Listen to the podcast:
Every week, we talk about important data and analytics topics with data science leaders from around the world on Facebook Live. You can subscribe to the DataTalk podcast on iTunes, Google Play, Stitcher, SoundCloud and Spotify.
This data science video and podcast series is part of Experian’s effort to help people understand how data-powered decisions can help organizations develop innovative solutions and drive more business.
In this #DataTalk, we talked with Dr. Ayesha Khanna about the future of artificial intelligence.
Here’s the full transcript:
Mike Delgado: Hello, friends. Welcome to our weekly #DataTalk, where we talk to data science leaders from around the world. Today we’re talking about the future of artificial intelligence. Which I know is a generic headline, but with Dr. Khanna, this conversation could go anywhere. She speaks on so many different data science topics, AI. I don’t even know what to title this. The future of artificial intelligence is very broad. We’ll just see where this conversation takes us. For those who don’t know Dr. Khanna, she is the CEO and co-founder of AD … Is it ADDO AI?
Ayesha Khanna: Yes. ADDO AI.
Mike Delgado: If you’re searching on Google, it’s A-D-D-O space A-I. Check that out. She’s also the founder of 21C Girls, which is a charity offering free coding and AI courses to girls in Singapore, which is an awesome initiative. Just shows her passion for teaching AI and helping other people get started. She’s been featured all over. She’s spoken at TEDx, she’s featured in The New York Times, Forbes, Harvard Business Review, etc. She has an amazing academic background. A B.A. in economics from Harvard University, an M.S. in operations research from Columbia University and a Ph.D. in smart city infrastructures at The London School of Economics. We are grateful and honored to have you as our guest today.
Ayesha Khanna: Thank you so much, and thank you for that glowing introduction. That was the best introduction I’ve ever gotten.
Mike Delgado: Thank you. It’s a pleasure. Our data science community loves learning from people like you who have been in the field for a very long time and started companies. You’re an inspiration to so many of us in the data science community. Our first question that we always ask is, can you share your journey? What led you to begin working in data science and then eventually start a company?
Ayesha Khanna: I’d always done a lot of math and statistics in high school. It was only when I was an undergraduate at Harvard, where I met a lot of mathematicians from Eastern Europe, specifically Romania. And my Romanian friends looked at math and statistics not as cold, engineering heuristics that would get you to an answer, but a wonderful way to understand the complexities of science. And for them, it was beautiful.
Listening to them talk to each other made me think of data science and all the statistics and econometrics that I was studying as really a tool for problem solving, for pursuing my own curiosity and as something that’s creative. Which, growing up in Asia, one tends not to associate science and engineering with creativity.
When that clicked in my mind, I went to Columbia University and I did a master’s in operations research, which was a lot of applied math and statistics and computer science. And I did some computational biology. I just fell in love with these new fields and the potential to do interesting things and reach insights with data and with technology.
Then I spent 10, 12 years on Wall Street. I started off as a software engineer … and then when I moved to Asia, I saw more and more digitization of services, more and more people starting to use mobile phones. Such a lot of data, but a great gap in talent. And that’s when I decided to start ADDO AI.
But of course, it requires deep talent. I am by no means good enough alone and definitely not even compared to my awesome colleagues, who are all Ph.D.s in AI. And incredibly curious, wonderful team players. I got very lucky that I got a fantastic team. It all came together with ADDO AI.
Mike Delgado: Aside from all your amazing accomplishments, you’re extremely humble.As you were building this company, what were you looking for to help you be successful? Because you’re helping companies with AI initiatives. What skill sets were you looking for, what personality types were you looking for?
Ayesha Khanna: The first important thing is the theoretical foundation. It’s important that you just have not learned the functions to apply, which are sometimes [inaudible 00:05:21], but that you’re understanding the concepts. You understand gradient descent, for instance. You just haven’t used the function and the methodology. And you understand what a logistic regression is. And you understand what reinforcement [inaudible 00:05:39] model is. And that’s important. And that kind of interest is important to take. And then, for artificial intelligence, your computer science and your programming skills need to be really good as well.
The second thing that counts is are you a team player? And the reason that matters is because nobody, no superstar in any consulting firm or any team, can work on his or her own. You need to be somebody who can collaborate with other people. And in our firm, these other people are not just AI specialists or data architects. We also have social scientists. Because you can’t build customer-centric products and services without understanding human psychology and behavior. And we also have people who are domain experts. Because no AI model works without knowing [inaudible 00:06:32] insurance or transport, what are the problems of that industry. And then you have visualization experts. These are experience designers. You have to get along not just with people in your own field, but a lot of people. It’s pretty much like real life.
And then the third thing is what is your potential for leadership? That, of course, is something that emerges over time. And that has three components to it. First, are you inspired yourself, with passion? Because if you don’t have passion, it’s very difficult to lead. It’s difficult to convince clients what would be a good solution. It’s difficult to inspire your team. Secondly, are you organized, which means you can manage. You have discipline, which is the foundation of creativity. And systematic discipline is especially important in our field.
And the third one is if you have resilience.
Because in the start of a young company, I get people who come and they’re very easily disappointed. We listen to “no” a lot. Everybody’s watched Black Panther and expects magic to come out of data science and AI. And you know, you just have to roll with the punches. And your job is to be a trusted adviser. To have your core ethics strong … Because with data, there are all kinds of issues and biases. And you can’t do it, not for the money, not for anything. Because our clients trust us.
So those are the three things that I look for in the team.
Mike Delgado: That’s amazing. Because I’ve heard so many people talk about the programming part of it, and understanding certain languages, and having a background in STEM, statistics. And you’re talking about a lot of the soft skills. The artistic side, which is a big part of innovation. You need to be creative, you need to be passionate. A lot of people don’t think about that artistic, creative side to data science. That’s amazing. Because, and I’ve shared this in a previous show. When I was in school, there was the humanities section, right?
Ayesha Khanna: Yes.
Mike Delgado: And then you have the STEM, mathematics section. And we rarely cross paths. You choose one or the other. And I feel that whenever I talk to data scientists like you, data science leaders, you have that mixture of both. The creativity of the humanities is so pivotal, along with understanding the math and the science behind making things happen. That’s brilliant to hear that. And very cool to hear how you’ve structured your team. Tell me about the work that you’re doing now with ADDO.
Ayesha Khanna: At ADDO AI, we help clients think of how to use artificial intelligence to either solve some of the challenges they are facing or some disruptive competition they’re facing. Even though we build algorithms, that is not only what we do. First of all, we spend a lot of time with a client, identifying the problem statement. Why do you want to use this? So I often have [inaudible 00:09:56] and they’ll come to me and they’ll say, “We just want some AI. We just want some data science.” I’m like, “But why?” And they’re like, “We don’t know why. Please tell us why.”
It’s very misguided. Because new technologies will keep coming. But ultimately, every business is there for its customers and the strategic vision of its employees. Once you get that, then it’s easier to think about how to systematically go after that, including using AI data science.
The second thing is we build the distributed data architecture. People don’t appreciate that to do enterprise AI, you need to have the full architecture with it. It’s not some cutesy little program on a laptop or one instance of … Some storage space and AWUS. You need a proper … for streaming data, batch data or depending on the frequency it’s coming in. And all of that needs to be set up. Without that, you cannot run the algorithms. And you need all the scalability. You need search mechanisms.
And then finally, at the same time, there are out-of-the-box algorithms. But, depending on the quality of the data, you add what you’re looking for. You have to hybridize them. You also have to have an intuition for big data. And that’s why we have so many professors who work with us. Because they have so much experience, and they have an intuition. “This doesn’t sound right, guys.” So that helps us avoid rabbit holes.
We were lucky we’re working with the largest transportation company in Singapore. And we’re integrating a lot of different vehicles into one platform and then looking at mobility patterns. We’re working for the largest insurance company in Japan, Sompo. We’re working with the governor of Dubai. We’re going to start working with the largest bank in Pakistan. What distinguishes these companies is that the leaders have a lot of bold vision and they’re willing to take a risk and use AI, and all the pain and joy of that process.
Mike Delgado: The pain and joy, yeah. It’s funny that people come to you and they’re like, “Dr. Khanna, we need your help to implement AI,” and you’re like, “What’s the business problem here?” You’re not just here to sell an AI solution. Because if it’s not going to add business value, it’s going to go nowhere, right?
Ayesha Khanna: Exactly. And then everybody’s disappointed. It is to nobody’s benefit. It’ll be very short-term thinking for me and very bad advice for me to the client, or to any company, or any government, if they did not have some goal in mind. And of course, the beauty of AI is, as you know, that you see unexpected correlation. You see interesting insights.
My co-founder won a major award from The World Bank last year. Because in Pakistan they were trying to have more men and women from the outskirts of cities come into the city to work. And they were going to build all these feeder bus routes inside. But they couldn’t understand why more women weren’t taking them.
You put the buses, why aren’t the women coming? And then he started looking at all kinds of factors, from the position of the light at the bus stop to how much light there was in a bus to how far it was from certain stops to how the seats were organized to the weather that day. And he was able to come up with very interesting insights on how to increase women’s participation by changing some parameters of how you put urban mobility infrastructure in place.
New ideas come up. Innovation comes up, but you need to start with some question. And without that question, you’re lost.
Mike Delgado: That is so cool. It sounds like you have an amazing team of individuals who are winning awards for the work that they’re doing and helping to solve problems like that. Starting with a problem and then figuring out if we have the data to help solve it. That’s beautiful.
In the news, the headlines that sell, the headlines that get all the attention, are those scary ones. The ones about the robots taking over the world, job loss … We know that there’s going to be job loss as automation is helping humans with certain tasks. I know this is the million dollar question. What do you see as the future, 20 years from now, as AI is more incorporated into the work that we’re doing? Helping us with tasks, as we have voice assistance that can answer questions quickly. I’m excited about this future. I’m excited about new jobs being created. Obviously, there’s also the down part of certain jobs going away. I was curious about your view of how AI is going to impact our future.
Ayesha Khanna: The important thing is to understand that any technology that’s going to have such a massive influence on humanity is a double-edged sword. You can’t be naively optimistic, and you also can’t be depressingly pessimistic. You have to take on the responsibility, not as passive observers, but as citizens of the world. As employees of companies that have access to data, and technologies like this, to steer it into the right direction.
First of all, will there be displacement of jobs? Yes, there will be. I see a lot more in the developing countries. A lot more optimism. In Malaysia, Indonesia, in China. People are excited about it, because they see this as a tool for social mobility. They see this as an opportunity to leapfrog into the more developed countries. And of course, the populations are bigger. There’s a huge emerging middle class. And for the first time, they have access to relatively cheaper, high-quality education because of [inaudible 00:16:18] and others.
In the developed world, there’s a lot more hesitance and fear about automation because they are not seeing the opportunity of new jobs that easily. And yet we know that new jobs will come.
For instance, if it is healthcare. The promise of personalized healthcare has been with us for so long. But I think its age is now going to come upon us. With greater computational power and the ability to apply models, like deep learning and reinforcement learning. And so personalized medicine will happen. There’ll be many new jobs that come up in this field. But does that meant that traditional jobs like accounting will stay? No. Probably not.
We have to guide people who are just coming into universities on what the future jobs are so they feel optimistic and they’re guided in the right way. And then we have to help people who are losing their jobs for some time and give them some skills upgrading [inaudible 00:17:22] so they can transfer their skills to the new kinds of jobs that will be available. Overall, all said for the next couple of decades, I’m very positive. I think this will do a lot of good. Of course, I do worry about accountability and transparency.
There’s other things, like fake news. There’s also fake voice and fake video. The potential to manipulate people is more than ever before. And that’s where we have to be very conscious; we have to be very careful. We have to invest in AI research that counters this kind of manipulation by AI as well. We need a two-pronged approach toward it. And a very balanced and reasonable way to move forward.
One thing is for sure. There’s no moving back. You can’t stop the march of progress.
Mike Delgado: That’s right.
Ayesha Khanna: If you start with that assumption, then you deal with it. Then you try to take the right path.
Mike Delgado: In what fields are you seeing the most movement right now?
Ayesha Khanna: We see a lot in computer vision. Is that what you meant?
Mike Delgado: Emerging fields that are using AI or …
Ayesha Khanna: Industry-wise, one area where we see a lot is in smart cities. Smart cities is the application of technology to urban environments. Healthcare, infrastructure, police, education, government services. And that’s because there are two major trends that are going to transform humanity in the 21st century. One we’re familiar with, which is technology. The other is urbanization.
200 million people are going to move from villages into cities in India. 400 million people are going to move into cities in China. Over 70 percent of the entire population of the world by 2070 will live in cities.
And cities are groaning under this influx of migrants, who are looking for interesting places for social mobility and access to infrastructure. So we see a lot of work with [inaudible 00:19:39] transportation. We see a lot of work in healthcare. And private, public sectors both are investing in the kinds of digitization and other things you need to make these services better for customers and citizens.
Another area where we see a lot is in security. Because there’s a great need for policing, for security, and all kinds of things going to that. In Asia especially, we see a lot in retail, with the emerging middle class coming up. And in financial inclusion. So, in some countries, less than 20 percent of the people have a bank account. But over 80 to 90 percent have a mobile phone. When you want to bring these people into the financial services sector, you need to give them a credit score.
But you have no transaction history, like we do in the U.S. You know, you get a credit score, which is both the bane of our existence and the advantage, depending on where you are.
So you can use other data, such as data from your mobile phone to other geographical data to other information, and correlate that to find a new credit score for a person. This is going to change the lives of people. You can use satellite data for farmers. You can help them understand when to plant seeds at a certain time. Or how much their crops were destroyed by a typhoon, using deep learning and satellite imagery. You couldn’t do that before. There is tremendous potential across all fields for AI to truly change the business models and the democratization of access to basic services.
Mike Delgado: That is exciting. We have an office in India. We had an interview with one of our project managers there who was talking with a lady who came from a remote village. And she was trying to get a small loan for her business. And she, by trade, was a sewer. She wanted to buy some sewing machines. She was talking to him about how she can’t prove her identity because she lives in this remote area; she has no birth certificate. So you have all these different problems. But like you said, there is alternative data out there that can help prove who this person is and their income stream, maybe through their mobile devices, to help them get the loans they need.
Ayesha Khanna: That’s why I think us data scientists and AI engineers have a really exciting future in front of us. Because there are so many problems to solve in the world. There’s no end to problems. In the developed world, there’s Alzheimer’s, there’s diabetes, there’s chronic disease. There’s overpopulation in certain areas that would be helped from better infrastructure. I’m really excited about the future because I see so many ways that people can make money and do good at the same time. You don’t have to sacrifice one for the other.
Mike Delgado: That’s right. That’s awesome. Using data for good. Data philanthropy. That’s beautiful.
I know we’re out of time, but just two quick questions. The first one is, what is your favorite programing language?
Ayesha Khanna: Mine is Python. Because it’s easy. It’s so readable. It is also powerful. In data science, machine learning, Python is the most popular. And it has some great libraries as well, like NumPy and SciPy. And I find that for beginners, especially, if they want to join, it’s a great language to start with. We use it a lot for our clients as well.
Mike Delgado: Awesome. And then the last question is, what advice do you have for people who are interested in data science?
Ayesha Khanna: First of all, I’m really happy to hear that. For people who are interested, I would highly encourage you. This is a field that’s only going to grow. The question is where to start. And you can start at Coursera, and you can start Udacity. I often audit a lot of courses from there. Don’t just jump to very applied courses. Take the time, and take the statistics course, that is a prerequisite. Take the computer science course. It’s really worth it. Because, in the long run, you will learn that that’s the foundation on which you can pivot a little bit, as the demands of the job change.
The second thing is what we talked about earlier. Look at the humanities. Look at business. Look at geopolitics. All of these things inform the decisions that you make for your clients or for your own company. Keep an interdisciplinary approach. Don’t just be very siloed. Because you will have to work in interdisciplinary teams in the future.
Mike Delgado: I love that. And Dr. Khanna, it’s so cool to hear, with all your education, your success, you’re still auditing courses. You’re still going to Coursera. So for those who are listening in and looking to be data scientists, there’s no excuse. Look at Coursera. Take Dr. Khanna’s lead. Thank you so much for your time. Where can people learn more about you?
Ayesha Khanna: They can go to my website: AyeshaKhanna.com. They can also find me on LinkedIn. I’m also on Twitter, Instagram, Facebook. I’m on all the social media outlets. I love to hear from people, and I love to keep in touch.
Mike Delgado: Awesome. On our Experian blog, we’ll have this video and podcast and a full transcription. I will put links to all your social profiles so people can connect with you, follow you, keep up with all the work you’re doing. And for those listening to the podcast, the short URL is just ex.pn/datatalk42, and that’ll bring you over to Dr. Khanna’s blog post that’ll have all those links, all those resources. And the other things you mentioned, Coursera, I’ll make sure to put links there too for people who are interested in that.
Thank you for spending your Saturday morning with me and the #DataTalk community. I hope you have a wonderful weekend, Dr. Khanna.
Ayesha Khanna: Thank you so much. You too. Thank you for having me. Bye.
Dr. Ayesha Khanna is Co-Founder and CEO of ADDO AI, an artificial intelligence (AI) advisory firm and incubator. She has been a strategic advisor on artificial intelligence, smart cities and fintech to clients such as SMRT, Singapore’s largest public transport company, SOMPO, Japan’s largest insurance firm, and Smart Dubai, the government agency tasked to transform Dubai into a leading smart city. In 2017, ADDO AI was featured in Forbes magazine as one of four leading artificial intelligence companies in Asia and Ayesha was named one of South East Asia’s groundbreaking female entrepreneurs by Forbes magazine in 2018. Ayesha is also the Founder of 21C GIRLS, a charity that delivers free coding and artificial intelligence classes to girls in Singapore.
Prior to founding ADDO AI, Ayesha spent more than a decade on Wall Street developing large scale trading, risk management and data analytics systems. Ayesha was co-founder of the Hybrid Reality Institute, a research and advisory group established to analyze the social and economic impact of accelerating technologies. She directed the Future Cities Group at the London School of Economics, and has been a Faculty Advisor at Singularity University.
Ayesha has been named one of Singapore’s leading female entrepreneurs and a leading Asian fintech influencer by Fintech Asia. She also served on the Singapore Ministry of Education’s Steering Committee on future skills and applied learning for emerging industries.
Ayesha is author of Straight Through Processing (2008) and co-author of Hybrid Reality: Thriving in the Emerging Human-Technology Civilization (2012). She has been published and quoted on technology, innovation and smart cities in The New York Times, BusinessWeek, TIME, Newsweek, Forbes, Harvard Business Review, Strategy+Business, and Foreign Policy. She has presented at major financial, technology and other industry conferences, provided high level government briefings, chaired symposiums such as AI Asia, and spoken at TEDx events.
Ayesha has a BA in Economics from Harvard University, an MS in Operations Research from Columbia University and a PhD in Information Systems and Innovation from the London School of Economics.
Ayesha is a Fellow at the Institute for Ethics and Emerging Technologies and on the Board of Advisors for Humanity+. She is an advisor to the startups Octa (a chatbot for young travellers) and Arro (a delivery robot for sports). Follow her on LinkedIn
Check out our upcoming data science live video chats.