There is no doubt that the advent of new Tech allows those lagging in earlier Tech the opportunity to enjoy a level playing field, even leapfrog the incumbents. This happened in India with mobiles and then with e-commerce. The same will undoubtedly happen with AI as well.
Any Tech endeavours to bring personalised and real-time solutions to its users. For a large and diverse country like India, AI can legitimately claim to make that happen better than any other Tech. It can overcome the constraints so far in terms of cost and infrastructure.
But again, I focus in this article on the challenges. Some of them will be caused by AI. Some others are the significant challenges in achieving the AI benefits that we all hope to see.
Unemployment due to GenAI
The biggest concern regarding the advent of AI is the loss of jobs it’s going to cause. To be fair, it has been seen over centuries that technology changes the way we work and live, which causes the nature of jobs to change, resulting in the loss of jobs due to automation.
But then this new way of life creates new jobs that didn’t exist earlier. People adapt to the new jobs, and in general, the new jobs created are more than the jobs lost.
Is it different this time? The impact of AI seems to leave no facet of life untouched. Industry, social media, and even the arts are all facing the onslaught of AI. The rate of change has been faster than ever, leaving no time for people to prepare for the new jobs that may be created.
In recent times, India benefited a lot due to the creation of IT jobs building Technology. The hope seems to be that India will play a similar role in building AI Technology as well. Ironically, these Tech jobs themselves will be at risk, as more and more software development activities will get automated with AI.
IT is one sector where the gap between the experts and the non-experts can be very huge (100x to 1000x!). The experts will use AI technology to create much better outputs – they will participate in building the technology not just for India but for the world. Even the non-experts will be able to cover their gap in expertise to a significant extent if they are creative and think critically while using AI/GenAI.
Even if their %age is small, given the numbers in India, there can be remarkable successes with AI-enabled apps and services benefiting a much larger number of users than before. Those using AI technology without creativity or critical thinking will find their abilities actually diminishing, and will eventually lose their jobs to AI.
An important fact to remember is that in India, many sectors such as healthcare and education require last mile human touch (think nurses or teachers) to provide real solutions to the requirements. AI cannot do these jobs.
So a lot of jobs may not be creating the technology but bringing the technology to the end-users – people or businesses. And if AI increases the number of technology users significantly from the current levels, then indeed more jobs could be created than lost to AI, through the last mile solution, deployment, personalisation/customisation for these new users.
These jobs would be more domain-based and hence Techies would have to learn the domain (such as healthcare). Or the domain workers would have to learn how to use AI/GenAI. In fact, GenAI will also be the technology to skill / re-skill them faster and at the scale required.
Can we not avoid job losses since we all know there is so much more to do? The risk is that the poor users of AI will produce poor quality and sometimes even harmful output! And as we have seen with mobiles, Tech usage grows exponentially. So, time will tell soon enough how dramatic the impact of AI will be on jobs in India.
Challenges for India in realising benefits of AI / GenAI
Whatever the rate of adoption of AI / GenAI globally, there could be reasons for the growth of AI in India to be slower as compared to other countries. This means slower realisation of the benefits to India, and also losing ground to other countries in this paradigm shift in technology.
The fact is that all the LLMs that drive today’s GenAI are foreign, mostly American, not Indian. So, in terms of bringing personalised and real-time solutions to Indians, they will not work that well in India.
We all know the problem of biases that can happen with AI. We expect data to be objective, but we have learnt that data from the past can have biases. And if the data and the LLM’s based on that are not even Indian, we can certainly expect problems.
We need LLMs for Indian languages. However, the big prerequisite for that is having proprietary access to data. Because, even with Indian languages, if the data is public the Tech giants will be able to build the LLMs better.
Kritrim.ai, with its proprietary data from Ola, certainly seems well-placed to do that. Taking an open-source approach, Sarvam.ai is creating the full stack for GenAI with its Indian language LLM OpenHathi.
The one big opportunity for India is the Digital Public Infrastructure and Digital Public Goods. The huge data being created by the India Stack (UPI, OCEN, ONDC etc.) can certainly be used for an open-sourced Indic LLM, if the data privacy issues are taken care of well.
Of course, challenges remain in terms of the cost of the supervised training (the manual data annotation, labelling, cleansing) that is required for creating the LLM, but this can be considered an opportunity as well. The challenge of the cost and even availability of GPUs themselves may be solved by new architectures or by competition in due course of time, but till then we will have to figure out our way. The huge energy consumption required by AI / GenAI could be another big challenge for us in India.
The solution may be to build smarter AI / GenAI by combining LLMs with our “domain” knowledge of Indian people and businesses. This means a judicious balance of rules and patterns as I mentioned in my previous article, that will bring down the volume of data required and hence the energy and other resources required.
Beyond AI/GenAI for India, there is a bigger opportunity in AI/GenAI from India for the world. But for that, we may require more original research, more patents, more talent, and more investments, if we need to rise beyond the labour-intensive jobs such as training data preparation.
In conclusion
Challenges are opportunities for entrepreneurs and we at Pentathlon are confident that the Indian entrepreneurs will overcome these challenges to grab the huge opportunity in building AI in India. They will bring the benefits of India to the world, very much in line with our thesis.
So, what can possibly democratise AI and GenAI so that Indian companies – big and small – can hope to play a leadership role in the AI story? We see a path and that will be the topic of my next article.
