The government, from next month, will start collection and aggregation of datasets available with the ministries, government departments and public and private records to build a database with all non-personal information that can be used for the India AI (artificial intelligence) datasets platform, sources told ET.Once collected, the raw data will be sifted through to make it usable by improving access, quality and utilisation of the information, a senior government official said, requesting not to be named.

“The idea is to provide seamless access to startups, companies, academics and researchers to large swathes of domestic data. Once we have all the data in the same place, it will be easier for us to apply artificial intelligence (AI) and machine learning (ML) to create localised large language models (LLM),” the official said.

The collection and sifting of all the data is likely to take up to six months following which the government will open up the collated database for training of the LLMs, small-language models, ultra-large language models, foundational models, and others, he said.

The ministry of electronics and information technology also plans to allow the trained data to be used by domestic startups and companies to build application programming interfaces, which can interact with the language models built on top of the database and provide specialised services, another official said.

Under the Rs 10,372-crore ‘India AI Mission’, launched in March this year, the government plans to spend Rs 199.55 crore for the India AI datasets platform, while Rs 689.05 crore is planned to be spent on the application development initiative.

The India AI mission is a larger central government initiative under which the government plans to spend Rs 4,563.36 crore to set up computing capacity of up to 10,000 graphic processing units (GPUs), Rs 1,971.37 crore to set up the India AI innovation centre, Rs 882.94 crore for the future skills programme, Rs 1,942.5 crore for the financing of startups under the India AI Mission.Among the seven pillars of the mission, the government also aims to provide streamlined funding support for deep-tech startups in the AI space, set up an India AI innovation centre, and an AI datasets platform.

The government, from next month, will start collection and aggregation of datasets available with the ministries, government departments and public and private records to build a database with all non-personal information that can be used for the India AI (artificial intelligence) datasets platform, sources told ET.Once collected, the raw data will be sifted through to make it usable by improving access, quality and utilisation of the information, a senior government official said, requesting not to be named.

“The idea is to provide seamless access to startups, companies, academics and researchers to large swathes of domestic data. Once we have all the data in the same place, it will be easier for us to apply artificial intelligence (AI) and machine learning (ML) to create localised large language models (LLM),” the official said.

The collection and sifting of all the data is likely to take up to six months following which the government will open up the collated database for training of the LLMs, small-language models, ultra-large language models, foundational models, and others, he said.

The ministry of electronics and information technology also plans to allow the trained data to be used by domestic startups and companies to build application programming interfaces, which can interact with the language models built on top of the database and provide specialised services, another official said.

Under the Rs 10,372-crore ‘India AI Mission’, launched in March this year, the government plans to spend Rs 199.55 crore for the India AI datasets platform, while Rs 689.05 crore is planned to be spent on the application development initiative.

The India AI mission is a larger central government initiative under which the government plans to spend Rs 4,563.36 crore to set up computing capacity of up to 10,000 graphic processing units (GPUs), Rs 1,971.37 crore to set up the India AI innovation centre, Rs 882.94 crore for the future skills programme, Rs 1,942.5 crore for the financing of startups under the India AI Mission.Among the seven pillars of the mission, the government also aims to provide streamlined funding support for deep-tech startups in the AI space, set up an India AI innovation centre, and an AI datasets platform.