WEAVING WORDS ACROSS WORLDS: THE SOCIO-TECHNICAL DANCE OF MACHINE TRANSLATION IN PRESERVING CULTURAL NUANCES
Abstract: Neural machine translation (NMT), powered by natural language processing and transformer models, has revolutionized global communication, achieving high accuracy in high-resource languages like English and Arabic. However, it struggles to preserve cultural nuances—context-specific linguist...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Ziglôbitha
2025-03-01
|
| Series: | Ziglôbitha |
| Online Access: | https://www.ziglobitha.org/wp-content/uploads/2025/05/22-Art._Hadja-BOUFRIOUA-pp.299-308.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract: Neural machine translation (NMT), powered by natural language processing and transformer models, has revolutionized global communication, achieving high accuracy in high-resource languages like English and Arabic. However, it struggles to preserve cultural nuances—context-specific linguistic elements such as proverbs, idioms, and legal terms—in low-resource African languages, including Algerian Arabic (darja), Tamazight, and Yoruba. Through a literature review and case studies, this study evaluates NMT tools like Google Translate and ChatGPT 4, highlighting their precision in data-rich languages but noting failures in translating darja proverbs, Tamazight sayings, and Arabic legal terms. It addresses three questions: How do data biases influence translation outcomes? What factors limit NMT’s effectiveness in African languages? Can human-AI collaboration preserve cultural heritage? Drawing on Algeria’s linguistic diversity, the study proposes community-driven archives and ethical AI frameworks to enhance cultural fidelity. Aimed at translation scholars, AI ethicists, and advocates for Africa’s 2,000 languages, it confronts the risk of linguistic erasure while envisioning technology that safeguards cultural identities in the digital age.
Keywords: Neural Machine Translation, Cultural Nuances, African Languages, Human-AI Collaboration, AI Ethics, Algerian Arabic, Tamazight.
|
|---|---|
| ISSN: | 2708-390X 2709-2836 |