Desi AI app challenges ChatGPT, Gemini:Handles 2cr translations daily; CEO says ‘Impossible to predict where AI will be in 5 yrs’

Artificial Intelligence has transformed the way people work, with tasks that once took days or even months now being completed within minutes. Today, AI is being used across almost every sector, making it an integral part of daily life and business. India has also emerged as a major AI player. According to the SIDE 2026 report, the country ranks fourth globally in AI adoption, ahead of the UK, Japan, Germany and France, with only the US, China and Singapore ranked higher. India’s AI market is now estimated at ₹1.2 lakh crore to ₹1.5 lakh crore in 2026. To accelerate AI development, the government is investing over ₹10,300 crore under the IndiaAI Mission to build a nationwide network of supercomputers and Graphics Processing Units (GPUs). The multilingual AI platform Bhashini is one of the flagship initiatives under this mission and is available on both Android and iOS. In this regard, Divya Bhaskar spoke to Amitabh Nag, associated with the Government of India’s Bhashini project, to understand India’s position in the global AI race and where he believes AI will be over the next five years. Govt’s Bhashini platform records 2 crore translations daily across 22 Indian languages Bhashini is a Government of India initiative that uses Artificial Intelligence to bridge language barriers. The platform develops AI models for translation and speech recognition across 22 Indian languages, which are integrated into digital systems and applications. Launched by Prime Minister Narendra Modi at the Digital Summit in Gandhinagar in 2022, the platform became operational for public use by the end of 2023. Since then, Bhashini has processed nearly 8.5 billion translations. It now handles around 20 million (2 crore) translations every day. The platform has issued API keys to nearly 3,000 users, while around 1,000 websites use its translation plug-in. ‘Our focus is removing language barriers, not competing with global AI platforms’ Amitabh Nag said, “Bhashini has built a strong presence in Indian languages, where its AI models are being widely used. The platform’s objective is not to compete with any other AI system but to eliminate language barriers.” According to him, “It does not matter whether people use Bhashini or another platform, as long as technology helps bridge communication gaps. Users naturally choose the platform that performs best, and Bhashini’s usage figures reflect its growing acceptance.” The Bhashini mobile app has been downloaded more than 20 lakh times. Nag said, “A new feature will soon be introduced to make the app even more user-friendly. The platform is already being used extensively across government departments, from gram panchayats to Parliament.” ‘India uses AI for inclusion, making its impact unique’ Comparing India’s AI ecosystem with countries such as the US and China Nag said, “The goals are fundamentally different. While many countries primarily use AI to improve productivity, India is leveraging the technology to promote inclusion.” He said, “This approach has resulted in a much larger and more diverse user base, making the platform’s impact far-reaching across the country.” ‘Bhashini was built with research institutes across India’ Amitabh Nag said, “Most of Bhashini’s AI models were developed in-house with support from 17 leading research institutes across the country. Instead of following a centralised approach, each institute worked on languages spoken in its own region.” He added, “IIT Jammu helped build the AI model for the Dogri language because local expertise and grassroots data collection were essential. This decentralised approach enabled the government to develop accurate AI models for India’s diverse languages.” ‘AI becomes smarter as it learns from more data’ Explaining how AI models are trained Nag said, “They require large amounts of language data. A monolingual corpus helps the system understand a language, while a bilingual corpus enables translation between two languages.” He compared Bhashini to a child who has just begun learning. He said, “Today it has read 100 books. If it reads 1,000 books, it will become smarter. If it reads 10,000 books, it will become even more intelligent. An AI model’s capability depends entirely on the amount and quality of data it receives.” ‘Open-source approach ensures digital inclusion for every Indian language’ Nag said, “Bhashini was created to ensure that no citizen is left behind in India’s digital transformation because of language barriers. The government has made its datasets and AI models open source, allowing industries and developers to build new applications on the platform.” He added, “Bhashini continues to expand into more local and tribal languages. Recently, the platform added Bhili, a tribal language with no standard script and primarily spoken orally. This focus on Indian and regional languages makes Bhashini fundamentally different from global AI platforms such as Google.” ‘Voice AI will help bridge language, digital and literacy gaps’ Amitabh Nag said, “Bhashini’s biggest goal is to ensure that people at the last mile are not left behind in India’s digital transformation. The biggest barriers for such users are language, lack of digital skills and low literacy.” He said, “Voice-based AI is helping overcome all three challenges, making technology easier to access for millions of people. While the platform also has significant business applications, its primary objective is to make digital services accessible to everyone.” ‘Building AI for Indian languages was the biggest challenge’ Nag said, “The project initially faced challenges similar to those encountered by any start-up. Since AI language technology was still new, there was little awareness about its potential and even the need for such a platform had to be established. As the project grew, the team had to plan its scale, infrastructure and manpower.” He added, “The biggest technical hurdle was the lack of language data. AI models rely heavily on large datasets, but Indian languages have limited digital data. Even when combined, Indian language datasets account for only around five per cent of global language data. Collecting, refining and continuously improving this data remains one of the team’s biggest ongoing efforts.” ‘AI leadership depends on the area being measured’ On India’s position in the global AI race Nag said, “There is no simple answer because every country leads in different areas. While some nations are ahead in certain AI capabilities, India has built strengths in specific use cases, particularly those focused on public services and digital inclusion.” He said, “India has recently launched several large language models and AI systems that directly benefit end users.” According to him, “Global comparisons depend entirely on the parameter being evaluated, with different countries leading in different aspects of AI development.” ‘AI is evolving so fast that even five-year predictions are difficult’ Amitabh Nag said, “It is impossible to predict where Artificial Intelligence will be five years from now because the technology is evolving at an unprecedented pace. The Government of India has made significant investments in the AI sector, which have already delivered encouraging results and are expected to create even greater opportunities in the future.” However, Nag stressed that AI is advancing so rapidly that long-term predictions are becoming increasingly unreliable. Instead of focusing on where AI will be five years from now. He said, “The priority should be preparing for tomorrow by making the right decisions today. Prediction does not work in a technology that is changing this fast.” ‘India is using AI to improve public services and citizen welfare’ Amitabh Nag said, “AI was still in its early stages when he joined the project, with even private companies only beginning to adopt the technology. India has consistently embraced new technologies with a focus on benefiting ordinary citizens rather than just improving efficiency.” According to him, “AI has already been integrated into key government institutions, including Panchayats, the Ministry of Rural Development and NITI Aayog, making it an important tool for public service delivery.” While private companies are also using AI to improve their last-mile reach Nag said, “India’s primary objective is to use the technology for people’s welfare and well-being. In contrast, many other countries largely use AI to boost productivity and business efficiency.” ‘AI Kosh serves as a national repository for data, models and applications’ Amitabh Nag said, “The IndiaAI Mission includes a key initiative called AI Kosh, also known as the AI Datasets Platform. ‘Kosh’ means a treasure of data, AI models and applications.” He added, “The platform enables organisations and developers to host and share their datasets, models and AI applications so that others can easily access and build on them. Bhashini plays a significant role in strengthening this national AI ecosystem.” ‘DPDP Act ensures AI development with strong data protection’ Nag said, “The Government of India has introduced the Digital Personal Data Protection (DPDP) Act to safeguard citizens’ personal information. Under the law, only data that can be legally processed is permitted for use, while unauthorised data remains protected.” He said, “Bhashini’s development is aligned with these data protection norms and that the government has provided adequate time for organisations to implement the new legal framework.”