Bengaluru Tech Summit 2021: AI for Evolution & Transformation in India

Bengaluru Tech Summit 2021: AI for Evolution & Transformation in India

Artificial intelligence is emerging rigorously in India. AI technology has penetrated almost all industries, from ecommerce to BFSI and from Manufacturing to Agriculture. Data Science and Deep Learning are more employed to solve complex business challenges. AI is frequently utilised across several B2B, B2C, and C2C channels. To build AI programs that evolves generation after generation without human input. The program replicated decades of AI research in a matter of days, and its designers think that it could discover new approaches to AI one day.

AI Solutions:  For Voice Recognition and Language translation in India

India is residence to more than a billion people who converse numbers of different languages and accents. Approximately 75% of the population can read, write, and speak in their native language. Communication difficulties present challenges for many in accessing online services, including those services and schemes proposed by the Indian government for the masses. Aware of this, the Indian government started the National Language Translation Mission announced in the Union Budget 2021-22 and intended to develop next-generation, voice-based government apps and websites available in all Indian languages. This move has fully underlined the potential of AI to solve many challenging problems. It creates excellent scope for cognitive AI, speech recognition, and language translation services in a multilingual market like India.

Language services have been roughly for a while, and we feel some of them in everyday life. Translator apps use end-to-end Deep Neural Network (DNN) models to translate over 90 languages and accents. Then there’s Cortana, Alexa, and Siri – all of which are based on speech recognition technology enabling speech-based services deployment across many industries.


VIVEK RAGHAVAN, Chief Product Manager and Biometric Architects, UIDAI, shared his experience in Bengaluru Tech Summit 2021 - I think that the focus, in the area of the judiciary at this time, is to try and figure out that how you can make things more efficient. "I mean, justice delayed is justice denied". And we want to improve both the access and the improvements to reduce the kind of delays involved and have some of these processes, which take a long time and simplify them. And sometimes, they involve AI. And sometimes, they don't involve AI. I had the opportunity to work in developing, and we ended up building something that translates orders and judgments into several different Indian languages. The court system is using it to translate this thing. And it's not just being used in India. The Supreme Court of India uses and actually, Bangladesh also uses some of these technologies, which we have developing open-source related to Indian languages.

I think the focus is on how the systems are more efficient and more effective. Quite astonishing with it, you know, especially talking about a fundamental shift and how India is using solutions to its complexity and scale, lead, and pioneer some of the work, especially when you talk about localization of technology and regional languages, the digitalization of content, like legal documents, and the entire judiciary content, I think that's brilliant. And especially when you kind of layer it with technologies that can be reused, that are open, it makes it all more exciting because the exponential adoption of this would be mind-altering.

How do we blend ethical thinking with AI technological problem solving thinking?


S SADAGOPAN, Former Director, IIITB, INDIA, explained this question at Bengaluru Tech Summit 2021 in this way. We need to sit across different disciplines. As we know the techies, the social scientists need to understand some of the broader issues because we are not trained in social science. And social science should also start with faith in technology. They should not start with the feeling that you know technology is terrible, and then they won't talk. If we come together, I will take only one example when we created the aeroplanes. We could fly, and everyone knows there is a significant risk associated with flying right; rays can crash. But what we did over some time, to current decades, year after year, safety has improved so much for all practical purposes, you know, planes, we kind of we get into the plane and then send a message that I meet you at this time, assuming the time landing. So I think we will go to that stage. But it will take time; it is not going to happen in one day. I'm sure many safety issues were not necessarily our or known by the engineers, but we have to come to terms with that. So sit together, be open-minded and learn. And together, we will win. 

I think blending Arts and Science, liberal arts, and social sciences with technology is imperative as we tackle some of these more morally questionable implementations of technology and the uncontrolled proliferation of technology and use of data, etc.

What kind of AI thinking is emerging in terms of ethical balancing & considerations?


K ANANTH KRISHNAN, Executive VP and Chief Technology Officer TCS, answered this issue at Bengaluru Tech Summit 2021 in this way, the question is now coming more and more frequently, and I'm sure a lot of the work by the Hollywood people on the dystopian future has a lot to do with it. But I think it's important and I agree that opens a view that we as computer scientists should not try to do this alone at all. And we must collaborate with all the other sciences, social sciences, for sure.

I'll give one example that TCS is CO-Innovation Network, reached out to the National Law School of India University in Bangalore. We have done several interactions with them to understand the perception of law school in this context. So it is multidisciplinary.

Many corporations, especially business leaders, even at the CIO level, CTO level and definitely at the board and the business leadership level, are still not getting their head around the fact that many AI solutions are inherently probabilistic. So, you know, we've been trained to instinctively say that a computer will always say one plus one is always two. Whereas, an AI algorithm will give you a probabilistic answer with an accurate estimate and a consistency estimate and say that within these bounds, yes, this is what the likely forecast will be. So, that part is coming more and more to say how accurate your system is? How consistent is your system, and then there is an available answer to be given? But we are getting questions on the robustness of data privacy, especially fairness in its defined sense. And there are all kinds of confusion on the fairness issue because, inherently, human beings are not fair. And culturally internationally, you ask what is fair and let's say in Australia or Japan or India; context may not be seen to be fair in Germany or the US and so on. And we can even contractually say that okay, and then it's up to you to figure out whether it was good or bad. We are taking a very engineering view with explainability, a little more engineering view on accuracy, forecast, and consistency, part of the AI answers. We need to be very interdisciplinary and work with all parts of society to arrive at a good answer.