Skilling and Education in AI

In two significant areas, one is in agriculture, which is the highest employer, biggest employer anywhere. It’s also one of the least productive of sectors that we have anywhere. And that productivity gap in agriculture, if we can narrow it even by a small percentage, you will move the needle by a significant amount. And AI can do that. Just think about much of agricultural output in the global south comes from smallholder farmers who lose 40 % to 50 % of their crop because of pests. Now, if a farmer can identify what the pest is and use a homemade remedy that is given to them in their own language and using local ingredients, if I can move that 40 % down to 30 % or 20%, suddenly a huge swing in the farmer’s income.

So there is no question from a human perspective, if my income is going to, if my crop loss is going to, go up by 10 or 20%, you know, I will adopt it. So. So that’s the first thing, which is purpose. Now, in addition to agriculture, small businesses. I don’t really need a whole bunch of employees if I can essentially harness AI to do market research, to do analysis, and almost be an employee. And I can be a one -person shop and employ and really build a business. Now, beyond that, there are several other areas of application, which, you know, we’ve done the analysis to kind of see, you know, where are some of the biggest opportunities.

So there’s agriculture, small business. After that comes education and skill building. Another very powerful use of AI. And a fourth area is health care. Now, for each of these areas, there is an element of a major chasm that the humans need to cross. And that chasm doesn’t have to do with technology. It doesn’t have to do with how big the pipe is. It doesn’t have to do with whether I have access to, you know, any of the devices. It doesn’t have to do with the various elements of the digital public infrastructure. In fact, India is one of the shining examples. of the distribution system, the rails having been laid. But the key chasm, the big jump that we need to make is across a trust gap, which is in addition to digital infrastructure, in addition to other forms of infrastructure that includes talent and data and compute, there is a trust infrastructure that needs to be built.

Because from a human perspective, I will use a piece of technology if I can trust it. Now, there are many reasons why people are, on the one hand, very excited about AI, as is very evident over here, and at the same time, there is a lingering concern. There is a lingering concern because I don’t quite understand what’s inside that black box. I don’t quite understand how the hiring algorithms work. Why did I get rejected from this job? Why did I get that diagnosis? From a healthcare system. What is the language system telling me? Is something being lost in translation? Can I trust an image that has just been sent to me on social media? So the issues of trust are a very important set of questions.

And then the data that I’m submitting into the system, simply by interacting with AI, I’m submitting data and providing input. I’m actually acting as labor for the AI industry. What’s happening to the data? Who’s using it? Where does it go? Can it be used against me? Is it going to be used in my favor? So the whole question of trust is going to be an enormously important part that we need to consider. So first, purpose. Second is creating a trust infrastructure. And the third is recognizing that no matter what we say, no matter what rhetoric we put on our screens, no matter how many alliterative slogans we have in our meetings, AI is going to be a force for inequality.

There are many reasons why AI is going to create an unequal playing field, not the least of which being the fact that the algorithms are feeding on data. Data is simply a reflection of the past, and as we know, the past is not a terribly equal place. So that algorithm is going to essentially act as a mirror to our past, and maybe part of the risk is that the inequalities of the past get reinforced into the future. There are inequalities in terms of who has access to better tools. Now, even with open source and people being able to, you know, vibe code themselves, there’s an element of democratization, but there could be very different levels of access across a society.

So the usage context itself could be unequal. There could be inequalities when you go into different parts of the world, when you go to different parts of the country. So geographically, there is likely to be inequality. There’s inequality in terms of who’s providing you AI. So today, much of the frontier AI models are coming from two places, United States and China. And much of China’s AI infrastructure is built on top of a foundation from the United States. Much of the foundation of the United States, the leaders of the companies that are producing it, they’re all over here, really small. So it’s a tiny industry that’s providing us the foundation from which we are building the rest of the system.

And then one last really major source of potential inequality has to do with the resources that AI is absorbing, primarily energy, water, space, and even kind of our environmental resources, enormously important. Now, none of this means that we should stop the train. But we need to understand the human impact. that AI is going to have, both positive and negative, as we move forward and put the relevant policy systems in place, the relevant trust -building systems in place. Otherwise, we might be not only wasting a demographic dividend that India has got, but a trust dividend that India has got. One critical and really important aspect of an ecosystem like India is that it’s a very trusting society, very trusting in terms of digital.

It’s a very trusting infrastructure. Trust levels in India is in the 70 % range, whereas in the United States is in the 25 % to 30 % range. That’s a huge platform to build on. And it’s going to be really important for us to follow through with that trust that users, our potential consumers are giving us, and for the policy and the technology sector to be able to make sure that that trust dividend is not wasted. So with that… But I’m going to sit down, and I look forward to learning from my colleagues on the panel about how do we make AI more purposeful and not just powerful. Thank you.

Thank you, Professor, for the insightful remarks. Very exciting, and at the same time, you know, you raise some concerns around inequality. Let me first go to Arunji. You know, as we said, like the demographic dividend that India holds, we will be adding a million plus to workforce every year. How do we make sure that they’re skilled, they’re ready for what the market is asking, the skills are continuously shifting, as CEO of NSDC with the mandate of skilling the population? How do you look at this? How is, you know, do you see AI as a threat, as an enabler? How are you approaching this?

So, good morning. AI is an opportunity and an enabler. So let me begin with a few words about NSDC itself. So this is a national platform institution under the Ministry of Scale Development. And we work through two arms, 36 sector scale councils and close to around 400 training partners. So these are the two arms through which we have been working in the scaling space for the last two decades. With AI coming in, of course, it’s an opportunity, as I say. But primarily in four areas we have started work. One is, of course, AI and how career trajectories are getting shaped. So we require some kind of guidance, direction, et cetera. So work on that front. Second is creating scaling programs for AI, AI scaling programs.

Third is how does AI itself affect the entire value chain of scaling when it comes to, say, training, assessments, counseling and the other. areas. And the last is since we do large scale program management, how do we use AI to evaluate or monitor outcomes? These are the four primary areas we are working on. Obviously I will just pick for each of these areas in brief. The first one, setting the agenda or setting the direction with respect to careers, NSDC and the sector skill council, specifically the IT sector skill council, we have brought about certain reports, how jobs get shaped by AI, the new jobs and how the existing jobs get changed, etc. Within that is career counselling.

Once you know that this is the way a certain job would get transformed or a new job would come, a lot of career counselling is required for students. So how do we create AI enabled career counselling tools, models, etc. So that’s one area of big work. work. Coming to AI skilling programs, clearly there are three trials. The first is of course where we talk about AI for all skilling, which is more like AI awareness and AI usage. So we have this skilling for SOAR program under which we work with schools etc. The second is where we talk about how does skilling affect practitioners or people in the workplace. And this is where our sector skills councils are busy putting together how do we make the current programs, how do we bring in AI modules in it.

Of course to begin with how AI affects job roles to start with and then translating that into how the new programs would look like. The third area is AI for engineers where we skill engineers and this is where we work with engineering colleges. We have something called the Future Skills Centers and we work with close to right now around 10 ,000 students and around 50 ,000 students. And we work with 10 ,000 students and around 50 ,000 students. So we have a lot of different things going on. We have a lot of different things going on. We have a lot of different things going on. We have a lot of different things going on. We have a lot of different things going on.

We have a lot of different things going on. We have a lot of different things going is to create close to around 22 companies work with us, including Microsoft, Google, and Amazon, and Schneider’s, and Siemens, et cetera. And we create these kind of skilling centers within engineering colleges. The good thing about it is that this is part of the credit -based system. So students can pick up over the four years they are doing the engineering every year, every semester they pick up a course, and you string together a course, then you have a kind of a program for, say, an AI architect or something like that. So we look at the entire skilling program. The third is, as I said, AI is changing the way we skill, you know, the way we train, the way we assess.

Early days, again, pilots on, how do you use AI as a training assistant, you know, to our trainers, you know? So what do we do, how do we work with that? Similarly, assessment is a big area. Please see, many of our training involves vocational training, which means hands -on training. So we use AI for hands -on training. Hands -on training requires a lot of… piloting etc so towards that we are working can we say for example just giving an example say what’s a good weld or what’s a bad weld if the AI is trained on that then it can help the current assessor in actually you know actually providing a better assessment and also augmenting the number of assessors we are currently having the last piece is we have our skill platform for large scale program management today it is called SID and we are now bringing elements AI into it so that how do we how do we monitor outcomes better so big challenges in a country like India is monitoring outcomes they are facing how to use AI on that area also so

very interesting and exciting to see what you have brought to the table Neenaji if I can move to you Anandji spoke about the skilling programs but certification standards are very key and how do you do that in an environment where skills are you know the courses are becoming outdated in months the requirements are shifting from your vantage point like a lot of content being created a lot of initiatives all around how do we define qualified professional in AI? Is there a plan for certification or standard setting? How should we think about it?

Thank you so much. Thank you for inviting me. I’m a former executive member of NCBT. One minute about NCBT. NCBT is a regulatory body NCBT is a regulatory body under the Ministry of Skill Development. So something on AI since we’re sitting in an AI conference. Around two and a half years back we came up with a skilling framework for AI. And the framework actually talks about three layers of skilling for AI. It talks about skilling for all. It talks skilling for few and skilling for many and skilling for few. all of the initiative we started working as part of the SWOT initiative that was mentioned by Arun also. Now what does that mean? Like all of us know how to use payment gateways or payment UPI etc.

Can we actually use AI in a similar way? So our thought was can we take AI to everyone every nook and corner a radiowala or a plumber or somebody who is a beautician etc. So what did we do on that? And I’m going to take a minute before I come to a certification question. We actually have tried creating a small nano -credential for something which a beautician can use and how can she use AI for giving a better service to the customers. We’ve actually created a virtual try -on for a tailor. How can a tailor use a virtual try -on concept for actually taking it to, you know, telling a person which design or what kind of a color suits a person.

We actually created basic courses, of course, on AI, which also have been done and they were launched sometime in July. We’ve got around two lakhs plus people who registered on those courses. But idea was, how can we take it to everyone? So how can simple things like, how can a plumber find out if there’s a fault in the pipe? So can AI be used there? So one of the points which I think Professor talked about was, how do we take it to masses? How can AI make an impact in our lives or everybody’s life, like internet is doing or anybody else is doing? So that’s what we’ve tried doing as part of some of the courses that we already are in a state to launch.

In fact, some of the courses have been launched. We all been talking about in this conference that AI is going to replace coders. So there’s going to be lots of jobs, etc. But still, how can actually you use AI for helping to launch? We actually are stalling. we’ve actually demonstrated how AI can help in coding also. How can it help me to learn coding? How can it help me to test a particular program? So AI doesn’t stop at just being AI and taking away jobs. I think we have to groom and possibly diffuse, I mean, the word which has been said, the concept of AI which is happening. Now let’s look at certification in the courses.

Very wonderful question he asked. Things are changing almost every day. I think the carpenter’s role is going to change. In fact, we have from Furniture and Putting, the Sixth Grade Council, a small little model which says, how can I design a particular furniture better if I have AI? Knowing the wood, amount of wood, and the space for which I have to design the furniture, can the carpenter actually use AI to design the furniture better? That’s the way it’s going to make an impact. Now how can I embed this course in a course which I’m teaching a carpenter? That’s the impact it’s going to make. And that’s what we’re trying to do. So a wonderful question from that point of view.

So what we’ve done is we came up with the concept of stackable micro -credentials or stackable nano -credentials. Which means based on the changes which are happening, you could actually stack the small, small modules together and make a skill that is required that needs to be done. And in these skills, you could also have something like employability skill, which could be design thinking or others. And it could also be an AI module which can be embedded, which actually tells you how AI can be embedded in a particular course. For example, our ITIs already have a small seven and a half hours module which have been embedded as basic concepts of AI which are now being taught to every ITI student.

Now what we want to see is how can we actually do our late machines or other machines, how can they be done in a better way with AI. So that’s why we are trying to do it. Now certification is now, the way we are trying to do is small, small certificates. You know, which a person earns, can I. actually lead to credit and the total of that actually can also take you to a larger credit and that’s how it’s been planned as part of something called National Credit Framework that we actually came up with from NCVT and the Ministry of Education. This was launched around two years back. So that’s how it’s been planned. I hope that answers your question.

To Rakesh, historically, you know, whenever general purpose huge technological change comes in, it ends up increasing divide for some time till people evolve and you know, you learn new skills and get off the curve. Now in India’s vantage point, we spoke about the starting point we have. In your view, what needs to be done differently? We heard a few initiatives in motion to make sure that we cross this transition and manage this transition very carefully and aptly.

Thank you so much. It is a very, very pertinent question. And I truly believe that in the past, we have been able to do this. And I truly believe that in the past, we have been able to do this. And I truly believe that in the past, we have been able to do this. era, India was at a disadvantage. It was very difficult when we had internet coming in. India was not as uniquely placed as it is placed today. We have ubiquitous connectivity. We have low cost of connectivity. We have a huge amount of internet penetration. We have applications which are general purpose applications used by a billion people like UPI and others. So today our starting point is very, very different, not only for our users, but also for those who are making these applications.

I think journey started about 10 years back when India realized that we have to make applications for our own people and not just rely on the world to make applications for us and take them forward. So that’s where we are today. Hence, the opportunity for us is immense, given that we are a billion people with access to low cost connectivity, reliable connectivity, already using these applications for our financial transactions and other use cases. And especially for all the programs that we just heard, NSGC and others are today making. So I’ll just talk about three things that I thought I’ll bring to your notice is, I think the point that we should move from digital literacy to work literacy.

What that typically, the point that I’m wanting to make here is that, just a minute, it’s, my phone is misbehaving. So the idea here is how can we remove friction to learning? And friction to learning, typically, the point that I wanted to make is how can it be anytime, anywhere, any media, any duration. And more and more, we are seeing that our population, our people are used to one minute, two minute consumable content. Giving a one hour lecture or one hour content may be difficult now. People really need to consume it in two minutes. If they like, they might go ahead with the two hours. content. So are we creating the right content for our users or are we just trying to reshape the content we have a lecture somewhere and we give a video on a platform and say consume it.

So this whole content strategy has to work if the skill is to be really imparted for people in a consumable manner where the as I said friction to usage is least and obviously if we dovetail the programs that sir was just talking about into those 600 labs that digital AI mission is going to set up across India then somebody who’s interested and you hook on by this small meaningful content on an Instagram can the person really find it interesting and get to somewhere where they can really see the benefit of it and we heard Nandan talking about it is that we should lead from the first principles it is not only giving black boxes to people and saying work with it if you really want India to progress people should understand also a lot of people should understand what goes into this.

Black box so that they can start innovating around so I think that’s one which is remove friction. the second is I think we are all talking about agents and we are talking about physical AI it is not going to be easy for any worker including trust me for me if tomorrow I am told that my secretary is an agent and not a physical person although the productivity of that agent may be much better now imagine we are getting into a workforce where we are talking of lights out factories it’s a reality in China where the factories are totally autonomous how do you get your workforce to work with this physical AI you are working and besides you there is a robotic arm working doing half the work how do you work in such environments where a lot of work is being done by physical AI it will take a lot of mindset shift it will take a lot of role shifts and it is not only telling them what AI is how their roles are changing what is now expected of them once that is clear then only you will be able to train them better so therefore how do we move towards this journey and in our mind be very optimistic of the fact that it is here and it will impact us.

We have to be ready. There is enough and more for a country like ours to become the torchbearers for the world of how to engage 1 billion people on this endeavor. We will create data which will train apps but for that we will have to take on the mantle of creating our own apps which are purposefully made for India, made in our languages, made for our specific use. Thank you.

is going to add to the broader AI ecosystem that we need?

So when I look at the work that we’ve been doing across board and across product areas, and speaking to some of the announcements we’ve made this week, what we’re looking at is how do we bring the full stack of AI into India, right from the foundational level. It’s creating a secure, resilient infrastructure. How do we bring that computational power that India needs in India and not having to rely on that power, that compute in other markets? And that’s why we started out with the build of the data center, the AI data center in Vizag. Adding to that is how do we then ensure connectivity with the rest of the world? That’s where we’ve looked at the subsea cable investments that we’re going to do, which is going to connect Vizag to, to the, to, uh, the U .S.

directly, you know, circumventing through the southern hemisphere. And then as you go up the stack, like you’re talking about, is how do we build out applications and solutions that are actually delivering value to the last mile citizen on the ground, whether it’s in agriculture or health. But every time we look at it, we are looking at how do we kind of, you know, complete the circle. So if you look at it in the education space and the work we’ve been doing with Charn Singh University, is how do you have AI not just at the skilling level, but how do you bring in AI at the learning level and the administration level so we can create a more effective and efficient way of actually delivering AI to the students.

And we’re addressing every part of that chain of learning. How then do you connect it to the workforce, correct? So as you look at the professional certification. And it’s important that these loops start to close because that’s when you actually see impact. That’s key. Similarly, as we talked about in agriculture healthcare, correct? How, in agriculture, you go from seed to market. How do you give information to a farmer to understand the weather pattern so he’s better able to know when to sow, when to harvest, information on the market, information on financial support. So the whole aspect that we are thinking through is how do we kind of help connect the dots? How do we come in and provide our support and our technology to ISVs to create these solutions to connect the dots?

Because that last mile connectivity is actually going to determine the success of this technology. Thank you.

I think we’re quite at time, but I want to take just a minute for a last question to all the panelists together. If you had to identify one decisive action which India should take looking at 2030, something that we can be proud of, what action should India take to make AI as an equalizer? Let me start with you, Professor. five second response is rapid fire.

Five second response, I think the one action that we need to take is improve the trust infrastructure and make sure that the human at the other side of the AI understand what’s inside the black box, at least to the extent that it makes them feel comfortable accepting the output and the decision that the black box is offering.

I think in the next three years I think every Indian should have access to an AI assistant whether it’s a farmer, a student or anything. We have the platforms, we have the DPIs in place, we have the language thing in place, we also have the SIDs, the skills platform in place. It’s time we put it all together and create every Indian having one assistant working with them.

I think what I would love is that if ethics and value is also part of every AI course that’s taught, at least in India, we will possibly create a different kind of AI creators in that field. That would be my opinion. Thank you.

I would believe that it is access to affordable compute which will be important for India to succeed.

I think I’m going back to my first point is on the flywheel. I think a lot of the investments are coming into the compute side. How do we look at bringing in investments and creating economic models for the diffusion of AI and that’s going to be important.

Trust, ethics, compute, human at the core of everything. Thank you so much for the exciting discussion. Thank you. Can I have a round of applause for the panelists and our moderator please.