India’s Digital Imprint on the Globe — Part 2 | Empowering Communities
‘AI for All’: A Peek into India’s Deep Tech Strategy
The India Stack infrastructure provides the scaffolding for innovators to build scalable solutions that empower individuals. This begs a question. Can technology also help us provide solutions that allow us to go beyond an individual and empower entire communities, states and nations?
Niti Aayog’s Arnab Kumar, in a recent summit on Future of Work organized by YourStory, addressed how India aims to use AI and Machine Learning to ensure “AI for All” — a motto that requires us to be aggressive in our ambition and inclusive in our focus.
Empowering communities: A 3-pronged approach
Kumar proposes a 3-pronged approach to develop deep tech for high impact. First: structured and democratised data access. Second: AI-specific computational capability. Third: Research and Development.
Access to structured data is a game changer for building AI solutions, and Niti Aayog’s efforts in the space of Healthcare and Agriculture are telling.
The Imaging Biobank for Cancer is primed to give oncologists with the set of modern tools to curate and annotate tumor imaging, discover and validate quantitative imaging & pathology biomarkers and build clinical-grade decision support systems.
The Indian Agriculture Stack is similarly geared towards achieving interoperable data sharing between agricultural platforms.
The National Data and Analytics Portal (NDAP) is another flagship initiative to be a one-stop platform to utilise, analyse and visualise all published government datasets, overcoming current challenges of lacking standardised or user-centric datasets.
The Three-pronged approach to develop deep-tech for high impact includes: Structured and democratised data access, AI-specific computational capability, and Research & Development.
For computing, Arnab Kumar recommends AI-specific computing, over High Performance Computing (HPC) and the AIRAWAT Cloud Computing platform is Niti Aayog’s pitch to provide this brute computing power. With 100 PetaFlops of computing available to every single researcher, academician, and startup for both medium and long-term use, it will be useful in image recognition, speech recognition, natural language processing for research, development, and creation of a variety of new applications for the support of advancements in the fields of agriculture and healthcare.
The final arrow in the quiver is boosting R&D around AI. With both Basic and Applied research feeding into product development, the target is to soon become adept at offering AI as a service (AIaaS). This research augments the National Knowledge Network.
Enriching lives, one algorithm at a time
With a keen focus on the above 3 aspects of deep tech, Niti Aayog sees multiple application areas being boosted by the use of AI — Precision Farming, Health Diagnostics, Personalised Education, Road safety and Traffic regulation amongst others.
Some of the India specific challenges such as enabling Natural Language processing directly on Indic languages, or using AI apps for postnatal care of newborns and mothers show a lot of promise and can be scaled to other localities and populations.
As an example, India has one of the highest incidences of low-birth-weight babies in the world, and the highest percentage of infants not weighed or inaccurately weighed at birth. It isn’t too difficult to come across birth registries in rural maternity clinics where every child is noted to have the exact same weight — an indication of newborn weight measurement being a mere afterthought. This is courtesy of health centers being under-staffed, or frontline healthcare workers being under-equipped. Inaccurate weight measurement could deprive babies weighing less than 2.5Kgs of critical care and shackle them to a life of prolonged and serious ill-health.
Wadhwani AI’s pioneering work in AI-Powered Anthropometry is an example of inspired humanitarian application of AI, built to solve the above challenge: A smartphone-based anthropometry technology — a computer vision and AI-powered virtual weighing machine — allowing frontline workers to screen for low-birth weight babies in rural homes. Such an application — the first of its kind in India — will provide accurate, tamper-proof, geo-tagged measurements on a smartphone, without additional hardware or data connectivity. The potential for such a cutting-edge technological development to benefit the most vulnerable sections of our society is immense.
Bharat builds responsible AI
Each of these deep-tech driven innovations have a fundamental responsibility to adhere to the call for ethical use of AI.
AI solutions need to shift away from the Black-box models and start focusing on AI explainability, prevent unfair bias in deployment of ML based solutions, create data management guidelines that protect privacy and lastly establish clear accountability in the development of AI solutions and procedures for liability.
Ultimately the pitch is straightforward — India provides a perfect “playground” for enterprises and institutions to develop scalable solutions which can easily be implemented in the rest of the developing and emerging economies.
Given this garage approach to AI, building for India means building for 40% or more of the world — a modern day technological opportunity imbued in the Indian ethos of Vasudhaiva Kutumbakam, if done right.
In Part-1 we covered India Stack and how the products and services leveraging it empower the individual.
Vignesh Venkatachalam is a Tech Policy enthusiast, interested in using Data Science for Policy Making. He is intrigued by developments in the areas of Internet Governance, AI Ethics and System/Organisational Design and their applications in Public Policy.
Like what you are reading?
We write at the intersection of emerging tech, politics, culture and us every fortnight. Subscribe for free : www.bit.ly/IntechDispatch
