NITI has partnered IBM to develop AI-enabled yield-prediction and real-time advisory to the farmer on productivity, pest-warning, etc, using data gathered from remote-sensing satellites, soil health cards, IMD and various other sources.
Many economies—France and China, most prominently—have either formalised strategies to harness and realise the potential of artificial intelligence (AI) or are already heavily investing in AI. The NITI Aayog’s National Strategy for Artificial Intelligence #AIforAll that outlines focus areas for adoption and possible challenges, therefore, merits close attention from policy-makers if India is to leverage AI for growth, development and greater inclusion.
Indeed, adopting AI means a 15% boost for the gross value added (GVA) for the economy by 2035—NITI estimates that AI could potentially add $957 billion to India’s $6,397-billion dollar GVA projected for that year. The focus areas that NITI has in mind—healthcare, agriculture, education, urban development/ smart cities & infrastructure, and mobility and transportation—hold both great promise and risk for the country. Consequently, the rewards from adopting AI in these areas and the opportunity cost of deferring such adoption will thus be either transformative or ruinous, respectively, for the country.
Advancements in technology over the last couple of decades—computing evolution (cloud, big data, machine learning, etc), falling costs (cheaper data storage), growing digitalisation, among others—have ushered in an “AI Spring”. With access to technology easing for the masses, the gains from AI are any nation’s for the asking. For instance, the NITI paper speaks of how tackling cancer could get much sharper, and easier, if AI were to be harnessed in oncopathology.
While India sees 1 million new cases of cancer every year, it has barely 2,000 pathologists with experience in oncology and less than 500 expert oncopathologists. However, machine learning could let a general pathologist correctly identify cancer, if annotated and curated pathology datasets are made available for algorithms to match patient slides with images of slides from the data repository. NITI is likely to soon launch a programme to develop such a database.
In agriculture, Microsoft, in collaboration with the International Crops Research Institute for the Semi-Arid Tropics (Icrisat), has developed and AI-enabled sowing app that sends advisories to farmers on the best date to sow, soil-test based fertiliser application, manure application, seed treatment, optimal sowing depth, etc. In 2017, 3,000 farmers in Andhra Pradesh (AP) and Karnataka used the app, resulting in a 10-30% increase in kharif yields across crops.
NITI has partnered IBM to develop AI-enabled yield-prediction and real-time advisory to the farmer on productivity, pest-warning, etc, using data gathered from remote-sensing satellites, soil health cards, IMD and various other sources. To tackle school dropout, the AP government has partnered Microsoft—an app powered by the company’s Azure Machine Learning processes data on a student’s gender, demographic details, past and current academic performance, teacher skills, etc, to identify those likely to drop out. This has helped the state identify 19,500 probable dropouts from government schools in Visakhapatnam district alone, for the current academic year.
While the gains from AI abound, the challenges are many, too. To truly harness AI’s transformative potential, India must address its lack of expertise in AI research and application—the demand for AI and machine learning specialists in India could rise by 60% this year alone, and a demand-supply gap of 2 lakh data analytics professionals by 2020 is likely—as also its lack of high-quality data ecosystems across sectors. With data being the spine of AI adoption, there will be privacy and data security concerns that the government must address on a war-footing; though, with the Srikrishna committee’s report, this could soon be addressed.