ICKUs) yang merepresentasikan identitas historis, sosial, dan budaya masyarakat adat. Penelitian ini bertujuan mengembangkan Cultural Meaning Preservation Assessment (CMPA), yaitu metode multidimensi untuk mengevaluasi kemampuan sistem AI dalam melestarikan makna budaya pada terjemahan ICKUs dari Kutai Adat Lawas. Penelitian menggunakan desain deskriptif kualitatif dengan korpus sebanyak 25 ICKUs yang diperoleh dari Badan Registrasi Wilayah Adat (BRWA) dan Indonesia Kaya. Sebagai acuan evaluasi, dikembangkan Human Gold Standard Translation (HGST) menggunakan teknik peminjaman dan deskripsi (Newmark, 1988). Selanjutnya, terjemahan yang dihasilkan oleh Google Translate, DeepL, ChatGPT, Gemini, dan Claude dievaluasi menggunakan enam dimensi CMPA, yaitu pelestarian leksikal, semantik, budaya, historis, pragmatik, dan identitas pribumi. Hasil penelitian menunjukkan bahwa meskipun sebagian besar sistem Akal Imitasi (AI) memiliki kinerja yang baik dalam kesetaraan leksikal dan semantik, pelestarian makna historis, pragmatik, dan budaya masih menjadi kelemahan utama. CMPA memberikan kontribusi metodologis sebagai kerangka kerja holistik untuk mengevaluasi pelestarian makna budaya dalam terjemahan berbantuan Akal Imitasi (AI). Keywords: Artificial Intelligence-Assisted Translation, Cultural Meaning Preservation Assessment (CMPA), Indigenous Cultural Knowledge Units (ICKUs), Translation Quality Assessment, Cultural Meaning Preservation, Kutai Adat Lawas. ABSTRACT Recent advances in Artificial Intelligence (AI) have significantly improved translation quality through the development of AI-assisted translation systems capable of producing faster and more linguistically accurate translations than previous approaches (Koehn, 2020; Toral & Way, 2018). Nevertheless, AI continues to face challenges in preserving cultural meaning, particularly when translating Indigenous Cultural Knowledge Units (ICKUs), which embody the historical, social, and cultural identities of Indigenous communities. This study aims to develop the Cultural Meaning Preservation Assessment (CMPA), a multidimensional assessment framework designed to evaluate the ability of AI-assisted translation systems to preserve the cultural meaning of ICKUs derived from Kutai Adat Lawas. The study employed a qualitative descriptive research design using a corpus of 25 ICKUs collected from the Badan Registrasi Wilayah Adat (BRWA) and Indonesia Kaya. A Human Gold Standard Translation (HGST) was developed as the evaluation benchmark using borrowing and descriptive translation techniques (Newmark, 1988). Translations generated by Google Translate, DeepL, ChatGPT, Gemini, and Claude were subsequently evaluated using the six dimensions of the CMPA framework: lexical, semantic, cultural, historical, pragmatic, and Indigenous identity preservation. The findings indicate that although most AI systems perform well in achieving lexical equivalence and semantic accuracy, preserving historical, pragmatic, and culturally embedded meanings remains a major challenge. The proposed CMPA makes a methodological contribution by providing a holistic framework for evaluating cultural meaning preservation in AI-assisted translation. Keywords: Artificial Intelligence-Assisted Translation; Cultural Meaning Preservation Assessment (CMPA); Indigenous Cultural Knowledge Units (ICKUs); Translation Quality Assessment; Cultural Meaning Preservation; Kutai Adat Lawas.]]>
Copyrights © 2026
