The great Indian AI push

MeitY has come up with a multi-year plan to develop an India dataset platform and build indigenous compute infrastructure

The Ministry of Electronics and Information Technology (MeitY)’s push for the creation of dedicated data platforms seeks to shape the future of Artificial Intelligence (AI) in India. This, going forward, can help create a centre-backed data platform, with data from Indic languages, to progress innovation that will drive the next generation of corporate data infrastructure in the coming years.

WHAT IS THE MINISTRY BACKING?

On 13 October 2023, Rajeev Chandrasekhar, the Union Minister of State for Electronics and Information Technology, in an interaction with the media, elaborated on the Centre’s India AI strategy, which was based on reports compiled from seven working committee reports. There are two clear prongs to the strategy—first, a multi-year roadmap to develop an India Datasets platform, and second, the development of an India AI compute platform to indigenously build chips that will power super complex and powerful data operations.

“The India Datasets platform will be one of the largest and most diverse collections of anonymised datasets to train multi-parameter AI models… It will be deployed in real-life use cases that span across agriculture, healthcare, education, fintech, cyber security, governance through digital public infrastructure, and languages through the Bhashini program. We believe AI can have a transformational effect in these areas, and catalyse the startup ecosystem that is already fuelling Digital India,” Chandrasekhar said at the press briefing.

Chandrasekhar also said that a key part of India’s AI strategy will be to create curated datasets as part of the broader platform, which will be managed by an independent National Data Management Office (NDMO) within MeitY. Further, there will be both government as well as private sector datasets within this platform, and only anonymised and non-personal information will be a part of it.

FUTURE OF INDIA’S DATA OPERATIONS

Such a platform is likely to play a crucial role in the future of India’s academia, private sector and public sector data operations. In turn, data centres will begin playing increasingly crucial roles in storing, securing and processing with low latency the extent of indigenous data in India—all of which can potentially become equitable as a digital public good for a wide range of use cases.

“The India datasets platform will be one of the largest and most diverse collections of anonymised datasets to train multi-parameter AI models.”

Why, though, is this important? First, most of the world’s datasets are based in the English language—which represents the only universal language that offers organised data across various industries such as economy, finance, medicine, energy, and others. In due course of time, each industry will feature multiple business intelligence platforms, which will process anonymised corporate data with trillions of parameters to process industry-wise data analytics.

These data analytics operations, in turn, will play increasingly important roles in every sector—including corporate decision-making, automation of sales and ancillary operations, and internal automation of various sub-segments. To do so, the creation of datasets, especially in diverse language subsets, will be key.

The backing of the government to create a public datasets platform, in this regard, could be key to offering a uniform base for corporations in the country to leverage a public data platform in future. This, in turn, could be particularly key for academic institutions to pursue AI research in applied and theoretical divisions. Startups and small enterprises can also leverage such platforms to create the base for AI operations, even without having substantial revenue for investment.

PUBLIC COMPUTE INFRASTRUCTURE

One of the biggest challenges to developing and processing AI models and infrastructure is access to compute power. Today, US tech firm Nvidia has a lengthy waitlist for corporates to get access to their graphics processing unit (GPU) chips—which require multiple billions of dollars to put together and build.

Under such circumstances, having an indigenous public compute infrastructure, akin to India’s supercomputer infrastructure, can help companies and entities with limited resources have the ability to create advanced AI applications.

It is this that can help India leverage AI datasets and compute infrastructure as a corporate public good—enabling development at unprecedented levels. Going forward, this can help resolve one of the biggest challenges of AI development—how expensive compute resources have been. In turn, this can be further built upon by private firms, and in turn, be developed into an industry the way the domestic fintech sector was built by the advent of Unified Payments Infrastructure (UPI).

Author- Vernika Awal

feedbackvnd@cybermedia.co.in