Agility in adapting to technology really important, says Amith Singhee

In a video interview, Amith Singhee talked about Research Lab's role and generative artificial intelligence (GenAI). Edited excerpts

AMITH SINGHEE, director of IBM Research Lab India and chief technology officer (CTO) of IBM India and South Asia, believes all companies have to be “very aware” about technology. In a video interview with Shivani Shinde, he talked about Research Lab’s role and generative artificial intelligence (GenAI). Edited excerpts:

How does IBM Research’s work help IBM?

IBM Research Lab works very closely with all the business units of the company. We have two main goals in research. First is to be the growth engine for IBM. Our organic innovation drives growth in different businesses, and the second is to shape the future of computing.

Our role is like a spotlight for the company to understand what's happening with the state-of-the-art technology and to shape some of that, especially in the computing stack and link that with the businesses and with the product units.

Just because somebody figured out how to read images with a computer doesn't mean that it's ready for consumption in the enterprise system. A lot of research challenges that need to be addressed.

How should one look at technology, especially the speed with which its changing?

It is a significant challenge, but also an opportunity because all this tech offers promise. Any technology adoption is driven by business priorities. Getting into any of these areas — AI, GenAI, quantum computing, Cloud — for a client is an investment and frankly the amount of disruption each of these bring is a significant investment. Second, they need to set up their organisation so that they are very aware of the technology. I don't think it's an option anymore for large businesses to say someone else will figure out the technology, and I'll only do what I do in my business processes. They need to understand the technology because it's becoming a critical driver or enabler of business.

Clients need to be willing to change and relook at their strategy every six months, or one year…agility has become really important in making these choices.

There is a debate on large language models (LLMs) and should India look at investing in it. Comments

It has to be all of these. Our needs are diverse. If you are serving a regional market, you don't need an LLM, you can serve that market using SLMs (small language models). But if you're creating something like the Digital Public Infrastructure-style national AI capability for farms everywhere or migrants who move across the country…then you cannot look at it just vernacular by vernacular issue, you need an at scale DPI style capability. India will benefit from doing all of the above.

People are getting skilled in the latest technology trend. What will be the impact of this?

There is a big benefit to it. If the young talent in the company, or for that matter people across segments, are up to date with the new trends and understand how to implement them, they can definitely partake in the new opportunities that arrive. I think where we can have a soft spot because of this is, if we are only going to deliver services or use somebody else's tech, it's fine, you can learn that, or use it to create value.

But if we want to be a nation that creates the tech that others use, that cannot be done just by keeping up with the trend. You have to define the trend, you have to know the starting points and work end to end and that's where the foundational depth and rigour is very important. We do need to protect, incubate and grow a section of our skills that does the deeper research, and innovation that is new to the world. This is important because technology is going to move faster and faster, and if we don't have that internal competency at the scale that India is to the world then that's the lost opportunity for the nation.

How do you see AI shaping the information technology (IT) services industry?

There's definitely an inflection point in terms of automation that's coming. There is no denying that a number of mundane repetitive tasks that don't have an aspect of creativity, or some kind of thinking involved will get automated. If automation accelerates the business processes, I think businesses will do more. It is disruptive for individuals who now have to think differently about reskilling.

This impacts the IT services industry as they work with people. Many of the employees are going to use these automation tools whether it is for generating code, or analysing invoices. IT services vendors will also look to pass some of that benefit to the clients. At the same time as clients are looking to adopt AI and automation themselves, they are not looking just for the same IT services they used to get. They are looking at services that allow them to use AI in their business – or how do they integrate processes. These are new opportunities opening for IT services companies.

How do you see Watson has fared when compared to some of the other AI players on the global platform?

IBM has been in the AI space for years. While Watson itself was a big mile-marker for us, as a brand, as a set of technologies, some of our efforts were highly successful such as WatsonX assistant, which is an enterprise chatbot and is well recognized as a leader. Another one is Watson knowledge catalogue and others. A lot of the narrative around GenAI has been led from a consumer and creator mindset. Like writing poems, creating images etc…that's fine. If you're somebody who quickly wants to create something using OpenAI or any other service out there that's great. On the enterprise side, there are a number of unique requirements that are not served by these publicly available commercial AI services. They need the ability to customize these things in-house, they need to build the AI on their data in-house and control it end to end rather than somebody else controlling them, and other issues. Their requirements are just not served by any of the Copilots or software bots out there. We have to develop these GenAI solutions that cater to those specific needs.

While it's not always visible in the consumer world, there is a tremendous need in the business side for AI Solutions that caters to them.

What are the top three technologies that IBM Research is working on?

At a high level, the three big things for us. One is foundation model, which includes the area of AI, GenAI but using foundation models and foundational models do not mean just LLMs. There are non-language models such as time series data, where we are working or geospatial data etc.

Another big area of our focus is the multi-Cloud computer. The EkStep Foundation also recently launched a programme called Open Compute Cloud, it’s the same concept and we are also partnering with them.

Third, is quantum computing. IBM looks at it from a full stack approach. In the Indian Lab we are more focused on the software layer, we're not doing too much on the hardware side.

What is your take on quantum computing?

There has been an inflection point in quantum computing. There have been trillions of circuits that have been run on IBM quantum computers and millions of users across the globe have used it. The other side of the story is that it is still a technology very much under development. It is not a mature technology that you can roll out to all your technology departments and start using it. It is where normal computing was perhaps in the middle of the last century where computers were showing up, and businesses didn't really know how they're going to use it etc.

I think we have 250 or so members now of the IBM Quantum Network. Many of those are large enterprises and academics, who are seriously using the best quantum computers to do research work. So definitely there's a demand from that perspective.

In India, we are in a green shoot phase. We have IIT-Madras sign up as a Quantum network member and then we have LTIMindtree too. I think it's at the phase of interest but not yet at the level where businesses would invest into it. With the National Quantum Mission and building competency in the country more understanding around it will happen.