WORLD NEWS

It took 5000+ years to go from invention of the wheel to the bicycle. Another 70 years to gas-powered cars and 100+ more to the space shuttle. Hundreds died in “discovering” these steps forward. Now imagine the wheel ITSELF being able to “learn” from varied sources – travels, observing birds etc. then “remember” and not repeat its mistakes, and “evolve”. Add ever improving computational skills. This “Smart Wheel” will reach the space shuttle stage in a fraction of the time. This is Generative Artificial Intelligence – Gen AI. It is already here and represents a once in centuries opportunity for India.

GenAI operates through “deep learning” and self-directed knowledge searches in very large and disparate data universes. This learning is used to autonomously solve highly complex problems. Unlike “automation” where humans “delegate” tasks, GenAI creates its own, ever-expanding work sphere. Five properties differentiate GenAI from past technological leaps:

• It can “think” and decide for itself
• It can come up with new ideas
• It can be easily integrated into daily lives of “normal” people interacting in natural languages
• It will continually expand the scope of what it can do
• It will keep accelerating the speed of its evolution as it “learns”.

GenAI’s impact will be ubiquitous. The structure and economics of entire industries will transform, affecting sectors as diverse as Financial Services, Entertainment, Health, Agriculture and even Environmental Protection. For the first time in history, white collar roles will be replaced. Functionally, GenAI’s effect will be seen in three big areas.

Customer service – it can take-over almost the entire primary interaction –being seen with chatbots for starters.

Research, Innovation and Creativity – the ability to pull patterns out of hitherto unconnected datasets and “birth” new ideas will lead to more Eureka moments and reduce the drag of repeated mistakes.

Production and Process Optimisation – creating brand new patterns of production that break the mould of existing practices. The recent cancellation of software contracts to our IT majors is a canary in the coalmine about the imminent existential threat.

The financial impact is enormous – McKinsey estimates that GenAI can add upto $4 trillion annually to the global economy. Likely an underestimate, it is expected to add $1.5 trillion by 2030 to India’s GDP. Naturally, investments are flowing in. The US government invested $1.7 billion into the field as part of its non-defense investment, and another $1.1 billion as part of its non-classified defence budget in 2022. Private investment has been higher. Between 2013-2022, investment grew nearly 20 times. In 2023, investment in GenAI and related areas was about $27 billion with $18 billion coming from Amazon, Microsoft and Google.

India is at a crossroad. There is a path of inaction leading to results paralleling the Industrial Revolution. Then, some countries “invested” in industrial technology and gained a huge advantage over those that didn’t. The result – colonisation of Asia and Africa and the exploitative colonial pattern of trade. Those that do not invest will become the new exploited colonies.

The more attractive path parallels the discovery of oil by Arab countries. Arab countries have been a dominant voice in the global economy based on their control of “crude”, and having a major share of GenAI will have a similar power. Global dominance in GenAI needs large datasets to help “train”, trained manpower for research and product creation, and capital. Serendipitously, India has two – data and people, and “crude” offers a way to solve for capital.

First, India has to grasp the scale of the GenAI opportunity and create the ecosystem to access it. It is revealing that NITI’s strategy on AI didn’t mention Large Language Models (LLMs). LLMs are the foundation of GenAI developments – and not having “Indian” LLMs is creating long-term “nirbharta”. This is just to illustrate the difficulties in strategizing about GenAI in the absence of either private behemoths or a huge military-industrial complex.

So, what are the specific tasks for us?

• Inventory our resources – our huge pool of engineers and scientists that can be made “fit for purpose” for GenAI work; and the gargantuan amounts of data we generate that is a natural training set for GenAI LLMs
• Choose Priorities – India cannot focus on all potential GenAI uses. A smaller set – health, local language interfaces and education, delivery and measurement of public services and environment protection could be initial priority areas
• Go “Upstream” – Instead of building applications, focus our companies and research on creation of relevant LLMs.
• Availability of LLMs will unleash the creative energy on an unimaginably broad scale as the “power” of GenAI is placed in the hands of larger user groups. The upstreaming will also provide a “seawall” of protection against larger western competitors, and naturally lead to the creation of full stack AI solutions.
• Subsidise and secure computing infrastructure and related supplies – The computing resources needed for GenAI are unlikely to be affordable for most corporates. The government has to create a “national” computing infrastructure. Similarly, the government must actively secure international supplier partnerships to ensure we have access to the requisite hardware for building out GenAI ecosystem. For this, we need to proactively switch resources to this within our budgetary processes.
• Make our education “GenAI ready” – establish a separate task-force to immediately introduce changes in the New Education Policy necessary for making our students ready to be GenAI users and builders.
• A “new” PPP model for GenAI – Microsoft et al have invested heavily into GenAI ventures while leaving them independent – so that their innovativeness and quest for the cutting edge is not diluted. The “Nationalised” infrastructure has to be combined with “investments” to support budding private GenAI champions. Returns for the government can be a mix of subsidized use of resultant applications and share of profits.
• South-South cooperation – PM Modi’s strategy on providing vaccines to the world is a template. India should become the “go to” provider of GenAI technology and products for the global south. In addition to the strategic value it brings to our security, it also creates a natural market for our GenAI enabled IT industry.
• Non-paranoid regulation – allow the industry to build up steam. Just like a thoroughbred horse, allow the GenAI ecosystem to gain scale initially through looser regulation. This will allow fast growth, and give regulators time to understand the industry better.

Now to draw inspiration from “crude” to secure capital. Western capital rushed into the Middle East after realizing that everyone needed crude, and only they had it. It will be the same with GenAI. If we harness our human and data “reserves”, the universal need for GenAI products – LLMs, applications will pull capital in.

India was once the centre of global production – textiles, hand-tools. Then industrialization killed our advantage. GenAI offers an opportunity to recapture our place in the global economy, and for longer.

END OF ARTICLE