Study reviews community-centric translation methodology

AvodahConnect commissioned expert translator to review its approach

---

Communities around the world strive to retain their cultural identity. Limited by lack of educational materials, high costs, and the challenges of traditional translation models, people in these overlooked communities often struggle to preserve their language. 

AvodahConnect’s AI-assisted translation empowers communities and reshapes their stories. To dig deeper into the organization’s approach, partnerships, and technology, AvodahConnect commissioned Dr. Gilles Gravelle, a respected translation consultant with decades of experience and a PhD in linguistics. 

In a new white paper, Dr. Gravelle describes how community-centric translation is a faster, more cost-effective method of preserving low-resource languages.

Community-based translation saves costs and time 

Cost and time are two barriers that keep communities from translating more resources into their own languages. AvodahConnect reduces both costs and time by using the Bible to begin every community-centric translation project.

Seven times longer than the average novel, the Bible is full of rich stories, themes, and emotions, all shared through the use of a massive vocabulary. At over 783,000 words, its length, breadth, and global availability make it an effective foundation for underserved communities’ translation efforts. 

AvodahConnect’s unique combination of technology, local expertise, and common purpose centers the community and creates unity. 

Shared priorities guide translation processes

Dr. Gravelle’s review reveals AvodahConnect’s community-centric practices. While AI-supported translation is a technical endeavor, at AvodahConnect, it’s rooted in connection and collaboration. 

In the white paper, he describes how the organization emphasizes:

• Human-centered technology 

  Prioritizing lasting connections over automation, AvodahConnect strives for collaboration. The AvodahConnect team works with local people to design processes that encourage community-led ownership of the Bible translation projects. 

• Deep partnerships 

  AvodahConnect partners with local organizations and defines clear roles for their contributors, plus AI developers and translators. People from the community drive the translation effort.

• Iterative translation workflows 

  Translators collect data, edit AI-generated drafts, and undergo peer reviews and quality assurance cycles. These processes foster continuous learning that yields high-quality translation and builds local translation capacity.

Throughout their translation projects, local people do more than participate — they run the processes that preserve their language. AvodahConnect’s approach draws people into the work, bridging the gap between technology and humans and helping everyone feel invested in their community’s success. 

Technology excels at translating low-resource languages

Analyzing the methodology, Dr. Gravelle found that AvodahConnect uses small language models (SLMs). An SLM learns from a small number of written materials, paying special attention to unique words, grammar, and style in the language. 

This narrower method is highly effective for low-resource languages because the SLM trains on specific, curated material. Using the Bible as training material gives SLMs a wide range of vocabulary, grammar, and narratives from a culturally significant text. 

When combined with AvodahConnect’s years of iteration, review, and community validation, translations become faster, more cost-effective, and of higher quality.

Expert evaluation highlights process and partnership 

Dr. Gravelle conducted an in-depth study of AvodahConnect’s translation model and community partnerships. His white paper explains how community-led, AI-assisted Bible translation ensures high quality and contextual relevancy.

Discover more about how communities are reclaiming their languages with AvodahConnect. Download the white paper now.