India to Launch 'IndiaAI Datasets Platform' by 2025 to Boost AI Innovation

IndiaAI Datasets Platform To Go Live By January To Accelerate AI Innovation

Data from central and state governments and the private sector will be a part of IndiaAI Datasets The IndiaAI Datasets Platform, a part of the IndiaAI Mission, will go live by January, National eGovernance Division (NeGD) chief executive Nand Kumarum said The IndiaAI Datasets Platform is aimed at streamlining access to high-quality, non-personal datasets that can facilitate AI innovation Speaking during an event, Kumarum said that the platform will mirror the US-based collaborative and open-source forum HuggingFace, where developers can create, train and deploy their models, according to a report by Economic Times. This unified platform will serve as a repository for a diverse range of datasets, making them easily discoverable and accessible to researchers. Highlighting the use cases of generative AI in the government, Kumarum said it can be used for drafting requests for proposals or schemes if trained on past documents, improving office productivity, and enabling easy comparison of different states' policies. Similarly, he added that AI trained on norms around buildings, roads, irrigation systems and others can help partly automate the verification and approval process. Kumarum said that data from central and state governments and the private sector will be a part of IndiaAI Datasets, adding that private partnerships are a work in progress. This comes after the Ministry of Electronics and Information Technology (MeitY) presented a report last year suggesting that the government introduce a pricing model for sharing data, including non-personal data, under the India Dataset Platform (IDP). Last year, a working group of the government's artificial intelligence (AI) think tank, IndiaAI, recommended sharing of non-personal data. It is pertinent to note that NeGD provides the technology for the implementation of the ministry's egovernance projects. The union cabinet approved the IndiaAI mission in March with an allocation of INR 10,372 Cr over the next five years. The Mission aims to position India as a leader in the global AI landscape and foster technological self reliance and ethical practices. The Mission has six pillars - computing capacity, future skills, an innovation centre, application development initiative, startup financing, and safe and trusted AI. It not only aims to create 10,000 CPU Processing Units under AI computing but also emphasises creating a public-private model for funding and implementation that will further enable setting up data centres and AI services in the country. AI, especially GenAI, has taken the world by storm in the last few years. India is among those countries which are leveraging AI to improve efficiency and productivity. This has spawned the rise of several AI startups in the country, which offer various use cases for the emerging technology. India is currently home to over 100 generative AI startups. These startups have also attracted a lot of investor interest. The Indian AI startups have raised over $600 Mn in funding since 2019. Spearheading this transition are names like SarvamAI and Krutrim, which are focused on building Indic LLMs.

Indian Govt to Launch Open-Source AI Datasets Platform in 2025

What's the news: India will have an open-source forum for hosting datasets called 'IndiaAI Datasets Platform,' by January 2025, said National eGovernance Division (NeGD) Chief Executive Nand Kumarum on October 9, 2024, reported Economic Times. Stating that the platform is one of the pillars of the government's Rs 10,000 crore IndiaAI Mission, Kumarum said the goal is to create a space for developers to create, train, and deploy their own models. "The idea primarily is like HuggingFace - you have models, you have datasets, and you have people coming up and using those datasets and building models. We are trying to do something similar," Kumarum said, as per the report. Data from central and state governments, as well as from the private sector, will be part of the platform. Last month, Additional Secretary to the Ministry of Electronics and Information Technology and CEO of Digital India Corporation, Abhishek Singh, informed that the Indian government, under the IndiaAI Dataset Platform, was setting up Data Management Offices which would be tasked with ensuring that data collected is AI-usable. Additionally, there would be corresponding Data Management Units in every government department, he mentioned. On September 30, the Ministry of Science and Technology inaugurated BharatGen, a government-funded generative AI (GenAI) initiative focused on creating a multimodal Large Language Model (LLM) for Indian languages, indicating the country's growing interest in building its own AI ecosystem. Private companies have also announced LLMs trained in Indic languages. Tech Mahindra's 'Project Indus', initially trained in Hindi and around 37 dialects, claimed that it would cover other languages and dialects in a phased manner. Ola in December 2023 announced Krutrim, a family of LLMs trained in Indian languages, to cater to "1.4 billion India." Krutrim claimed to be similar to ChatGPT in reasoning capabilities and that its base model could write in 10 Indian languages while understanding 20. Ola CEO Bhavish Aggarwal also claimed that the LLM could outperform GPT4 in Indic language performance. However, several users reported inaccurate answers generated by the model.