Amitabh Nag leads the Digital India BHASHINI Division as CEO and oversees the National Language Translation Mission, an effort to create a multilingual AI system accessible across India. BHASHINI, or Bhasha Interface for India, ranks among the largest multilingual AI projects globally. It unites startups, academic institutions, research bodies, industry partners, and government agencies to develop local language technologies. Nag brings more than forty years of experience in technology, digital change, business planning, and public innovation. He previously held senior positions at HP and TCS, managed major IT projects for government and companies including the Passport Seva initiative, and earned an engineering degree from the National Institute of Technology, Kurukshetra. In an interview, Nag discussed BHASHINI’s progress, the difficulties of creating domestic language tools, and the potential of language technology to support broader AI inclusion. The project began in 2022 with the goal of removing language barriers. Success requires attention to technology along with content, data, standards, community involvement, education, and adoption. Officials anticipated strong demand in governance and public services. Work started with twenty-two languages but extends further. Roughly seventy research institutes across India collaborate on localized solutions, handling data creation and model review at regional levels. Many Indian languages lack adequate digital records, so teams gathered spoken data directly from people describing topics or images to build text collections. Five core challenges were addressed: speech recognition, text translation, text-to-speech conversion, character recognition, and entity identification. Models now cover twenty-two languages and more, supported by over forty services that enable conversational use. Automatic language detection, voice activity detection, and a glossary of three million context-specific Indian terms have been added. Since the 2023 launch, the platform has handled eight billion inferences and processes about twenty million daily. Usage spans nearly all states and numerous government departments. Applications include automatic transcription in 120,000 panchayats and nearly two million downloads of the BHASHINI app. Farmers in several states receive voice advisories in local languages. Future work focuses on completing transactions through voice alone, with eight hundred aspirational districts already using the tools.
