New Delhi, September 2(APAC Media): BHASHINI and Kathmandu University have discussed possible cooperation in language AI, with a focus on Nepali and other underrepresented languages.
The discussions took place during an interaction between the Digital India BHASHINI Division (DIBD) and Kathmandu University’s Center for Digital Public Infrastructure and Artificial Intelligence (DPI-AI), followed by a meeting with the university’s Vice Chancellor.
Focus On Nepali And Local Languages
The talks covered language models for Nepali, Newari, Tamang and Gurung. Participants also discussed traditional scripts such as Prachalit Lipi and Ranjana Lipi, which remain less represented in digital language technologies.
A major part of the discussion was the availability of language data. Building reliable AI models for local languages requires large and representative datasets covering different speakers, languages and writing systems. The two sides also discussed translation capabilities and the computing infrastructure required to develop and scale language models.
Voice AI And Digital Applications
The discussions also considered voice-based applications that could make digital services easier to access in local languages. The focus was on connecting language datasets and AI models with practical applications rather than limiting the work to model development.
For Kathmandu University, the engagement brings its research and digital public infrastructure work into discussions around language technology. For BHASHINI, the talks could support wider cooperation on multilingual AI beyond Indian languages.
BHASHINI operates under the Digital India Corporation and MeitY and works on AI-based multilingual digital access. The platform currently supports Indian text and voice languages as well as several international languages.
The meeting comes as governments and academic institutions look at language AI as a way to improve access to digital services for communities that are often underserved by mainstream AI systems.
