top of page
Writer's pictureAnna Mae Yu Lamentillo

Inclusive AI can preserve endangered languages


Nearly half of the world’s 7,164 languages are at risk, and experts predict that 95 percent could vanish by the century’s end. This crisis extends beyond words; each language embodies cultural knowledge, unique views, and the essence of community identity. The rapid progress in artificial intelligence (AI) offers an unprecedented chance to preserve these languages if AI development becomes inclusive and accessible to all communities.


However, funding discrepancies between affluent and developing countries hamper this progress. In 2022, the United States led with $679 billion in research and development, followed by China and Japan. By contrast, the Philippines and other developing nations operate with limited resources, making significant AI investment challenging.


The need to make AI research more accessible and inclusive is crucial, especially to support low-resource languages and marginalized communities that stand on the edges of digital access.


NightOwlGPT is working to address this imbalance by using AI to preserve endangered languages and bridge the digital divide. Incubated under LSE Generate and accelerated under NVIDIA Inception, the platform offers real-time translation, cultural insights, and interactive language learning.


Its initial pilot in the Philippines focuses on Tagalog, Cebuano, and Ilokano, but its global mission reaches broader. NightOwlGPT exemplifies how AI can support linguistic diversity and empower communities to actively engage in the digital era on their terms.


Low-resource languages face unique challenges in natural language processing (NLP) compared to widely spoken languages like English and Mandarin. AI language models excel with abundant data but struggle with languages like Tagalog or Cebuano, which have intricate structures and cultural contexts.


Models trained solely on high-resource languages often miss the nuances, leading to mistranslations that dilute meaning. NightOwlGPT tackles this by creating NLP models tailored to low-resource languages, ensuring translations retain cultural and linguistic integrity. In doing so, it not only preserves language but also reinforces a sense of identity among its speakers.


The digital divide in developing economies further isolates communities with endangered languages, as these areas often lack AI resources and technology initiatives.

By focusing on accessibility, NightOwlGPT aims to bridge this gap, making AI affordable and adaptable. The platform is designed for those who might otherwise lack access, with features that prioritize social equity and respect for diverse cultural landscapes.


For NightOwlGPT, AI is more than technology — it’s a bridge to digital inclusivity and empowerment for all, especially those historically sidelined in the digital world.

Addressing the imbalance in AI funding is essential for inclusive technological progress. Wealthy nations lead in AI development, yet the cultural and linguistic needs of lower-resource regions often remain overlooked.


This lack of inclusivity deepens inequalities and risks the loss of entire cultural and linguistic heritages. By prioritizing research that values linguistic diversity, AI can evolve into a tool for empowerment, as NightOwlGPT’s mission demonstrates.


As NightOwlGPT expands, aiming to include over 100 endangered languages by 2029, it champions a future where AI accommodates diverse cultural needs rather than forcing communities to conform to dominant tech paradigms.


Ultimately, true AI progress is not solely about innovation but about fostering an inclusive digital environment that celebrates every language and culture.

bottom of page