[image: Attachments]7:00 PM (2 hours ago) New Delhi, Aug 12 (INB) The government is leveraging artificial intelligence (AI), large language models and digital language tools to promote mother-tongue-based education and improve access to learning resources in tribal and multilingual areas, Minister of State for Education Dr Sukanta Majumdar said on Wednesday. In a written reply in the Rajya Sabha, Majumdar said the National Education Policy (NEP) 2020 places emphasis on multilingualism and the promotion of Indian languages, with the government undertaking several initiatives for education, preservation and research in these languages. The government has initiated the BharatGen project, a multimodal large language model aimed at developing inclusive AI solutions supporting all 22 scheduled Indian languages, he said. The project is anchored at IIT Bombay and implemented through a consortium involving IIT Kanpur, IIT Madras, IIT Hyderabad, IIIT Hyderabad, IIM Indore and IIT Mandi, according to the minister. Majumdar said AI-based translation initiatives such as ANUVADINI, developed by the All India Council for Technical Education (AICTE), and BHASHINI, under the Digital India Programme, are also helping expand access to Indian-language digital content. BHASHINI has developed AI models for Indian languages in collaboration with more than 70 research institutions and hosts over 360 AI-based language models and more than 22 language services, including automatic speech recognition, machine translation, text-to-speech, optical character recognition and transliteration, he said. The platform also supports tribal languages including Bhili and Santhali, with its language resources comprising 246 million parallel sentence pairs and 3.7 million monolingual text entries, Majumdar said. The minister said the Central Institute of Indian Languages (CIIL), under the Ministry of Education, has developed foundational primers and linguistic datasets in several tribal languages in collaboration with NCERT. A 12-week Santhali language course has also been launched on the SWAYAM platform to promote mother-tongue-based education, he said. Under the Linguistic Data Consortium for Indian Languages (LDC-IL), resources have been developed for languages including Santhali and Mundari. The Santhali resources include a 50-hour speech dataset recorded from 60 speakers, besides parallel text corpora containing 5,200 sentences each in Santhali and Mundari. CIIL has also developed foundational primers in Mundari, Kharia, Ho-Hindi, Kurukh, Korwa, Bhumij, Santhali-Hindi and Malto, while its Bharatavani portal provides Santhali resources including language-learning material, encyclopaedic content, textbooks and dictionaries, the minister said.
