 Okay. Let's get the slides. Welcome while we get the slides. Good morning. Yeah. How's the day been so far? Are you guys coffee tea, all well tally, caffeinated? I am. So, great. Before we kickstart this, right? A little bit about the genesis of this topic. So, I think yesterday Katrina's talk, you must have attended. She mentioned that as AIBM designers, we are constantly experimenting to figure out new ways of looking at AI. And this is another thought experiment, right? So, we realized that AI was sitting in people's labs and in policy makers' rooms, and all of a sudden, we see JNAI just bursting the bubble. And everybody's like, my God, what just happened? And as this whole AI took a surge the last few months, we weren't prepared, right? We weren't prepared. A lot of questions, a lot of conversation, a lot of dialogue is happening between AI experts, designers, social scientists. Everybody's figuring out those very pesky questions, hard questions about what role will it do? Is it AI for commercial purposes or AI for social good? So, we are exploring a completely different dimension here of how can AI be built to build an equitable world in the future? So, AI for social good. How can we do that? That's an exploration we're making. And it's a very discursive approach. You might be wondering what the word discursive means. So, discursive means to have a discourse. And discourse is basically having a dialogue, a provocative conversation using design. So, taking a discourse, taking a discursive approach, a provocative approach to challenge status quo, challenge norms about AI, that's the approach we're taking today. Okay. So, am I audible? So, as Ruchi has set up the stage now, we all know that good design can solve problems, right? Good design provides good solutions. But what if the design helps us improving some of the societal problems with, you know, prompting some of the positive ideological changes, right? I'll repeat, prompting positive ideological changes. If you do that, then what would be the benefit of that? That is what we are going to dive into. So, basically, this talk is an effort to introduce all of you to the amalgamation of three things, right? Provocative design thinking with the intelligence of AI and using the design thinking approach. So, that is what we are going to delve into. Yeah. So, again, we will be talking about a lot of concepts here. Keep your notebooks ready. Okay. This is by no means an exhaustive session. These are very deep topics. What we will do is we'll share some concepts. I would encourage all of you to go deep, read about it, and make your own intersections as well, and then we can have a chat offline. Okay. So, coming to the structure of this 15, 20-minute talk, we'll talk about foundation. So, we'll just take a quick review of what is discursive, which we just explained, what it is, what's AI, we all know what's design thinking. Then we'll look into the intersection of these three. As I said, an amalgamation of these three. How can we use all of these three together? We'll not just leave you with the theoretical part of it. We will also take you deep into a narrative where we'll look into a real-world scenario of how it can be used, and then we'll also talk about the execution part of it. So, that's going to be the basic structure. Okay. So, talking about the foundation. As Ruchi just explained, discursive design is all about something that focuses on creating design, which provokes critical thinking. As she said, that challenges your status quo, it challenges your public beliefs. It actually provokes the dialogue and discussions amongst yourself. So, something that triggers a discussion and something that triggers an idea of change. And all of this can be done through some of these components. It takes a while. It's a lag, right? Let's go there. So, value of changing screen rights, as we said, a couple of things. This is different from commercial design. We are commercial designers. Largely, I myself am as well. We build apps to create business goals, business outcomes. I think let's just park that for now. Let's build something to provoke new conversations, to actually answer very hard questions, questions around financial inclusion, questions around building an inclusive credit scoring mechanism, questions around mental health stigma. Let's answer those very hard questions. And of course, we'll show you in the end how those questions can be woven in to your commercial design as well. So, just bear with us. Yeah. So, talking about the components, this design has to be provocative, right? It should be challenging your conventional wisdom. It has to be narrative. It has to have some storytelling behind it. It has to initiate a dialogue. It has to have a critical reflection, which means that it should initiate that think critically kind of a mindset. It has to have a context. So, you have to understand the cultural, social and ethical context of the problem of the situation. And then it has to have materiality. So, basically, physically or digitally, this design has to exist because that's how it's going to pass on the message, right? And it could be anything again, like just open up for imagination, right? So, it could be an object, a digital object. It could actually be a physical asset. It could be a solution. It could be anything. It's up for your imagination. Yeah. Yeah. And there's a very good example that, you know, we really want to bring it up here. It's a very hard-hitting example. So, this is a project that is done by three students of National Taiwan Arts University. What you see on the right is not a chocolate-flavored popsicle, right? It's not chocolate, by the way. So, what you see here is polluted water popsicle, okay? I'll first play this video for you. Yeah. Did the students create this design because they wanted to get awareness amongst the citizens, amongst the government authorities that what is the importance of having clean water resources? So, it's a very good example of what we are trying to explain, which is discursive design. Yeah. And imagine what they did, right? They made this, so these are all popsicles from all Taiwanese rivers. They collected water samples. They froze them in a popsicle mold. They packaged, they had packaging, right? So, they can actually experience how dirty the water is. Very discursive, okay? It's not, there's no digital solution here. There is no mobile app here. Thank God, right? But there is a solution which actually just shifts the way you think about this, right? Yeah. Yeah. And again, we spoke about discursive. The second concept is AI. We're all talking about AI in so many sessions. I thought I'll take the opportunity of defining that a little bit for you because I'm a person of first principles. I need to know what this really means. So, when you think of AI, you think of two dimensions, okay? The first dimension is how the systems think like humans. They mimic to act like humans. A lot of work is happening in this space. Cognitive scientists, psychologists are running multiple experiments. All of us are in the last two years, how brain works. You can look at the book titles on the bookshelves. So many books because everybody to kind of have an AI which works, we have to understand ourselves first. How do we work, right? So, that's one part. How do they act like humans? So, this whole Alan Turing's Turing test, you would have heard of that. So, a test, he wrote that if AI is able to talk in normal English, reason and explain, is able to tell you and read your data and find patterns for you, is able to see things for you, is able to move around, then that's a perfect AI. We haven't cleared the Turing test, yes. So, Suleiman from CEO Deep Mind says in five years, guys note, five years, we will clear the Turing test in AI. So, that's the human side, right? But on the other side, systems are think rationally. So, they're logical agents. They solve your problems logically. Now, you will realize it is two very different and friction and what different poles apart point of views. So, there is a human here which is very unpredictable like ourselves and there's a rational mind. When you understand AI, you have to understand these two dimensions. And in your products and services when you're building, which dimension has to be more predominant? When do you behave like a human? Like an emotional coach? Like an emotional bot? But when do you behave like a rational decision making tool? That's a decision we have to make as designers. And talking about the spectrum of problem spaces, talking about design thinking. Now, we all know what is design thinking. So, we are not going to explain that here. But, you know, this is the spectrum of problems. Somehow, you know, we think that we are right now here. This is standard design thinking that we are following. Focused on designing the products, the services, the platforms. But if you see the problem spectrum is, you know, going towards the high complexity, right? There are questions being asked about organizational change, behavior change, sustainable future, etc. And then there are higher problems or high complex problems which talk about complex climate change, driving sustainable actions, financial inclusion, a very, very important point which we are going to cover in our narrative. So, these are some of the high complexity problems. And we really think that design thinking needs to evolve going ahead. Yeah. And Don Norman's book, right? Designing a Better World talks about that. The designers need to have a different toolkit. Now, the larger problems they are solving, they're no more just designing web applications or mobile applications. We're actually solving these wicked problems. And you know what? AI could be your friend to help you solve that. Yeah. Okay. So, we spoke about these three concepts in isolation, but they are some parallels. They are places where they intersect. They are commonalities. They are overlaps, which you should know about. You should stand there. So, again, design thinking we all know, right? So, how can a discursive lens be applied to design thinking? Now, when you're doing design thinking, if you're opening up provocations, it can give you different frames. A different eyes to look at the world. Not just one version, but multiple version. Because it gives you different frames and actually challenges your assumptions and all the stereotypes we have in our heads, it actually helps you create multiple scenarios. There is no one future, by the way. There are multiple futures depending on what your context is, what your vantage point is. So, it helps you create those multiple scenarios. And while you build those scenarios, you can handle tackle sensitive issues. So, that is where the discursive design and design thinking can intersect and build on each other. And now, how can AI augment, right? So, what role will AI have in this? Now, the beauty of AI is humans can do everything, but not at scale. AI gives you scale. So, let's use that to our advantage. So, today, if I want to empathize its scale, today when I conduct research, I pick a small sample called research, pick up a sample of 50, go deep in their lives. But how do I know larger trends? What are the emotional journeys users are going through? AI can absolutely get you the data sets to kind of make those inferences. That's one. Second, it can help you generate new ideas. You can actually give prompts and kind of experiment much faster with that. And last, it can help the prototype much more faster and simulate the future you want to build. Simulate the future which will actually answer those pesky hard questions nobody's answering today. And then, how can AI be leveraged for discourses? Oh my god, this is a godsend, right? When we look at discursive design, we struggle to do a scale. Imagine that popsicle? That was an AI, I think, in the world. All the rivers I can see. How much mud, how much dust we have, right? It's a physical object. How come you can get AI enabled? So it can give you broader sentiments. It can learn what people are saying, how are they feeling. It can actually model big topics which can actually be woven into your AI models. Today, AI models are trained with data from the past. With this, we are creating new knowledge for the AI's to learn about human value, human connection, which is called the thick data to AI. And this is an interesting one, right? So this is one of the person I'm very inspired by, Rafiq Anadol. Hello. He's a live media artist. Okay? He doesn't paint with brushes. He paints the data. So if you look at his website, he actually collects public data. He actually builds them into a narrative around the disappearing wonders of nature. This is a coral reef. You all know coral reef is a disappearing, right? So he collected all that data and he's showing how slowly it's disappearing in life. It's living data and he does that through massive visualizations on buildings, facades, like massive screens to really hit you hard. How are these patterns? How is data changing on a daily basis? So that's a lot of theory, right? It's a lot of concepts that you design thinking, AI, discussive, but what can all of us do? So we have a narrative. I was hoping to build five narratives. That was ambitious, right? We have three minutes, so let's go next. Yeah. And this narrative actually talks about the financial frontier. So people in rural areas, they find it very difficult to access loans and there are various reasons about it. We'll directly delve into some of those reasons. So, you know, there are lack of collateral, so they do not really are able to produce what are the collateral required to get the loans. There are informal moneylanders around around the rural areas, you know, who lead the cycle of debt for these villagers or rural people. There are lack of financial literacies, so people really find it difficult to understand the terms and conditions and they fall into problems because of that. Then there are a lot of policy and regulatory barriers also, which do not let the financial services expand in some of the rural areas. Just one second. So we're designing our financial tools for people like you and me. Okay? I know there's a lot of focus and inclusion and a lot of startups working in that space, but these issues still prevail. Right? People do struggle to make decisions. People do fall in the debt of this whole cycle of debt. So basically, those local moneylanders give you 20 percent interest rate and you spend your entire life just paying back one lakh rupees to that moneylender. That's how extreme it goes. Right? Right. And then there are limitations of awareness of government schemes. The current government has actually brought up a lot of schemes for the rural areas, but then people are not aware about those things. So that is there. And then we have loan defaults. Now various reasons for the loan defaults as well. So there is financial hardships. So there are a lot of unexpected financial difficulties that people get into inadequate savings. People don't realize the importance of savings and they don't do that. And then the misuse of funds. So the loan fund is actually used for some other purpose than you know what is really intended for. Yeah. A lot of problems, right? Designers have a lot of work to do. Trust me. A lot of work to be handled here. Right? You know, somebody asked me, hey, Dharana is doing the job. That's not any time soon, guys. So much of mess we have to clean up. Right? So when we did that, we did a scan of the environment. Okay, what are the emerging signals in this space? And what's the on-ground challenges? They was clear. There was a friction between the signals of technology and what people are facing. From a social perspective, the signal says there is increasing financial inclusion. There are media articles everywhere about that. But you know what? People are still struggling with financial illiteracy. There is a there is a friction there. There are signals about AI based credit scoring mechanisms. How fair they are, how equitable they are, nobody knows. But challenges around illiteracy is still existing. People are still getting interest at a higher rate. They still struggle to get financial support when they're building new businesses and alive for themselves. Again, all these passes, I won't do it down, but the idea is to analyze the macro view from a social lens, technical lens, economic lens, political lens, and environmental lens and look for those friction areas. And the ecosystem is fairly complex, as you can see. Right? In the middle, we have our lenders and the end users, but imagine the rounds of complexity around it. And the friction, right? Information asymmetry, illegal practices, their loan sharks sitting there, giving these unfair things to our farmers and our folks in the villages. And this is an interesting one, right? This is a quote I read from somewhere that a loan is not a financial liability. It's a ticking time bomb. Right? That's how hard these problems are. These are wicked problems. I want to establish the problem space here, so we understand how hard this is. And there's a lot more details. This is just a summary of what we found. But what can we do? Designers, right? We imagine. We build worlds which are equitable. So how about I imagine a world like a sphere of trust? You can imagine there's a night's fair sitting here in the center, which actually helps overcome these issues. It actually builds relationship. It connects bonds. And when I gave this prompt to my dear AI friend, I say, AI, imagine there is a world like this. What would it look like? This is what AI gave to me. That they are actually like a thought partner. It is helping me think through what is the world I want to build. And then I thought, okay, I build this world. What are the different elements and attributes we have to go? So go next. So they could have payment calendars, right? So right now, all the interest payments are as per banks prerogative. I am saying it should be mapped to farmers' harvest cycles when they can pay back to you. There are pledge walls where people can say, hey, I vouch. I want to give help to the community. The credit history builders. So women entrepreneurs there can actually build a credit over time. All those interventions can be experimented with using these discursive principles. Imagine the next scenario is it's a learning center, right? How people can learn. So there's not just selling products but learn. If you go next. And the prompt I gave to AI, this is what AI built for me. Like, isn't it beautiful? Like, isn't it like a dream come true you really want to build, right? And a couple of things you could do is you could simulate scenarios there locally to make them aware. You can have policy windows with policy makers. You can have financial safety nets so people don't save because they have limited funds but help them save in smaller ways. And the last one, it could be enable for women with sufficiency. So how can we enable women? And that's a prompt when I gave to AI. AI gave this back to me. So the beauty here is imagination is still a human thing. You imagine worlds. What AI today can help you build those worlds and think through the discursive questions which often you struggle at times to start from scratch. And things like scale-licked financing. So if I have a suing business or I'm running a local power business, can I get financing around that? Community rating systems and also story walls where women can tell their dreams, their goals and we learn and the AI learns good things about that. This could begin as I said a platform, an installation. It could be anything. And all this data understand all these themes we are extracting. For policy makers who are designing policies in isolation can look at all these streams of things that's happening and it curates into themes and topics and make much more informed decisions. Businesses can make much responsible products not for the only business gain but also for society gain and again trust metrics etc can be built. Now how do we do this? Right? There's a concept here I've thrown at you. A lot of ideas but how do we... Okay I'll do that. Okay good thanks. So how do we bring these ideas into this? Today we are training models with historical data but with all these themes emerging with these ideas we can train models with something called value-based embeddings. Embeddings are data which models learn from and build software and they can actually flow in to your commercial apps and business models to make it much more responsible for the end user. You can just take a picture of this slide because we really don't have time to talk about this. Yeah it was an intense topic. These are the takeaways and call to actions for you from this session. And again as I said in the beginning right as designers we should read more. Don't trust me okay. Go next. There's some reading references some books you should read and the more the designers talk about these lenses the better equitable world we will all design together until not this mess which has been happening for so many years thank you so much and we are hiring join IBM on these big problem statements. Thank you. Thank you so much.