Preserving Precious Tongues: Exploring the Role of AI in Safeguarding Dying Languages

Our world is marked by pervasive heterogeneity, as different people in different parts of the world have their own customs, traditions, languages, and perspectives, all of which are deeply unique and special. Approximately speaking, there are about a staggering 7,106 living languages across our planet. However, not all of these languages exist on a level playing field. Because languages are deeply embedded within the social sphere, social inequalities inform languages too, as some languages are more powerful than others.

Hence, the global dominance of some languages, because of the greater cultural power and relevance they enjoy due to their greater economic relevance, pushes other languages on the verge of extinction. Therefore, languages die out when more broadly used languages push them out, as they are unable to compete with these popular languages. This is because younger generations no longer find these languages useful, and therefore stop learning them.

When languages lose their social relevance, they are also unavailable on digital platforms like Google Translate, which further increases their irrelevance, as people from other languages cannot use these digital tools to understand these languages. Here, the burgeoning realm of AI offers great possibilities for the safeguarding and revival of these languages.

Because AI technologies are able to learn quickly, feeding vast amounts of datasets in these languages to AI algorithms is a viable strategy to save these languages, by generating newer content in these languages which can be used to promote them. For example, AI can generate stories specifically catered to children as a way of teaching these languages to children, thereby ensuring that future generations stay in touch with these languages. For example, platforms like StoryWeaver, a multilingual story platform for children that supports over 300 languages, is a great way to promote language learning in younger demographics. The USP of StoryWeaver is that it displays a story in two languages side by side, thereby also encouraging the skills of translation and mutual comprehension in children.

However, the most important problem that impedes these efforts to revitalize dying languages is the paucity of content to train AI. This results in inaccurate translations generated by AI. For example, while AI-generated translation between languages like English and Spanish has grown by leaps and bounds to generate high-quality results, AI-powered translation between languages like Swahili and Hindi is riddled with several problems.

What we need then is the work of skilled translators to train AI technologies on these dying languages, by providing AI with specific feedback to improve its translations. However, a further complication this might entail is the future loss of employment opportunities for human translators working in these languages. Hence, there is an important need to achieve a delicate equilibrium between AI and human translations to ensure the preservation of these languages, without the loss of relevance of human translators.

Sources

https://acutrans.com/using-ai-to-save-dying-languages/

https://deepgram.com/learn/language-ai-cultural-preservation