On AI, various government agencies have conflicting and confusing strategieshttps://indianexpress.com/article/opinion/columns/artificial-intelligence-india-niti-aayog-6063205/

On AI, various government agencies have conflicting and confusing strategies

Ultimately, India’s AI strategy narrative needs to change from being a reactionary step to “counter the charge” of countries like China, to a proactive one where policies and infrastructure serve as “a beacon of inspiration” to other countries that are further behind.

artificial intelligence, artificial intelligence in india, artificial intelligence research in india, Niti aayog on artificial intelligence, 
While it is clear that India is heading in a direction where both the private and public sectors are unified in their commitment to promote and upscale AI, most of these commitments have been made on paper, in budget speeches, proposals and heavily researched reports

To be a $5-trillion economy by 2025, India needs to build a cohesive national strategy around artificial intelligence (AI). While the government has been vocal about its intention to mainstream AI applications for social good, and ensure that AI research in India keeps pace with global developments, there is little evidence to show that even the basic building blocks to achieve this have been put in place.

Multiple calls taken by various governmental agencies have led to seemingly independent and often confusing strategies, resulting in conflict and a very real danger of ineffective execution. Till date, the Ministry of Electronics and Information Technology (MEITy), the Department of Industrial Policy and Promotion (DIPP), and the Niti Aayog have all released national strategy documents, each of them containing recommendations on governing structures, policy, as well as proposals on creation of new agencies. Surprisingly — and worryingly — there is no mention of how (and if) these structures will co-exist towards the goal of a unified AI strategy for the country.

For instance, the Niti Aayog’s “National Strategy for AI” report allocates a budget of Rs 7,500 crore and recommends setting up Centers for Research Excellence (COREs) in conjunction with academic institutions. It also recommends setting up International Centers for Transformational AI (ICTAIs) in association with leading industry players. It falls short, however, of clearly recommending the governance framework under which this should happen. The DIPP is next, with a budget of Rs 1,200 crore towards setting up the National AI Mission (N-AIM). The N-AIM is supposed to be the nodal agency for all “AI related activities” in India which will also set up their own “centers of excellence” to promote interdisciplinary research, and assess the performance of various AI-based products in India.

The MEITy on the other hand plans to allocate a Rs 400-crore budget for new technology initiatives as part of the Digital India Programme, including working with the Digital India Corporation to set up yet another apex body for AI called the National Center for Artificial Intelligence (NCAI). While details on this are sparse, it has recently emerged that the ministry is at loggerheads with the Niti Aayog in terms of who ought to ultimately spearhead this movement. While the Union finance ministry appears to have weighed in to resolve the tussle, the final policy call on who gets to lead the charge is shrouded in controversy and uncertainty.

Advertising

Moreover, sector-specific AI applications, as diverse as facial recognition and crop classification, are being supervised by different state and central-level ministries with seemingly no consolidation around these national AI strategies. This can lead to fragmented adoption of technology, duplication of effort, and a wasteful use of financial resources.

While it is clear that India is heading in a direction where both the private and public sectors are unified in their commitment to promote and upscale AI, most of these commitments have been made on paper, in budget speeches, proposals and heavily researched reports. In fact, none of the recommendations highlighted earlier have yet been implemented in any useful form — this is in stark contrast to countries like Taiwan, which went from announcing a $36-million project to build a supercomputing platform to boost AI research in June 2018, to launching the National Public Cloud Computing platform, based on the Taiwania 2 supercomputer, in June 2019.

It is important that policy-makers and agencies converge their ideas around the groundwork that has been laid to streamline the effective creation and implementation of the country’s national AI strategy. There is also a need for greater transparency in the timelines and roadmaps associated with these announcements, so that startups, non-governmental organisations and researchers can not only provide their input, but also understand when they can use some of this promised infrastructure if they are to compete at the international level.

Ultimately, India’s AI strategy narrative needs to change from being a reactionary step to “counter the charge” of countries like China, to a proactive one where policies and infrastructure made in the country serve as “a beacon of inspiration” to other countries that are further behind. As the DIPP policy recognises, “people, process and technology” are non-negotiable for AI to proliferate in India, but in the absence of the first two, much will still left to be achieved in the third.

This article first appeared in the print edition on October 11, 2019 under the title ‘Too many cooks’. Viksit Gaur is an entrepreneur based in San Francisco with an engineering background in artificial intelligence and Akriti Gaur is a lawyer working on issues at the intersection of law and technology in India.  Views are personal