“AI is the new electricity”—Andrew Ng.
Andrew Ng,[1] the Director of the Stanford Artificial Intelligence
Laboratory, equates the age of Artificial Intelligence (AI) to the discovery of electricity
for allowing inexhaustible applications into all walks of life. AI might be the single
largest technology revolution of our times, and we are now living in an AI-led era
already. Be it, Amazon Alexa, Apple Siri, or Google Assistant, our smartphones now
are equipped with a personal assistant powered by AI with nearly 90% accuracy and
dependability. From live “chatbots,” to smart suggestions for travel, shopping, or
restaurants, AI has invariably impacted our everyday lives. The buzz around this incredible
technology is beguiling and poised for a paradigm shift in our lives.
Can a Machine Be Made to Think?
Can a Machine Be Made to Think?
That question unanswered in the past has now been answered in the affirmative albeit
in a limited set of circumstances. Machines can now be trained to do tasks of various
complexities with a defined output of measure, known as “machine learning” (ML). This
ML can handle a small quantity of data and provide a defined outcome. Going to the
next level is “deep learning” (DL), which is similar to the network of neurons of
the brain, with stacked layers and connecting in a random pattern. This deep neural
network can handle large quantities of data, learns by itself, and makes intelligent
decisions and performs complex tasks; in other words, it mimics the functions of the
brain. Both machine learning and deep learning are a subset of AI. It simply means
the ability of machines to perform cognitive tasks such as thinking, perception, learning,
language skills, problem solving, and decision making akin to natural human intelligence.
Based on their capacities, they are also called “weak AI” and “strong AI,” which are
essentially machine learning and deep learning, respectively.[2]
The concept of AI is not new, and researcher Alan Turing, the great British code breaker
in the second world war had done considerable work in this field but it was John McCarthy
who coined the term AI in 1955.[3] This field however remained very limited in its scope because the data was very
small. Over a period of time the data has increased exponentially to the current estimate
of 4.4’ Zetabytes’ (ZB) and is likely to rise to 44 ZB or 44 trillion Gigabytes (GB)
by the year 2020. This is called “big data” and data mining is “data science.” The
cost of hard drive per GB has fallen drastically from $50,000/GB in 1980 to 2 cents/GB
today. AI is already in use in many industries and web applications. It is expected
that 70% of enterprises will implement AI by the end of 2019.
AI technology is at an emerging and evolving juncture. India, a growing economy and
the second-most populous nation, has a tremendous opportunity to carve its need-specific
and robust AI ecosystem. With the inauguration of India’s first dedicated AI research
institute[4] in Mumbai, and the NITI Aayog the national think tank bringing out a discussion
paper on AI in the year 2018, the ground work has been laid down in order to gain
an early momentum for India in the global development of AI.[5] Taking this forward, the government of India has announced plans to establish a
national centre for AI and a national program for AI.This year, the NITI Aayog has
drawn up a plan providing Rs. 7,500 crores for creating an institutional framework
for AI in the country.[6] This has been a much needed impetus for India to catch up with her international
peers.
Globally, during the last two years, countries like China, the Unites States, Japan,
the United Kingdom, and France have announced their national policy positions on AI.
It is estimated that by 2030, 26% of the GDP of China and 10% of the GDP of the United
Kingdom will be sourced from AI-related businesses.[5] China is leading the race and has strategized to become a global leader by 2030.
With a robust thrust from the Chinese government to invest and incorporate diverse
AI applications such as face recognition surveillance and AI chip technology, they
have got an early start to lead the way in harnessing A.I. Currently, China harbours
the biggest cluster of AI scientists and has a well-structured program in place. In
this technological race with high geopolitical stakes, both China and the United States
are head and shoulder above the rest.[7] This dominance over the technology is cornered by a handful of tech giants, which
includes Google, Apple, Amazon, Facebook, and Microsoft (GAFAM) of the United States
and Baidu, Alibaba, Tencent, and Xiaomi (BATX) of China. These leviathans together
have a market capitalization of $4.5 trillion ($ 3,438 billion for GAFAM and $ 1,132
billion for BATX) in the AI technology expansion. [8] Despite global attention and expectation, the benefits of this revolutionary technology
will be limited to less than 50% of humanity unless a wide range of societal domains
is included.
An Inflection Point in Indian Strategy
An Inflection Point in Indian Strategy
In what manner and for what use is a’ labour surplus’ developing economy like India
is looking at AI?
AI can be used both for harnessing information and knowledge and in turn improving
expertise in various fields. The NITI Aayog in its vision paper has focused on reforms
in five important sectors or domains. (1) healthcare to improve access, affordability, and quality; (2) agriculture to enhanced farmers’ income, farm productivity, and reduction of wastage (3) education to improve access and quality; (4) smart cities and infrastructure for the urban population; and (5) smart mobility and transportation to ease traffic congestion and improve safety.
Setup on a “public–private partnership” model, the Wadhwani Institute for AI4 has
already worked on three core issues—maternal, infant and child health, in detecting
low birth-weight babies with the virtual weighing machine, on multiple challenges
in tuberculosis control, and lastly in cotton farming to reduce crop losses which
is the third largest agricultural yield in India mostly of small farmers. There are
already more than 1,500 AI startups in India, to cater to local and global needs.[9]
The nations of the developing part of the world share common challenges in sectors
such as health, education, infrastructure, agriculture, and urbanization. With IT
capability, India has the potential to becoming a “global player” to produce scalable
solutions for similar problems. In effect, the solutions developed and applied to
India can benefit 40% of the rest of the world, and this is in fact a tremendous opportunity
for the country to assume its position as a trailblazer for the emerging economies.
The government is rightly exploring the potential to provide a destination for an
“AI garage.” The NITI Aayog estimates adding $1 trillion to the Indian economy by
way of new AI solutions in the next 15 years which will boost India’s annual growth
by 1.3 percentage points by 2035.[5]
AI in Health Care
With data revolution and effective data mining, the application of AI has brought
the health care sector vastly on the way of the transformation. This impact is seen
across the specialties, including radiology, dermatology, neurology, ophthalmology,
oncology, cardiology, genetics, emergency medicine, and so on during the past few
years. The medical image recognition technique by deep learning has produced predictable
accuracy in interpretation, which are at par with those of trained radiologists. Despite
the perceived threat of radiologists losing jobs, the near future likely to witness
the increased collaboration between radiologists and the technology to significantly
enhance clinical benefits in terms of accuracy, cost, speed, and also their reach
to remote areas. Similarly, Google AI has developed “computer vision” technique to
detect diabetic retinopathy and in the diagnosis of malignancy utilizing digital pathology.[10] Robotics powered by AI has been in use for demanding and high-precision surgeries.
The AI in healthcare can help to address issues such as poor connectivity and limited
supply of healthcare professionals in rural areas which at present constitute significant
barriers. This can be achieved through the implementation of cases such as AI-driven
image diagnostics, clinical decision support, personalized treatment, predictive population
risk stratification, early identification of potential pandemics, and tools for patients
to manage their own illnesses.
AI in Plastic Surgery
Currently, the impact on the field of plastic surgery is limited to specific tasks
such as quantification of burn wound size, monitoring of vascular perfusion following
microvascular surgery, and to facilitate accurate diagnosis of craniofacial anomalies
based on computed tomography (CT) images (with 98% specificity and 92% sensitivity).[11]
[12] It has been found useful in orthognathic surgery due to improved diagnostics, therapeutic
planning, computer-assisted appliances, intraoperative navigations, and follow up
of patients.[13] The potential for use in aesthetic surgery, the results of which are largely subjective
is also being explored with predictive tools for patient-perceived beauty. Intraoperative
assessment of symmetry and precise anatomical position can be made out using an optical
“head-mounted” display, and a higher degree of objectivity can be achieved.[11] AI has powered biomechanical neuroprosthesis to help in precise finger and wrist
movements to provide a better grip.
AI technology has loomed over the horizon and it is here to stay. At this inflection
point, it is a remarkable opportunity for all of us to be aware, and adapt to this
change to supplement our technical capacity as well as self-empowerment. Lastly, no
nation can now neglect this tool if it wants to keep pace with others and if it does
so, it will be at its own peril.
