PM Modi presents AI language platform Bhashini: What it is
Artificial intelligence (AI) tools are taking over the world. From businesses to governments, everyone is busy exploring how AI can improve their activities. In 2022, the Indian government launched 'Bhashini,' an AI-based solution to bridge the language barrier in the country. Now, Prime Minister Narendra Modi has presented Bhashini at the Shanghai Cooperation Organization (SCO) virtual summit. Let's see what Bhashini is.
Why does this story matter?
India is a country of diverse languages. According to the 2011 Census, the country is home to 121 major languages (spoken by at least 10,000 people). In the digital age, most of these languages do not find a place on the web. Therefore, it is essential to bridge the language barrier to make a fully-functional digital economy. That is Bhashini's purpose.
Bhashini is part of the National Language Translation Mission
Bhashini was launched as part of the National Language Translation Mission (NLTM). Through Bhashini, the government aims to develop a National Public Digital Platform for local languages. The government's aim is to deliver digital content in native languages instead of just English or Hindi. Bhashini is a project of the Ministry of Electronics and Information Technology (MeitY).
AI and NLP tools are available through Bhashini
Through the Bhashini platform, the government wants to make AI and natural language processing (NLP) tools available to Indian MSMEs, start-ups, and developers. Bhashini has a data repository and a model repository. The Universal Language Contribution API (ULCA) is also part of Bhashini. It is a standard API and a scalable platform for Indian language datasets and models.
ULCA is the largest repository of Indian languages
ULCA is the latest repository of Indian languages. Bhashini captures data and model contributions through ULCA. It will help researchers develop AI technologies, including machine translation, automatic speech recognition (ASR), text-to-speech (TTS), and optical character recognition (OCR) in different Indian languages. ULCA presently has 466 translation models and 80 ASR models, among others. It also has benchmark datasets.
Bhashini aims to create a multistakeholder ecosystem
Bhashini aims to create a multistakeholder ecosystem that includes governments, academia and research groups, start-ups, language missions, industry, data collection companies, publishers, and individuals. This ecosystem will develop a repository of data, training, benchmark datasets, open models, and tools.
'Bhasha Daan' is a language crowdsourcing project
'Bhasha Daan' is another part of the Bhashini project. The government is crowdsourcing language inputs for multiple Indian languages through Bhasha Daan. People can contribute to Bhasha Daan in four different ways: Suno India, Bolo India, Likho India, and Dekho India. Users can either type what they hear and speak or type what they see. They can also validate other users' contributions.
What are the uses of Bhashini?
The AI models developed as part of Bhashini would be used for document translations in various domains, including the judicial and education sectors, among others. People will be able to access government documents and communications in their native languages. Language barriers will stop Indians from being full-fledged digital citizens. Bhashini is an ambitious project that aims to transform India's digital sphere.
India would share Bhashini with other countries: PM Modi
"India supports proposals for reforms and modernization within the SCO," said PM Modi at the SCO summit. He added that India would be "happy to share" Bhashini with everyone to "remove language barriers within the SCO."
