AI often leaves out minority languages – but some communities are turning it into a lifeline
AI-summarised brief · reviewed before publication
Artificial intelligence is widening linguistic inequality as large‑language models favor majority tongues, leaving minority languages under‑represented. Languages such as Hula in Papua New Guinea (≈10,000 speakers) and Tetun in Timor‑Leste (≈1 million speakers) receive scant training data, resulting in poor or absent AI output. Yet the technology is enabling grassroots solutions. The Hula community created Vavanagi, a documentation platform built with AI‑assisted coding, now hosting over 80 users and 12,000 English‑Hula translations toward a translator app. In Timor‑Leste, the Tulun platform lets health educators adapt AI‑driven translations of materials into Tetun, maintaining local term control. A Dinka‑English dictionary, developed with AI support, expands resources for the 5 million‑speaker language. These initiatives illustrate how AI can be repurposed to preserve endangered languages.
💡 Why It Matters
- · Community‑driven AI tools turn a technology that marginalizes minority tongues into a practical means of language preservation.
