AI for All: How India Can Become an Artificial Intelligence Superpower
In 2015, Klaus Schwab, the founder and executive chairman of the World Economic Forum, announced the emergence of artificial intelligence (AI) as the dawn of the Fourth Industrial Revolution. The First Industrial Revolution was driven by water and steam. The second one was powered by electricity, and the third relied on electronics and information technology. While each of them enabled formerly inconceivable efficiencies in production techniques, the Fourth Industrial Revolution promises to pave the way for unprecedented automation technologies.
From waking up to Siri’s news updates to falling asleep to a movie suggested by Netflix’s recommendation engine, the technology underlying the Fourth Industrial Revolution has already penetrated our daily lives. As denizens of the 21st century, it’s now virtually impossible to avoid interaction with gadgets and software that use artificial intelligence, internet of things (IoT) and robotics. Products using these technologies have gained traction across all generations, from the centennials to the baby boomers.
AI is predicted to contribute $15.7 trillion to the global economy by 2030, and it has the potential to add $957 billion to India’s economy by 2035. So it’s worth exploring the possible impact of AI on India’s economy and workforce. While the media often focuses on the potential of such technologies to wreak havoc, this coverage rarely includes nuanced discussions about the potential economic impact of AI, and the policy environment required to reap its benefits.
Is India standing at the brink of a technology-driven “jobless future”?
The most heated debates around the impact of the Fourth Industrial Revolution are related to its consequences for the job market. A commonly cited 2013 research study, “The Future of Employment: How Susceptible are Jobs to Computerisation?” provides empirical evidence of the repercussions of automation by describing a phenomenon called “job polarization.” What this means is that middle-skill jobs (such as those in the manufacturing sector) will slowly disappear, whereas low-skill and high-skill jobs will increase. This is because middle-skill jobs will be easier to automate, as they involve routine tasks – whereas low-skill and high-skill jobs are harder to feed into algorithms, as they involve non-routine tasks that require emotional intelligence. India receives 65% of global IT off-shore work and 40% of global business processing work. Connecting the dots, the study’s authors claim that 69% of the jobs in India’s formal employment sector will be automated by 2030.
While these numbers paint a gloomy picture, one must realise that there are two sides to the same coin. While on the one hand, fears about job destruction are making headlines, on the other hand, labour economists have shown how technological progress engenders net job creation. Economists from MIT and Boston University advocate optimism around the labour market effects of automation. They claim that automation leads to the reallocation of jobs rather than their absolute displacement. While it’s true that the demand for certain types of jobs will shrink as they get automated by machines, production efficiency will simultaneously increase, augmenting the demand for labour to carry out other types of jobs related to it. A case in point would be an AI-enabled software capable of reviewing legal documents during the “discovery” phase of a legal suit. The software might be expected to replace legal clerks and paralegals – but it also would reduce the cost of discovery, thereby increasing the number of legal clerks employed in the U.S. by 1.1% per year between 2000 and 2013.
Interviews with business syndicates and several studies suggest that these optimistic arguments hold true for the Indian context too. Teamlease Services’ Employment Outlook Report 2018 gives a comprehensive overview of trends and forecasts in hiring and job growth across eight cities and 16 sectors in India. Rituparna Chakraborty, the co-founder of Teamlease Services, succinctly summarizes its key finding: “The data implies risk of automation. However, that does not mean it would lead to a job loss necessarily.” Debabrat Mishra, a partner at Deloitte, elaborates with examples of jobs like data entry clerk and cashier: They would be replaced by computerised machines, yet they would result in the creation of new jobs, such as data validation clerk and query handler.
Looking at the bigger picture shows that some type of jobs will take a hit in the short term, but the impact will even out as newer job profiles arise in the coming five years in India. The country’s National Association of Software and Services Companies (NASSCOM) and the Federation of Indian Chambers of Commerce and Industry (FICCI) commissioned a study titled “Future of Jobs in India” that sheds light on this change in the “workforce mix”: By 2022, 9% of the workforce will be employed in jobs that do not yet exist, and 37% will be working in roles that require radically new skillsets. And as more specialised jobs come up, their pay scales will soon start skyrocketing, due to a dearth of the talent required to undertake such roles. Even senior-level employees will be affected by these structural changes in the labour force demands. The case of Cognizant Technology Solutions laying off 200 senior staff as part of their “workforce management” strategy in 2018 bears testimony to this principle.
Feathers in India’s AI Cap
The disruptions and opportunities of AI are likely to accelerate in India, as it has already crossed some major milestones in the Fourth Industrial Revolution. Despite being a late entrant in the field, India has ranked 13th out of 200 countries with respect to the development of AI. According to analysts at GMF, the AI sector in India has grown by $150 million over the past five years (as of 2018), with private investments doubling from $44 million in 2016 to $73 million in 2017.
In order to leapfrog to greater heights of development, the Indian government has already established some state-of-the-art facilities dedicated to the research and application of advanced technologies like AI, and more are in the pipeline. In its working paper “National Strategy for Artificial Intelligence #AIforall,” the National Institution for Transforming India (NITI Aayog) proposed a two-tiered approach for achieving India’s AI aspirations, which involves the establishment of Centres of Research Excellence in AI (COREs) and International Centres for Transformational AI (ICTAIs). COREs will be hubs of knowledge generation for academics. ICTAIs will leverage private sector collaboration with industry practitioners to translate research evidence into social good for all. And in October 2018, the World Economic Forum (WEF) launched the Centre for the Fourth Industrial Revolution in India. WEF will work alongside NITI Aayog, business tycoons, startups and academia on projects pertinent to this sphere, starting with AI, blockchain and drone technology. Meanwhile, AI labs are being set up by Indian telecom companies such as Bharti Airtel and Reliance Jio. And a Centre of Excellence in AI has been launched in Karnataka and Telangana through a public-private partnership between NITI Aayog and NASSCOM, involving technology giants such as IBM, Microsoft, NVIDIA, Intel and Amazon Web Services.
To understand why India is so concerned with harnessing these disruptive technologies, take a look at some statistics which quantify their potential economic impact. According to estimates released by NASSCOM, AI could contribute $957 billion to India’s economy by 2035, with agriculture, education, healthcare and infrastructure as the primary sectors benefiting from it. In 2017, 4% of the country’s GDP came from the leaps made by digital products and services created using AI and IoT. According to predictions from Microsoft India, this number is expected to shoot up to 60% by 2021. Indian IT companies have already established themselves as pioneering technology solution providers to the world, and India could build on this model to provide AI as a service as well.
Making India an AI Superpower: #AIforall
With great computing power comes great responsibility, so it’s no surprise that policymakers are increasingly getting involved in the push toward AI. In his budget speech of 2018-19, India’s finance minister recommended that NITI Aayog craft a National Programme on AI to propel research and development of emerging technologies. And NITI Aayog’s aforementioned working paper provides a detailed preamble to the national policy which will be deliberated upon in the coming years.
The “#AIforall” working paper calls upon India to play to its strengths by investing in the “gold mine” – those sectors of the Indian economy where AI will have maximum social impact. It has identified five such sectors, namely: healthcare, agriculture, education, smart cities and infrastructure, and smart mobility and transportation. The paper recommends that NITI Aayog adopt a three-pronged strategy in pursuit of these objectives. First, it should develop AI pilot projects to establish a proof-of-concept. Second, it should draft a full-fledged national strategy to enrich India’s AI ecosystem. And third, it should partner with leading AI research institutes and organisations to execute these plans. As the name of its discussion paper suggests, NITI Aayog has branded this mission of deploying transformative technologies for social and inclusive growth as #AIforall.
But these efforts will also face key challenges on the road to realizing the full potential of AI. The biggest barrier is the absence of robust, interoperable systems to collect and store relevant data. The second key concern is an under-investment in fundamental research around AI, and in purposeful skilling to build AI expertise. Finally, the Indian regulatory framework continues to adopt a conservative approach to data privacy and security, and regulations on data anonymity remain nebulous.
The paper identifies many areas in which India must make large strides in order to become an AI superpower. In the long term, India must invest in building its competence in core AI research, by broadening its research base beyond the Indian Institute of Science and Indian Institute of Technology. The country’s IT giants, such as Tata Consultancy Services, Wipro and Infosys must then translate outputs from academic research into commercial products, which requires an integrated research ecosystem. In the short term, India has an opportunity to introduce national-level programmes and incentivize corporations to up-skill and re-skill the youth, enabling them to survive the job automation debacle. This will entail recognizing the credentials rewarded by informal training institutes, and formalizing these schools in the National Skill Qualification Framework. In addition, the government must welcome the adoption of AI through market-based initiatives, such as creating a marketplace for stakeholders to share data and develop AI solutions. It must also spur partnerships and collaborations, and fund the creation of open-source annotated datasets. Lastly, India’s regulatory bodies must establish a data protection framework and privacy laws along the lines of the European Union’s General Data Protection Regulation.
For a country in which millions still lack access to basic public services such as health, electricity, education and sanitation, establishing India as a leader on the global AI map may seem like a distant dream. And yet the time is ripe for India to take its place on the AI frontier. The necessary groundwork has already been laid. We cannot afford to miss this opportunity.
Photo courtesy of geralt.