Using Artificial Intelligence for Social Good @Neil_Sahota @IBM (Episode 35) #DataTalk

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.

To keep up with upcoming events, join our Data Science Community on Facebook or check out the archive of recent data science videos. To suggest future data science topics or guests, please contact Mike Delgado.

In this #DataTalk, we talked with Neil Sahota, an IBM Master Inventor and World Wide Business Development Leader in the IBM Watson Group. He is currently working with the United Nations to develop a model and set of metrics to encourage nations and organizations to pursue AI solutions for a more sustainable world. Learn more the U.N.’s Sustainable Development Goals.

Resources mentioned in the interview:

• University of California, Irvine
• InvestED
• United Nations Signature Conference [Video]
• Legalmation
• T-Cell Patent

Here is a full transcript of the interview:

Mike Delgado: Welcome to Experian’s weekly Data Talk, a show where we talk with some of the smartest people working in data science today. Today, we’re chatting with Neil Sahota. Neil is an IBM master inventor, he’s a worldwide business development leader in the IBM Watson group, he’s developed countless patents, he got his undergraduate work done in mathematics and political science, and then he went on to get his MBA at the University of California, Irvine, where he happens to be right now. He’s now faculty there. Neil, I’m curious about your journey, because your undergraduate work, mathematics, political science, were you leaning towards law back then? And what led you to start working with data and AI?

Neil Sahota: That’s a good question, Michael. The short answer, a lot of things happened by accident that opened opportunities. Long answer, math and political science are an interesting combination. I was always into science and math, the language of the universe, so to speak, but I always felt that people skills were important and understanding where people come from. Never was interested in becoming a lawyer, thought maybe politics, something I always had a passion for, but the pull of technology and the things that can be done was the lure that took me down the path. And I wound up where I am because I was always the kind of guy who likes to try and solve the problem in general and the problem on hand. And you start thinking about all the data and stuff that started getting collected, what you could do with that. And there are no machines or anything out there that can tell you, so learning the art of insight became a passion of mine, and that took me down the data science path.

Mike Delgado: That’s cool. It’s so unusual. When I saw your background, I was like, “Oh my gosh, he got his undergraduate work in both a BA in mathematics and political science.” It seemed like such different majors.

Neil Sahota: They are, but I think there’s actually a good mesh between them. Knowing the math, knowing the computer science gives you the hard skills of what you want to do, but when you’re trying to draw insight, particularly about people’s behavior and motivations, I learned a lot more from my poli sci education on that end. If you think about it, I could tell my students that I don’t believe you never actually predict the stock market, no matter how good your algorithm and stuff are, because one of the fundamental tenets of the market is people behave rationally, but is that really the case? How do you gauge irrationality? You have to understand people to be able to do that.

Mike Delgado: Nice. Christina just commented, “Yeah, the art of insight.” I like that. Because there is, definitely, that art that’s there, aside from the mathematics. Because you’re having to dig through and figure out, “Okay, how do I tell this story to make it insightful for business?”

Neil Sahota: Yeah, absolutely. And it’s one of those things where you can shift through the data all you want, but to draw those nuggets out, that definitely is an art form. To know what’s actually meaningful.

Mike Delgado: Now, today’s chat’s all about how organizations can use AI for social good, and you’ve been working with the United Nations, and I was wondering if you can share how you got involved in the United Nations and what you’re working on.

Neil Sahota: That’s another by accident–type answer. I’ve always had a passion for social good and giving back. I’ve been very lucky in my life, and my personal goal is to try and leave the world at least as good as, if not better than, I found it. And through my talks and work with various people, there were actually a few people who were working with the United Nations who said a few years ago, “Hey, AI’s becoming a bigger thing, and there’s a shortfall in terms of resources and funding for the United Nations’ sustainable development goals. Maybe we can engage Neil; maybe come out, give a talk in front of the delegation and help improve the awareness.” My first foray was actually at the United Nations signature conference a couple years ago, and I got to speak in front of the delegation. I did a 10-minute talk that apparently went over so well that they kept coming back, and I’ve been happy to help. I’ve become, over time, one of their AI subject matter experts.

I’m working with them right now. They’re putting together a report and a model to assess the economic impact of the sustainable goals, as well as emerging technologies. If I remember correctly, I think the UN has estimated there’s a $5 trillion shortfall in the resources currently available versus what we need to enable those goals. And they’re looking for ways to try and bridge those gaps, and AI is one of those options. We’re actually going to work with them to say, “How can we use AI for social good? How can we make AI an enabler for these STGs?”

Mike Delgado: That’s awesome. It’s so impressive to be invited to speak at the United Nations, and then you’re so good, Neil, that they’re like, “We need to keep working with you.”

Neil Sahota: Yeah. I do feel honored that they want my help, and I’m more than happy to help, and they’re doing some really good work for the betterment of humanity. And hopefully I can say in the near future that I made a big difference.

Mike Delgado: I was looking at the UN website that you sent me, and I noticed there are all these different projects going on. Are there any in particular you’re interested in?

Neil Sahota: I’m a big believer in the education, and zero hunger, and the inequality. I hate to say it this way, that while technology is a good enabler, it’s also causing a lot of stratification. And one of the things we’re finding out is that the people who have access to tablets and laptops and these other devices, and upper middle class, they can get computer programming classes and stuff. They’re learning the skills that the future jobs are going to need. But if you’re from more of an underserved area, it’s much more difficult to get access to these types of tools and learn what you need to learn to get those types of jobs. And as a result, as we’re going to more, I’ll call them higher-level types of skill sets, and less away from administrative type of work, that stratification is going to get worse. That’s going to, unfortunately, increase that income inequality gap that already exists.

Mike Delgado: What do you see as some solutions to that? Because income inequality’s a huge issue.

Neil Sahota: The gap through partnerships and maybe ecosystems, still. There’s a lot of companies like IBM and Google that are trying to provide education to these underserved populations, to try and enable some of the tooling. But again, you need to make sure they have simple things like even a Chromebook to be able to do some of these things. The other thing is, I got a chance to speak at the ICT4D conference last year, which is a conference for a lot of the large NGOs of the world. I talked about how they might have to restrategize, rethink some of their models going forward to try and leverage more of an ecosystem type of model, where they’re not just forming partnerships among themselves, but they’re trying to form partnerships among people who want to volunteer their time, be mentors and coaches, as well as private industries that can supplement some of their programming and some of their resources and funding. It’s one of those initiatives … it definitely takes a village to try and bridge this gap. There’s just no easy solution.

Mike Delgado: Well, it’s cool you’re helping to lead the charge, and it excites me to see organizations like DataKind that are out there to help get some data scientists rounded up for these data philanthropic efforts.

Neil Sahota: It is. And it’s one of my passions in life, working with people, trying to help understand AI, what they can do is help see … not just think about commercializing. And I get the commercialization. I think we all do, but there are opportunities to commercialize and do social good at the same time. There’s actually a startup company in Orange County called InvestED, where one of the things they’ve learned over time is you have these small villages where you’re away from even a bank, where they need a couple hundred dollar loan to improve their business, create more money in the local economy, and do a lot of good, but it’s this arduous process. You’ve got to wait for all these guys to show up in your region, do all these things, and they went down the path of, “Can we help facilitate these microloans through technology? Applications through a phone, which most people actually have, send pictures.” But then at the same time, they realized people actually have the financial literacy to understand what some of these moves might be or what the better use of the money could be.

And realizing they had a chance to help reduce economic inequality, they’ve instituted a lot of free financial literacy–type of programming, which they tell in the story. They’re doing a pilot in Gauda, they’re taking the story of a village woman in Gauda and having her walk through typical, normal lifetime things, but from a financial perspective for educational purposes.

Mike Delgado: Wow. You also mentioned one of your passions being education. How do you see AI fitting in with helping to improve education? Because I’ve always seen charts and graphs about which jobs could be automated, and education’s always been one of those things that’s less likely to be automated. But that could be totally wrong.

Neil Sahota: I think the goal is not to automate people out of a job with AI. I think it’s to help us free up our time to do more creative thinking, more complex tasks. But that also begs the question, “How do we get people ready for that type of work?” I think there are two sides to the education coin. One is that you can definitely use AI as a supplement, like a private tutor, self-learning, try and enable that, as well as help students, parents, educators understand a child’s or student’s strengths and weaknesses. And try and tailor the curriculum more in a personalized, individualized level, so they’re learning at their own pace rather than just as a group, and hope for the best, which may not be fair to all students on both sides of the spectrum.

But on the other side, what I’ve started seeing work with some universities, is they’re thinking about what the ramifications are and how they have to rethink some of the curriculum. For example, there’s a company called LegalMation that’s built an AI-powered associate lawyer that can do discovery work. This is core work that people at a law school do their first couple of years. They’ve been able to do it so that what used to take maybe 12 hours of an associate lawyer’s time can be done by an AI machine in two minutes. Which means, then, as this transformation occurs, what work do associate lawyers do? They’re going to have to do something different, which means that now law schools have to think about how they can then compared to associate lawyers.

There’s this ripple effect we have to think about. It’s not just the tooling to help us get better educated, but how do we then re-swizzle the curriculum so we’re preparing people for those jobs?
Mike Delgado: Yeah, no doubt. I’m curious about your role. You’re on the faculty at UCI, and I’m curious how you’re prepping your students for this new age of AI.

Neil Sahota: I actually warn them not to be complacent. I tell them that … actually, I try and expose them to some of the things that are going on, but I really try to hammer the point home that the world today is not going to be the same world in five years, 10 years. And I wish I could go to them and say, “Hey, 10 years from now, here’s what’s going to happen.” But I think things are changing so quickly that there’s no way to predict it. If you go back 10 years from today, how many people would’ve imagined a world of Uber, Airbnb and all that. Yeah, it’s really tough to say. So the best thing I can do for my students is to tell them, “Don’t be complacent. Keep your eye on where the trends are gonna be, but you’ve got to develop your critical thinking skills. Because no matter what job you have, you’re going to have to do more complex work.”

Mike Delgado: That’s really, really good advice. Now, I was looking at your speaking calendar, and you’re speaking all over the world. It seems like every week you’re in some different country. And I noticed that you recently spoke at UCLA, and it was at a health conference. And I’m wondering, because obviously health is a huge area where data and AI can be used to help people. I’m curious about what you shared there. What were some things you thought were insightful?

Neil Sahota: That’s a good question. I probably gave them a little scare at the beginning. One of the biggest challenges in healthcare is that there’s a massive amount of information; healthcare information is doubling almost every four years. And most doctors freely say they have less than five hours a month to read the latest medical journals. You think about all the clinical trials, everything going on. How do you keep abreast of that? And you think about more personalized type of medicine and trying to figure out medical history, family history, genomics, how all these things factor in. We have such amazing tools, but how do we weave them together?

We had a really good discussion around how AI can bridge some of this gap. That you basically have a tool that has an eidetic memory, and it remembers every medical journal it’s read, clinical trial it’s studied, all these things, and can help serve that information to the doctor instantaneously. And at the same time that the doctors and nurses are working with a patient, asking questions, you can have your little AI tool in the background, listening in, looking at what symptoms have been described, what symptoms might be missing, bringing in other information, pulling up medical tests so doctors have that dashboard to try and make the best diagnosis possible.

Mike Delgado: Wow, that’s amazing. To have an augmented doctor who has AI at his or her side to help make strategic decisions for health. That’s awesome.

Neil Sahota: And I think that’s just the way it’s going to be going. As a whole, you think about any kind of vertical today, there’s so much information that we can collect and are collecting about how you make sense of it and how we can apply it to, not just discover new insights and see new dots, but do something that’s actually going to be beneficial for us as a society.

Mike Delgado: I was looking at your LinkedIn profile, and I noticed that you have countless patents, and I even saw a blog post about a recent patent that you had developed just this last year called the T-CELL patent, which totally fits into this discussion on technology being used for good. Can you share about this T-CELL patent, what it does?

Neil Sahota: Yeah. I guess the best way I can describe it is we live in a dynamic world that’s rapidly changing. How do you adjust for changes? And we’re so used to traditional computing models where we write our executable code, and it’s basically following a path or scenario. But what happens if something in the system changes? You need a programmer to come in, take a look at what’s going on, maybe business analysts, identify the new path, exception path, whatever they are, and program that in. Well, that’s not becoming a sustainable model anymore.

With T-CELL, the idea is that you have snippets of code within your system that monitor for changes in the system. They’re looking for those new exception paths, or getting attacked by a computer virus can actually generate new code snippets to address that change. You think of it as more like dynamic code. But as much as we keep talking about machine learning, deep learning, we try and enable tools that allow machines to respond to environmental changes.

Mike Delgado: Cool. Now, you just used two terms, machine learning and deep learning. We’re talking about AI, and this is a question that always comes up, and I would love to hear how you define it. Because whenever I Google “What’s the difference between AI machine learning and deep learning?” I read different articles and different perspectives. How do you share with students the differences and how to practically look at how to define AI machine learning and deep learning?

Neil Sahota: That’s a good question, Michael. One that will probably plague us for years. Machine learning is definitely a component of AI, and I really see AI, as we know it today, as three pieces. And the first is the machine learning. That as the machine does something, it actually learns from its experience. You don’t have to teach the machine every little detail and every possible question. It knows enough that it can try and figure something out. And as it does something, it gets better at it. Unlike humans, where we need like 10,000 hours to master a subject, a machine can become a master in a matter of a few weeks.

This second piece is natural language understanding. We, as humans, don’t realize how much we draw from a context of a conversation, how much slang, idiom, jargon we use. If I say, “Hey, I’m feeling blue ‘cause it’s raining cats and dogs.” Most people would … right. But a machine is gonna be like, “Well, Neil is physically the color blue, because small animals are falling from the sky.” That’s huge. This natural language understanding is to be able to enable the machine to do that and say, “Okay, Neil’s not talking about being physically blue; he’s talking about being sad because it’s raining.” Really that ability.

And then the third piece is to actually interact with humans as if it were another human. For better or worse, search engines have trained us on keywords. If you want to buy a mountain bike, what are you going to do? Probably go to Google, type in “mountain bike,” get 50 billion webpages, start looking at them, learn a few more keywords, go to Facebook, see some fan pages, maybe visit a store. You take all this information and try to make a decision. But with a real AI, that interactive piece becomes much less work. I guess that’s the way to put it. If the AI knows about mountain bikes, you can say, “I want to buy a mountain bike.” And the AI probably will come back and say, “Well, why do you want one?” “I want to get back in shape.” “Okay, how often do you think you’ll ride it?” “Probably three times a week, an hour at a time.” “And where will you ride it?” “Probably around my neighborhood.” And the AI says, “Here’s the mountain bike I recommend for you.”

It’s more of a conversation. You happen to have a buddy who knows about mountain bikes, same kind of interaction with the machine. Those three things are really what comprise AI.

Mike Delgado: I love those examples. I read a study that was done in San Diego where they brought in these voice assistants, I think they were Alexa, into these senior living centers, primarily for seniors who lived alone, and it helped improve morale because they felt like someone was there with them. And even though Alexa and voice assistants are still in infancy stages, being able to have that digital companion made people happier.

Neil Sahota: That’s definitely true. I can tell you that there are a couple companies out there looking at using AI-powered robots as companions at elderly care homes. You have someone who’s literally there 24-7. Some of the AIs out there are smart enough they can actually detect the moods of people and respond by how they interact with those people. If they know someone is sad, they can try and cheer them up. They know someone’s happy, maybe they’ll crack some jokes. You essentially have the ultimate patient companion to these people.

Mike Delgado: I love that. I think it’s beautiful. Especially for those who can’t afford to have maybe a nurse care aid, because it can be very, very expensive. But to have a voice assistant to work with you, I think that’s a beautiful way. And I’m really excited to see what happens in five, 10 years, to see the progress.

Neil Sahota: Well, we can do this again in five years and do a recap and see what happened.

Mike Delgado: Yeah, we should do a recap, for sure. Neil, have you … I was reading about some voice assistants that are now available for B2B use. Have you seen that technology or know how that is being leveraged?

Neil Sahota: Yeah. It’s not that different from consumer use. They’re just using a corpus or body of knowledge specific to that company. A simple example would be law firms. They often have specific case roles associated with their work, like you can expense this or you can bill this. But these documents can be 20 to 30 pages long, and it’s really tough to remember it, especially if you’re working multiple cases. Now if you teach Alexa or something like that those case roles, you can just ask Alexa, “Hey, Alexa, I’m working on the Delgado case”, to try and simplify it.

The Irvine company has started a tech service at the Spectrum. They’re using IBM Watson for that, so you can text questions to this number and Watson responds to you. Like, “Hey, where can I find a good Italian restaurant?” Or “Where’s the closest restroom to me?” Or whatever it might be. They’re trying to enable these things to provide extra value at services.

Mike Delgado: That’s really cool. And I guess the more data they have from people, the better it’ll become, right, over time?

Neil Sahota: Yeah. And that’s the machine learning aspect of AI. The more it does, the more proficient it’ll actually become. It never stops learning.

Mike Delgado: That’s awesome. I definitely have to head to the Spectrum and try that out. That’s very cool. We’ve been talking about how AI’s being used for good, in the medical field, education, and at the same time there are a lot of people, and obviously there are all those tabloid headlines about rogue robots taking over the world and taking all our jobs. But we definitely see examples of machine learning doing amazing things in the personal finance space, with investing, as you mentioned, helping doctors. And I think there’s a lot of concern about AI taking over jobs. And I know no one can predict what’s going to happen five years from now, 10 years from now, what jobs will be automated, but I’m curious about how you see AI working with people in the future.

Neil Sahota: That’s a good question, and it’s a topic I hear a lot about. I get the fear from people. And the goal is obviously not to put people out of work. The goal is to give people a tool to help them do their jobs. The ultimate hope is that by letting the AI take more of the administrative, tedious type of work, we are freeing people up to do more creative thinking and tackle more complex things. So if you have an AI that can help you crunch some genomic sequencing and do some of the heavy lifting around trying to find a cancer drug, that allows the researchers to think more about different options. And these are the things we’re hoping will happen. The great concern, I think, is what will people actually do with this extra free time? Are they going to embrace the challenge of this more complex work, or do they more likely want to be entertained, maybe watch cat videos on YouTube?

And I think that’s what drives a lot of the fear, and that’s why you hear a lot of organizations, especially government organizations, talking about UBI, universal basic income. We know, unfortunately, that some people are going to choose the cat videos, so what do we do? And I don’t think there’s a simple answer to this. I think this is something we’ve started thinking about and we’re probably going to put a lot of thought into what’s going to make the most sense.

Mike Delgado: That’s good stuff. Neil, we have a couple minutes left, so I wanted to get to the final questions we ask all our guests just to get your perspective. What is your favorite programming language and why?

Neil Sahota: Favorite programming language. Man, I don’t know if I actually have one. These days Python and R are pretty good, especially for data science, so we’re all with Python.

Mike Delgado: You’re like children. How do you choose?

Neil Sahota: Yeah.

Mike Delgado: Okay, cool. What advice do you have for people who are interested in getting involved in data science?

Neil Sahota: That’s a good question. I would say the best advice I can give you is, I hate to put it this way, rise to the challenge. There’s a lot of need for data science, but it’s not an easy field. You have to know a bit about statistics, you have to know a bit about computer programming, you have to know a little bit about the domain space you’re working with, understand the data, you have to be a little bit artist to know how you can weave the patterns you see into meaningful insights. It’s quite an eclectic collection of skills, but it can be a very worthwhile endeavor to get a lot of those “aha” moments. If you want to be a data scientist, don’t get discouraged by the complexity of the work. Step up and try and make it happen.

Mike Delgado: Beautiful. I love the way you put that. Painting the picture of the mathematical side, the artist side, and the curiosity side have to all be there. Okay. One of the last questions: What advice do you have for leaders? I know you speak to leaders all over the world at conferences and as you go in to consult. What advice do you have for leaders who are looking to build a really solid data science team?

Neil Sahota: That’s a good question, too. And I tell them you have to think differently. When it comes to data science and AI and all this stuff, it’s not your traditional computing model. Don’t let yourself get locked into the traditional box. Be prepared to think differently about how things work. Which means that when you want to build a really great data science team, don’t just look for the raw skill sets; look at people who can think differently and find those hidden nuggets that are out there that are going to yield those “aha” moments and drive real value for you.

Mike Delgado: If you were hiring a data scientist, are there certain skill sets you’d be looking for from that person?

Neil Sahota: In addition to the traditional hard skills, I would actually see if they have that artistic side. If you give them a potential problem where most people will go down one path, but it’s not going to pan out, are they clever enough to step back and try a different approach? Because if they are, they’re usually going to be a great data scientist.

Mike Delgado: Nice. Well, Neil, I want to thank you so much for being our guest. Where can people learn more about you and your work?

Neil Sahota: They’re welcome to visit me on LinkedIn and connect with me or check me out on Twitter. And if they have any other questions, you can reach out and ask me, as well.

Mike Delgado: Awesome. I put up a URL on the screen, and for those listening to the podcast, the short URL is ex.pn/neal. If you go there, it’ll redirect over to our blog post; it’ll have links going over to where you can connect with Neil on LinkedIn, Twitter. And for those watching the video, we’ll make sure to put those links in the About section of the video as well as in the comments, so you can connect with Neil. Make sure you’re following him on LinkedIn. He’s very, very active. Check out what he’s doing there, comment, and … Neil, thank you so much for all the work you’re doing in the data science field, especially data science for good, and the work you’re doing for the United Nations. It’s awesome that you have a heart for that and making our world a better place. Thank you so much.

Neil Sahota: Yeah, my pleasure. Thanks, Michael, and thanks for the opportunity to speak with your audience.

Mike Delgado: Great. Take care, Neil.

Neil Sahota: All right. Take it easy.

About Neil Sahota

Neil Sahota is an IBM Master Inventor and World Wide Business Development Leader in the IBM Watson Group. He is currently working with the United Nations to develop a model and set of metrics to encourage nations and organizations to pursue AI solutions for a more sustainable world. He is also a Lecturer at the University of California, Irvine – The Paul Merage School of Business.

In addition, Neil partners with entrepreneurs to define their products, establish their target markets, and structure their companies. He is a member of several investor groups like the Tech Coast Angels and assists startups with investor funding. Neil also serves as a judge in various startup competitions and mentor in several incubator/accelerator programs.

He actively volunteers with nonprofits for event management, fund raising, grant reviews, and site visits. Neil is an active member of the UCI Alumni Association and serves on the Board of Directors for the Orange County Marathon, supporting their work with the OC Kids program in fighting childhood obesity.

Check out our upcoming live video big data discussions.