From creating models in different languages to setting up data labs and building applications across sectors, the IndiaAI Mission is picking up speed and scale.
The 7th Floor of Parsvanath Capital Tower in central Delhi can be mistaken for the headquarters of a new-age technology company. Swanky, elegant and state-of-the-art, with no trappings of a conventional government establishment, it houses the corporate office of Karmayogi Bharat, a special purpose vehicle incorporated in January 2022 by the Ministry of Personnel, Public Grievances and Pensions, which directly reports to Prime Minister Narendra Modi.
Karmayogi Bharat runs the iGOT (Integrated Government Online Training) platform — an Indian version of open online course provider Coursera — with a target of upgrading skills of the country’s 30 million Central and state government officials over the next two years. Of the 115 courses, 13 focus on a single area: artificial intelligence (AI). “iGOT Karmayogi has emerged as the world’s largest digital capacity-building platform for civil servants under Mission Karmayogi. With over 74.5 lakh Central and state government officials already registered, the platform has recorded over 2.3 crore course completions in the last 24 months since its inception,” says Rakesh Verma, chief operating officer, Karmayogi Bharat. Over a million employees have completed the general course on AI. iGOT also offers domain-specific courses (mostly prepared by private entities Upgrad, Genpact, Oracle, Microsoft and Wadhwani Institute of Technology & Policy) such as AI in Agriculture, AI in Disaster Management, AI-Powered Grievance Handling Applications, AI-Led Digital Transformation in Healthcare, AI and Computer Vision and AI Use Cases in Governance. “Given that Group C employees constitute nearly 90% of the government workforce, enhancing their skills in emerging technologies like Artificial Intelligence (AI), Internet of Things (IoT), Blockchain, etc., is crucial. These technologies are reshaping governance by enabling data-driven decisions, increasing transparency, and improving service delivery…Various courses on AI and other emerging technologies have been made live on iGOT Karmayogi through our government and private ecosystem partners,” says Verma.
The government is preparing a 30-million AI and data literate workforce not just because AI is the talk of the town. The employees, in fact, are being trained to play a big role when a much bigger India AI Mission announced by the Central government is rolled out at scale in the coming months and years. “We are in sync with the AI Mission,” says Verma.
The Mission
On March 7, 2024, the Union Cabinet sanctioned ₹10,372 crore for the IndiaAI Mission to bolster India’s AI ecosystem over the next five years. Half will be spent on two key pillars of the mission — building computing capacity and indigenous large multimodal models that will be trained on India-specific data. The other pillars are improving access and quality of public datasets to make them AI-ready and developing and scaling up AI solutions. Another is setting up data and AI labs in Tier-II and Tier-III cities to offer foundational-level courses in data and AI. Others include supporting the AI start-up ecosystem from product development to commercialisation, developing indigenous tools and frameworks and making self-assessment checklists to ensure implementation of responsible and safe AI projects. “The AI Mission shows our prime minister’s commitment to democratise technology,” says Ashwini Vaishnaw, Union Minister for Electronics & Information Technology.
The nodal ministry, the Ministry of Electronics and Information Technology (MeitY), says such efforts will make governance smarter and add $967 billion to the Indian economy by 2035. Industry body Nasscom considers the IndiaAI Mission as a key tailwind driving AI adoption in India. A Nasscom-BCG study says India’s AI market is poised to grow at a 25-35% CAGR from $7-9 billion in 2023 to $17-22 billion by 2027.
The mission is perfectly timed in view of AI adoption by public and private sectors. “In the financial sector, AI drives a lot of business decisions, primarily because fintech has a lot of digital data. AI has started percolating in a good way wherever data is available,” says V. Kamakoti, director, IIT-Madras. Kamakoti was the head of the government ‘Task Force on AI for India’s Economic Transformation,’ which gave its report in January 2018. “A lot of AI interventions are coming in healthcare. In the education sector, people are collecting and analysing a lot of data ranging from attendance to performance to give feedback to teachers and parents,” says Kamakoti. AI is also driving customer engagement and public utility services.
India has around 2,000 AI-based start-ups. Dozens of pilots to prove the utility of AI are also under way. Every state and Central ministry and department is looking at AI solutions to improve governance. Established tech companies are also investing in AI and building platforms and solutions to serve clients. Once these acquire scale, a digital revolution, much bigger than what India saw in the financial sector with Jan Dhan-Aadhaar-Mobile, is possible. An AI-literate bureaucracy, which Karmayogi is working on, will be the key to AI-based tools and solutions. But capturing data is just one part of the digital architecture that India is building. The other pieces of the AI ecosystem are equally important.
The Process
Days after Chinese start-up DeepSeek shocked the global tech world by developing an AI model that rivals Google and OpenAI at a fraction of the money spent by others, IT minister Vaishnaw expressed confidence about developing India’s own Large Language Models (LLMs) like OpenAI’s ChatGPT soon. “Within a short time, India has progressed well on establishment of the first major pillar of the country’s AI Mission, Common Compute Facility,” he said.
On August 16, 2024, MeitY invited applications for empanelment of agencies to provide AI computing and Cloud services to make available 10,000 graphic processing units (GPU) for start-ups, researchers, students and academicians under the IndiaAI Mission. MeitY has already received offers for about 20,000 GPUs (as opposed to the 10,000 it sought) to build AI computing capacity. A total of 19 companies — including cloud service providers, managed service providers, MSMEs and data centre service providers — submitted proposals. The list had Jio Platforms and Tata Communications as well. “The process was completed last week. Against a target of 10,000 GPUs, we have empanelled 18,693 GPUs, of which about 15,000 are very high-end — 12,896 H100s, 1,480 H200s, 742 MI 325 and MI 300x GPUs,” Vaishnav said on January 30. “Based on this common compute facility, start-ups which want to develop foundational models will get the opportunity to develop India’s own foundational models (LLMs) and models focused on particular sectors or applications or problems,” said the minister. “We have started the call for proposals, and with this we will start discussions with start-ups, IT companies, academia and researchers to give their proposals for the foundational models. Every proposal we receive by the 15th of every month will be given to a technical body which will evaluate it purely on technical parameters. Then, we will fund those proposals,” he added.
The response in other verticals is also encouraging, says Abhishek Singh, CEO, IndiaAI Mission, & additional secretary, MeitY. “For the application development initiative where we plan to deploy 25 AI applications, we have launched an innovation challenge, in which 14 problem statements across five themes have been published. The areas covered under the innovation challenge (where the winner can get up to ₹1 crore as grant) include healthcare, agriculture, assistive tech for learning disabilities, governance, climate change and disaster management.” The National Institute of Electronics & Information Technology (NIELIT) is setting up data labs in Tier-II cities to train data analysts, data annotators in the next one year, he adds. The first one, a pilot lab, is functional at Karkardooma in Delhi. This lab is built in partnership with Intel. We are working very closely with the industry, says Singh. The ministry is also finalising a scheme for supporting AI start-ups. “Under the India Start-up Pillar, we plan to provide funding support to pre-seed- and growth-stage start-ups. The categories are being finalised.”
Similarly, the government is funding eight projects under the “safe and trusted pillar” to build safety features into AI tools, remove biases and inaccuracies, discover deep fakes and identify misinformation. Karmayogi Bharat is not alone in strengthening the capability of the administrative machinery. Take Wadhwani Foundation, the non-profit organisation founded by Indian-American billionaire Romesh Wadhwani, which is working with the Central government to drive AI initiatives. It has trained over 5,00,000 officials on these emerging technologies over the last two years. Recently, it hired chief AI digital officers from the private sector to help top bureaucrats define the AI roadmap for their ministries and prioritise projects based on AI. “We offered to hire chief AI digital officers and put them in agriculture, education, skilling, labour, electronics & IT and women & child development ministries. They will work with secretary or joint secretary level officers,” says Ajay Kela, president and CEO, Wadhwani Foundation.
The IndiaAI Mission is on. However, massive public computing capacity, AI solutions that can be scaled up at a national level, public datasets that can be of use to start-ups to build innovative AI tools and skill development centres are several months or even years away. Meanwhile, there are dozens of standalone initiatives to develop real-world AI applications that can be scaled up to national level once the AI ecosystem of the mission is in place.
Leading the Way
Shekar Sivasubramanian, CEO of Wadhwani AI (an independent non-profit organisation led by Sunil Wadhwani, Romesh Wadhwani’s brother), says there are two ways to ensure an AI-assisted transformation. “One, you can come up with a grand architecture; two, a bottom-up approach, which is what we do. We implement solutions,” he says.
Wadhwani AI works with state and Central governments to improve services in priority areas — health, agriculture and education. Its clinical decision support system, integrated into eSanjeevani, the national telemedicine platform of the Central government, aids remote doctors by providing AI-driven diagnosis from health data gathered through an automated personalised patient assistance form (PAF). As of May 31, 2024, PAF had facilitated over 120 million consultations. Similarly, its AI system enables media scanning units at the Centre and state level to capture unusual health events reported in media, helping prevent disease outbreaks. This tool, deployed in the Integrated Health Information Platform of the Central government, had extracted more than 42,000 health events as of June 2024; 215 were confirmed as disease outbreaks. That’s not all. In February 2024, Wadhwani AI, with support from Google and in collaboration with Bhashini, EkStep, NIC and Samagra, developed and integrated AI and ML capabilities into the Kisan-eMitra chatbot, an AI chatbot of the agriculture ministry for grievance redressal, to ease the workload of officials and reduce delays in payments to farmers enrolled in the Pradhan Mantri Kisan Samman Nidhi, which provides ₹6,000 annually to over 90 million farmers.
Wadhwani AI has developed another solution, Vaachan Samiksha, an AI tool for assessing oral reading fluency, for the Gujarat government. Designed to help teachers and public education programmes conduct assessments, it supports teachers in tailored remediation efforts and reduces their assessment workload while facilitating the development of individualised learning pathways. The AI model serves 33,000 schools in 33 districts, has on-boarded more than 1,20,000 teachers and conducted over 3.6 million assessments. “We’re also a part of the IndiaAI Mission. We are helping them define all this in pragmatic terms,” he adds.
India has several entities building use cases for the Centre and states. World Economic Forum India’s Centre for the Fourth Industrial Revolution, in collaboration with the Union Ministry of Agriculture and the Telangana government, runs the AI4AI initiative (AI for Agriculture Innovation). Stakeholders include agri-input companies, consumer and food processing firms, finance, insurance and technology players, the start-up ecosystem and farmer cooperatives. The AI4AI framework was tested among 7,000 chilli farmers in the Khammam district of Telangana. Industry and start-up partners used state government data management tools to transform chilli farming using bot advisory services, soil testing technology, AI-based quality testing and a digital platform to connect buyers and sellers. The digital advisory services led to a 21% increase in chilli yield, a 9% dip in pesticide use and a 5% fall in fertiliser use. Quality improvements boosted prices by 8%. The income of farmers doubled. Last year, Telangana expanded the programme to include 50,000 farmers, covering five crops across 10 districts.
Some entities are helping governments in other ways, too. CivicDataLab makes governments look at data, helps them build digital data capacities and aids data-driven decision-making. “We want to improve access and usability of data to help evidence-based decision-making. Thus, we co-create open-source data platforms that make high-value datasets actionable and accessible for all stakeholders, including government agencies, civil societies, academia, businesses and citizens,” says Gaurav Godhwani, co-founder and executive director, CivicDataLab. Assam was the first state to use its services to standardise procurement data. CivicDataLab’s AI model looks at the Assam government’s public procurement, geospatial, damages and demographic data to give insights into which districts and sub-districts are more vulnerable to floods. The start-up is rolling this out in Himachal Pradesh, Odisha, Bihar and UP on a pilot basis.The new-age ventures are also using AI to transform judiciary. Adalat AI specialises in AI and LLM-driven solutions that address case backlogs and delays. Sarvam’s tool makes legal drafting, research and document review processes smooth. Almost every ministry has some service that uses AI. In February 2023, the Supreme Court adopted AI for translation of documents. The finance ministry’s ₹100 crore Project Insight leverages data mining and analytics and uses AI and ML to find trends in bank accounts. The Ministry of Earth Sciences uses AI and ML to improve weather, climate and ocean forecasts. The agriculture ministry has partnered with IBM to develop an AI crop yield prediction model. Almost every state government uses AI for multiple purposes. Assam has an AI-based attendance system in government schools; Maharashtra operates an AI-powered chatbot to provide information about public services; Uttar Pradesh monitors inmates of its 70 prisons through AI-enabled video analytics platform Jarvis; Tamil Nadu ensures safety of women and children in public transport buses of Chennai with an AI-enabled panic button-CCTV surveillance project; and Gujarat has an AI-based photo-identification software for understanding population dynamics and social system of Asiatic lions.
The use of AI is also becoming the new normal in the private sector. Gopichand Kattragadda, founder and CEO of Myelin Foundry, an AI company with a vision of transforming human experiences and industry outcomes, says we are entering an era where AI is a tool just like the Internet. “This will bring about a humongous productivity change,” he says.
Katragadda, the first president of UK-based Institution of Engineering and Technology, says an AI-literate workforce means more efficiency and readiness to take on a product or service that is globally relevant. Kattragadda is a member of the Nasscom governing council for the Centre of Excellence for Data Science and AI. “AI is the new IT at an exponentially scaled-up opportunity level,” he says.
The Future
The biggest challenge for India will remain its ability to offer computing infrastructure. Another is sourcing reliable and compatible data in digital format. The revenue potential of many use cases will also remain a question. However, there is no doubt that AI-led transformation has been accepted by the country as a whole. “A lot of things are going to happen, which will have a good impact on society at large. That is where ‘AI for All’ is going to play a very important role,” says Kamakoti.
Once the individual pilots go national, the problem of scale will be solved. That will be the tipping point for government, academia, industry and the start-up ecosystem to move to the next level and create the kind of change fintech brought about in the financial ecosystem, or even bigger.
“Depending on the sector, we are 12-24 months from seeing the first green shoots. Then we will start getting millions, crores of people (as beneficiaries of AI solutions). Once you reach that level, you start making a societal change. I believe the end of 2025 or 2026 is the time. When you already have scale, a new player who makes a delta change has an enormous opportunity. That is when you will see a lot of start-ups coming up,” says Wadhwani AI’s Sivasubramanian. For a transformation of this scale, a wait of a couple of years will be totally worth it.
Fonte on-line:
