Mumbai: 

As India works its way through the fast-changing AI world, Niranjan Rajadhyaksha, Noted Economist and Executive Director at Artha Global, offers his insights on the way AI is shaping the Indian economy, labour market, and society.

In an exclusive interview with Eenadu, Rajadhyaksha provides insights on AI and suggests that the country should work towards developing its own AI models. He also explains the opportunities and challenges of AI and his perspectives on leveraging its potential.

Here are excerpts from the interview:

Niranjan Rajadhyaksha: There is no doubt that AI will have a profound impact on all of us. There are three main areas which deserve policy attention — the geo-strategic implications of AI, its impact on the economy, and protecting the privacy of citizens.

India can seek to do frugal innovation as it has in the case of its space programme or pharmaceutical research. The economic aspect has two parts to it — investments in AI capacity right now as well as broader work on assessing what impact it will have on job creation, especially in some areas of traditional strength such as low-end coding. The third issue is privacy. AI still needs a framework of ethical governance to protect citizens of a country from the misuse of their data, and much more work is needed here.

In the case of the first area, there is a robust debate in India on whether we should develop our own foundational AI models or focus on using the LLM models developed elsewhere to build the best use cases. The debate began when the financial resources and computing power needed to build foundational models looked massive for a country such as India, but the development of DeepSeek now tilts the balance towards the argument for building our own foundational AI models.

Niranjan Rajadhyaksha: Prof. Blanchard meant that it would be a positive productivity shock since DeepSeek promises to provide AI capabilities to a wider community through its open-source variants as compared to the proprietary AI algorithms that are dominant right now.

There are many reasons for this, including the fact that it takes time for firms to learn how to use a new technology well, some activities die out because of a new technology, and sometimes firms are conservative in changing their internal production processes. On all this, the evidence on this issue is admittedly scant right now, so we will have to see how AI affects our economic lives.

The bigger question for economists is whether AI will become a general purpose technology that will lift productivity across the economy rather than in one part of it. Earlier examples include the steam engine, electricity and the internet. However, we must also remember that economic research into the impact of general purpose technologies shows that productivity in an economy does not immediately respond to the introduction of such a technology.

Niranjan Rajadhyaksha: That is one important reason why India can consider developing its own AI models and training them on data that is originally in Indian languages, a case of what is sometimes described as sovereign AI. Some Indian tech companies are already working on AI models that will operate well with Indian languages, which should hopefully also build a bridge across several parts of our rich linguistic heritage. There is reason to be optimistic, but we should also be attentive to the potential obstacles in the way of a wider adoption of AI in our country.

Niranjan Rajadhyaksha: Technological change is usually disruptive in the sense that it can make some types of work obsolete but also creates new types of jobs. This is especially true of general purpose technologies that are pervasive. The reality is that the waves of innovation that follow the development of a general purpose technology such as artificial intelligence have an impact on the economy over decades. Think of electricity. It opened the way for the entire consumer goods industry. The Internet has led to everything from social media to online travel bookings.

So, tempting as it is to forecast, it is very difficult to realistically say how a specific type of work will evolve under the impact of artificial intelligence. However, there is some interesting recent research that shows that the skills gap in the same type of work can narrow when everyone has access to artificial intelligence tools. In other words, the difference between the most productive and the least productive worker in a specific team or a specific type of work can shrink. So, in that sense, AI will actually help workers with lower skill levels.

Niranjan Rajadhyaksha: This is perhaps the most important question that Indian policymakers will have to consider in the context of AI. Economists see production as a process that brings together men and machines in some combination. The pessimistic view is that AI is different because it may develop in a way that it could replace human beings in the process of production, not in a specific type of work but more generally.

Economic activity is not just about producing goods but also consuming them. An economy that is completely automated — and with an extreme concentration of wealth — will not be sustained unless most people have incomes to spend. These are some of the more conceptual issues that come to mind, though the world is admittedly nowhere near such an eventuality in the near future.

The optimistic view is that new types of jobs keep getting created. Either way, government policy in the future will need to work out how to resell those who have lost their jobs to AI and build a robust income protection system in case job displacement is very significant. There is also the problem of demand.