Skip to content ↓

August 2025 - Year 27 - Issue 4

ISSN 1755-9715

Lost for Words? How AI is Preserving Endangered Languages

Nigel Balfour has been active in English language teaching and training since 1992, building a career across diverse roles and countries, including Thailand, Japan, Oman, the UK, France, and Spain. As a teacher, CELTA tutor, assessor, and examiner, he has collaborated with educators and learners from numerous cultural and professional backgrounds. Passionate about the role of new technologies in ELT, he likes to stay up to date with innovations in teacher development and language research. Based in Barcelona, Nigel is dedicated to advancing the future of English education. Email: nigel.balfour@gmail.com

 

Editorial

The text was originally published at tldrelt.

Although nearly 7,000 languages are spoken worldwide, one vanishes every two weeks. With 600 lost in the past century, linguistic diversity is under serious threat. You might be pleased to hear that, according to an article by Daniel Oberhaus in the Atlantic, AI is stepping in as a surprising ally to preserve & even revitalise these vanishing tongues.

Why it matters: Languages hold cultural identity, history & knowledge systems. But globalisation, digital exclusion & government policies are accelerating their extinction. Today, fewer than 5% of languages are captured digitally, leaving the rest largely excluded from online spaces.

AI to the rescue: LLMs like Google’s Gemini can help by processing endangered languages—even those with limited data. For example, Google’s 1,000 Languages Initiative used a single grammar book to train AI in an endangered Indonesian language.

By processing endangered languages, LLMs make these languages accessible online, providing educational tools for younger speakers to learn their heritage languages & resources that open economic opportunities, such as online platforms & tools in local languages.

The challenge: Endangered languages often lack written resources or standardised scripts. By some estimates, 50% of web content is in English, leaving most languages underrepresented. Despite efforts like IrokoBench, which benchmarks African languages [identifying specific weaknesses in LLMs for low-resource languages], AI’s ability to tackle these challenges remains a work in progress. Obviously AI is no magic bullet, but it could serve as a useful tool in a linguist’s toolbox [to mix my metaphors].

Teacher Takeaways: Some sources to engage with the topic in class-

The website Living Tongues is an interesting starting point for raising awareness of language preservation. Its Living Dictionaries app gives you a chance to explore languages, their location & what they sound like.

Woolaroo, part of a project by Google Arts & Culture, allows users to take pictures of their surroundings & hear the words for those objects in 17 endangered languages. Languages include Tamazight, Potawatomi, Māori, Maya & Sanskrit, highlighting the cultural significance & efforts to keep these languages alive.

Have you ever taught a speaker of an endangered language?

 

Coming soon! Please check the Pilgrims in Segovia Teacher Training courses 2026 at Pilgrims website.

Please check the Pilgrims f2f courses at Pilgrims website.

Tagged  Voices 
  • Lost for Words? How AI is Preserving Endangered Languages
    Nigel Balfour, Spain