Embracing AI is the best card to succeed: Ivana Bartolett…

‘It is important to have systems that are fine-tuned and trained on Indian data representing the cultural norms, languages, dialects, etc’

India is navigating a pivotal era as its “techno-legal” approach to AI policy moves in the right direction, according to Ivana Bartoletti, Wipro’s Global Chief Privacy and AI Governance Officer. Speaking in an exclusive interview with businessline at the India AI Impact Summit, Bartoletti emphasized that instead of fearing AI, it should be embraced as a vital tool for success. As a policy head, what kind of feedback and suggestions are you getting from the global companies here?

The atmosphere here is fantastic. Having attended other global events like the Paris AI Summit last year, I particularly appreciate the openness of this forum. Seeing students and young people involved is vital, as the core theme is the democratisation of technology. The idea that AI must benefit all of society is central here. It is not just about driving economic growth, but ensuring that growth is equitably distributed.

In this context, governance is not merely a compliance exercise or risk management; it is an institutional capability. It is about the power of a business or a country to adopt these tools and govern their use effectively. India’s unique approach to this is both impressive and important. Do you think that India should draw attention from global companies when it comes to regulations because India still has very soft regulation?

There is no single ‘right’ way to regulate. The EU has its framework, and the US uses a mix of state-level privacy laws and governance models. India’s approach is particularly interesting because it leverages existing legislation first. While there are questions about whether the DPDP Act is sufficient to address all potential harms, India is not pre-empting everything with new laws. Instead, it is focusing on ‘full-cycle accountability’—the techno-legal approach.

Unlike the European Parliament, which might pass 300 laws in a year, India’s legislative process is more deliberate. This makes the ‘soft law’ approach, using guidelines for responsible AI and cybersecurity, crucial. These guidelines can eventually evolve into hard law, but starting this way allows for agility. When you combine this regulatory style with India’s proven capacity for massive technological rollouts, like UPI, you get a unique model that the rest of the world should watch closely.” Do you think we are going with a techno-legal way because we don’t have our own AI models yet and later on approach for hard laws?

That is a good point, but the problem is that AI systems don’t just pop up. The challenge is that you are not going to change this overnight. These are issues that go way beyond India. They are challenges that every country at the moment is grappling with, including Europe, where I come from. It is not just large language models (LLMs); there are small language models (SLMs) too. This is very important: how we use these systems, augment them and fine-tune them for use cases.

I believe that there is big potential. Fine-tuning these models to be trained in India is key because I am fully aware that such models are built with data coming from the web. The data here does not come from the Global South. Even these LLMs reflect the culture of the Global North. It is important to have systems that are fine-tuned and trained on Indian data representing the cultural norms, languages, dialects, etc.

So, it’s the combination — use cases for the Indian economy, smaller models and local languages — that can be used by the people. Do you think we are going on the right direction where we are working on foundational models depending on the global models? No, I think this is a combination of things. The good thing is to develop smaller models – not just focus on large models. Open source is also important.

The other good thing is using the models that we have to create applications that can be used locally by companies and businesses in India. So, it is a combination; it is not just one side for India. What could be your suggestions to get into our own foundational models to start with? I think you have a mix. There’s some very pragmatic and good approach to say we’ve got a mix, we have the large companies and they have invested in India.

So, that is important – they are investing in data centres, the entire AI supply chain, they are also producing models that they cater for local India necessities, and it’s important to have these big companies investing here. On the other hand, it is really about looking at models that are trained with data that you feel comfortable with. They should represent the norms of society and the languages.

You can do this with smaller models too. The other thing is developing applications that can really support the growth of the economy. There are several discussions around impact on jobs because of AI, at this Summit. Being a part of a leading IT company, do you think that’s happening across sectors? AI is going to be a threat for people who do not use AI. My fear is that people are using AI in a different way, also by doing what is called automation.

They trust such AI because the answers look structured. That is not necessarily good. You have got to teach people to question. Skilling should involve adopting a critical yet positive approach to artificial intelligence. Secondly, companies have a duty to provide these opportunities, as does the government in influencing public policy at that level. The other thing which is really important is to make sure that we understand what new jobs are coming around us.

This is a completely new field that AI has created. We need professionals in this field, as much as we will need people who are able to translate the technology into public policy. So, there will be jobs that will be created and there will be jobs that will change. It is really important to look at this and make sure that people are aware, not worried. They should be ready to embrace it and be told that embracing is important.

It is the best way to succeed. To remain valuable in the market, one must learn to work with AI, not to fear it.

Summary

This report covers the latest developments in artificial intelligence. The information presented highlights key changes and updates that are relevant to those following this topic.

---