AI’s impact on people, organizations, and countries is now so far-reaching there soon won’t be any aspect of our lives left untouched by it. To even begin understanding how wide-ranging the effect of artificial intelligence is currently, we undertake a wide-ranging discussion with Neil Sahota, AI advisor to the United Nations. Neil takes us on a grand tour of some really exciting ways AI’s capabilities are being used around the world today. He also introduces us to a thinking framework he developed that can guide us on how to harness AI’s capabilities to create disruptive advances.
Along the way we'll learn what the biggest challenge is for organizations deploying artificial intelligence, the unheralded metric which best captures the impact of AI on business & IT operations, & what the biggest area of growth will be for artificial intelligence over the next 3 years.
Guy Nadivi: Welcome everyone. My name is Guy Nadivi and I'm the host of Intelligent Automation Radio. Our guest on today's episode is Neil Sahota. AI Advisor to the United Nations, IBM Master Inventor, and Chief Innovation Officer at UC Irvine. While at IBM, Neil was part of the team which created the AI system that won Jeopardy, beating famous all time champions Brad Rutter and Ken Jennings. Neil applies his domain expertise in AI to a number of business and entrepreneurial ventures, as well as to social causes such as the UN's AI for Good initiative. As if all that wasn't impressive enough, he's also the author of an Amazon bestselling book titled, Own the A.I. Revolution, a guide for how to take advantage of the opportunities being created by AI's disruptive impact on just about everything. We're always on the hunt for just those types of insights on this podcast, so we've asked Neil to join us today and provide our audience with some of his highly sought out advice. Neil, welcome to Intelligent Automation Radio.
Neil Sahota: Hey Guy, thanks for having me on, I love your show and I'm looking forward to a great conversation with you.
Guy Nadivi: Same here. Neil, why don't we start by having you tell us a bit about your background?
Neil Sahota: Well, I am the guy who always sought the path of most resistance. I know that's counterintuitive, but I've always been inquisitive and had a yen to learn. And I will be honest in that I found it's the more challenging classes, projects, or things in life that really drive the growth. And that's really enabled me to do well several kind of once in a lifetime opportunities, like being part of the original IBM Watson team, being an advisor to the UN, to helping all kinds of people. So I count myself lucky, but it's that mix of most resistance, putting those puzzle pieces together, and just a natural curiosity that's taking me down this path.
Guy Nadivi: You mentioned the UN and I'm curious, how did you end up being an advisor to the UN on artificial intelligence?
Neil Sahota: That's a interesting story. Well, Guy, it actually started with my good friend, Stephen Ibaraki and I've known him for quite a long time and we help each other out, and I was helping out with something with the financial services round table and he called to thank me afterwards. And he said, "Hey, I have an opportunity for you." He's like, I sit on this committee for the United Nations. They have this once every four years, big event with all the world leaders, all the ambassadors and, this is 2015, so they're interested in AI and I put your name forward to speak, would you be interested? And I'll be honest, I thought he was totally joking with me. I'm like yeah right, Stephen, he's like, no, I didn't have to even sell you. Two of the young guys said we've heard Neil and familiar with his work, you think he'd do it?
And I honestly thought he was just pulling my leg and he was taking the joke too far. It wasn't until that night, when I got an email from the Chief of Staff of the Secretary General, the formal invite to come speak at this event that I'm like, oh my God, this is real. So that, was my foray I was warned that most of the leaders thought AI was terminator time. Robots were going to rise up, conquer the world, eradicate humanity. So no pressure, but I give a very optimistic speech, but I didn't just talk about what AI was, I actually showed them how we're using it in public service and how it could be used towards the UN Sustainable Development Goals. And that seemed to really connect with a lot of leaders and ambassadors.
That evening at the reception, the Secretary General came up to me and said, "Neil, I never realized we could actually use AI. There's a lot of excitement around this. I want to figure something out and can we talk about it?" I'm okay, yeah sure. Set up something in the next week, I thought, okay, I'll probably meet with couple of his people who knows what will actually happen. No, Secretary General was actually there himself and committed the resources of the people, realized the potential here, because there's obviously a lot of challenges to try to make the STGs a reality, but that led to the creation of the AI for Good initiative as well as realizing the focus on solutions and then actually becoming the UN's AI advisor.
Guy Nadivi: I've heard you say that the United Nations suffers anywhere from a $7 to $20 trillion shortfall annually, preventing it from achieving many of its humanitarian objectives, but that AI, "Has been a great bridge in narrowing some of that resource gap and actually accelerating solutions for people." Can you give us an example or two of how AI has done that?
Neil Sahota: Sure. It's sad but true that there's this big shortfall. But we found that AI and other emerging technology could be a really effective way to bridge some of that gap. So one example, actually one of the oldest human professions of agriculture, we've actually been able to develop AI tools, very simple to use for farmers in impoverished areas. And so, analyzing weather data and top soil conditions, but also the potential of insect infestation, figuring out what kind of seeds would work the best, what would be food crop versus cash crop, and prediction of price. I won't bore you with all, 2,000 plus kind of variables going on here. But the long and the short of it is, we were able to help farmers, even in places like Bangladesh grow 30% more food using 20% less water and 10% less top soil resources. And they don't require super computers or anything like that.
Very simple mobile devices just to give them... And we've seen the AI even say stuff like just plant this seed two millimeters to the right to actually accomplish that. Flip side, you look at healthcare, which is foremost on so many of our minds during the pandemic. You look at Africa where there's one doctor for every about 2,000 people. If I remember correctly, the average person lives about 70 kilometers away from a doctor or healthcare facility.
What we learned was is that it's really tough, obviously for them to get care, but we've developed simple AI systems that, just using a tablet, so again, you don't even need an internet connection, just need to be able to charge it. The tablet, any kind of a villager could take it, put information in or use the camera, or the video feature, the AI would prompt some questions and help kind of diagnose and then help the villager be able to try, sorry, apply some sort of treatment. Now, if it's more serious, it can actually alert the nearest healthcare facility and even send a helicopter in, but essentially turned each local person, each average citizen into a physician's assistant. So those are a couple of the projects that have actually been rolled out from this initiative.
Guy Nadivi: Very interesting. Neil, you advocate for a disruptive thinking framework, you call by its acronym, TUCBO, T-U-C-B-O. What is TUCBO and how can IT leaders apply to digital transformation?
Neil Sahota: Well TUCBO, it's really about thinking differently, right? Everyone always wonders, how is it guys like Elon Musk or Jeff Bezos have all these amazing ideas? How do you get that light bulb moment? It's not magic. That's what I've learned, right? It's not just, you're just sitting there and you had this epiphany. It's actually, there were slowly seeing all these dots and we learn how to try and connect these dots or at least we should be. And as we connect that last dot, we get that, think different moment. That kind of light bulb moment. The challenge is we don't really teach that, and we always tell people, oh, if you want to be innovative, just think differently, but no one tells you how to do that. TUCBO is actually my framework and how you actually can think differently. It's actually something I developed in my days as a management consultant for global Fortune 500 Companies.
So all these amazing ideas, all these things that people, said I did for them, actually came through this framework I developed for myself and it's something I've been able to teach other people. And it really TUCBO stands for Think different, Understand different, Create different, Be different, Own different. And so it's a soup to nuts from essentially think different the ideation. So what's the idea, how do you get that? To understand, which is the validation of the idea. So is there actually value and is it feasible? To then create different, which is the actual design implementation, the Be different is really then, okay, how do you actually roll this out? And how do you actually get, the adoption for it? And the Own different is actually, how do you actually build the infrastructure to support it?
Which is unfortunately, something a lot of companies overlook. So I've seen people use it for their whole enterprise, spur up a new product line or solve some operational issues, or I've even seen people use it individually for their own career paths. But it's something that's actually a repeatable framework that actually teaches you how to think differently, how to actually find those dots, learn how to connect them faster, so you get to that aha moment quicker.
Guy Nadivi: What are some of the low hanging fruit best suited for AI applications within an organization?
Neil Sahota: Typically, the low hanging fruit is something where you have little variation, you have data, you have the trainers or subject matter experts to do it. So it's something that is reasonably known. So you think about, processing insurance claims or several police departments are actually now using AI chat bots to collect police reports. So if you have an issue, rather than talk to an officer now, you can actually talk to the chat bot, it'll actually collect all the pertinent information by prompting you with questions and all kind of stuff. And so it's those types of activities that are the low hanging fruit. It's not the whole universe. It's typically a good place for most organizations to get started.
Guy Nadivi: When you speak with business customers, NGOs, governments, and others, what are they telling you are some of the biggest challenges they're experiencing in deploying AI?
Neil Sahota: I have to laugh because it's almost universally getting people to trust the technology. For some innate reason, and maybe it's a lot of the books, the TV shows, the movies that we've watched, that we believe that we as humans are special, and we are special, don't get me wrong. But you always think about the human always triumphs over the machine, in all our literature or shows, that it's actually tough for people to believe that somehow this AI system can do this and can do it as well as, or maybe better than a human being. So I'll tell you Guy, if you'd asked me that question three and a half years ago, my answer would've been different. Because back then, most people didn't believe these capabilities even existed. Today it's different.
No one questions the capabilities, no one questions what's possible. They recognize the importance of meaningful data as well as good training. It's that last piece now where the people that actually will need to use it, do they actually believe this is going to be effective or not? And that's the biggest challenge and unfortunately it's an organizational change management type of issue. And that's something that unfortunately most enterprises tend to overlook. So I can tell you from my own experience, 10 years ago, as we started working with healthcare professionals like doctors and nurses - just give them tools, so they can do their work. We're not trying to replace them, but a first reaction was... Wasn't even that, oh, you're going to replace me with the robot doctor, it was like, oh, there's no way this machine knows the stuff better than I do.
I've got 20 years of experience and there're just subtle things the machine will never get and that kind of stuff. And it's not here to be perfect, it's not here to replace you. But if you're in the ER room and you have a patient running in, you're going through symptoms of things really quick. It's real easy to make a misdiagnosis. In fact, on a normal basis, 20% of patients are actually misdiagnosed. ER is even much higher. Just imagine you have a little system, they're just listening in, seeing okay what symptoms do they have, what symptoms do they not have to help kind of isolate the top five potential reasons for why they're ill or they're injured, right? Still doctors making the final call and knowing all these things, but there was so much resistance to that when we first started and we still see that challenge today, less so in healthcare thankfully, but for most organizations going down this path, that's the biggest challenge they actually face, it's a people problem.
Guy Nadivi: Speaking of people, there is a lot of talk these days about a labor shortage due to the pandemic. How can AI help organizations overcome this labor shortage and contribute to organizational resiliency?
Neil Sahota: I'll start with the obvious answer. I'm sure most people think, well we can automate some jobs to AI and there's definitely truth in that. You look at the fast food places where I think there's already now, several McDonalds where you go through the drive through, you're actually talking to an AI machine. So it's actually... even though it sounds humans, very conversational, it's actually collecting your order information and then transmitting that to the store to make your food. So the automation is the obvious answer. Then there's a couple of two more, I'll say less obvious answers where AI is actually helping with the labor market. One of them is actually around just the talent management and recruiting. And so if you ever try to hire somebody, we always look for people that are qualified and will be a cultural fit, they'll fit in with the team and their work style matches the corporate culture.
Well the qualification is very easy to check. You interview them, you can have them take a test, if you're trying to hire a software engineer, you can have them do a little programming or scripting. It's that second piece, the cultural fit that's more difficult for us to evaluate. I mean, let's be honest, if you ask the job candidate well, sometimes there may be some long hours. Do you have any issues with that? No one's going to say no, right? Well, no, I love that, job's report and got to get the work done, I'll work 18 hours a day, seven days a week, right? So it's tough to evaluate. What we've actually found is AI's actually really good at doing that cultural fit. So this is actually work I started back in 2014, ironically, with a law firm.
And AI actually has the ability to look and analyze the corporate culture and the work style and even each team, each division, but even each team level and actually do the same type of analysis on each job candidate. And so you can actually see how well they actually fit to begin with. And then you can look at those candidates and check their qualifications. And so it's actually now inverting the recruiting process. This has been taken a step further now where you have other companies doing the same thing and they're working now on the whole gamification. Where now some companies are requiring you to, before you can even apply for the job, you have to go through this series of games run by an AI, and it's all dynamically generated. So it's not the same each time, and it's not the same for each person, but the intent of the game was just to get a feel for again, who you are, your work style, your personality, how you might mesh in to see if you'd be a good fit or not for this particular group.
And this particular corporate culture, that element. The third is actually around the what's being called the employee experience now. Where now they're using AI to help managers not just evaluate the skillset and the competency of their employees, but also actually look for the opportunities where they can improve the experience for the employees. So get rid of some of those small trouble spots, create higher morale, more productivity, but also work with each individual employee based on their strengths and weaknesses to identify the most optimal career path that the employee would actually be interested in and then help identify opportunities as they start popping up. So that that employee can actually get the knowledge, get the competency to achieve that next level on the career rung. That's something that's actually really hard to do, and most managers, I myself, am guilty of this, it's tough to juggle that, right?
Because sometimes the opportunities just pop up and you may not remember so and so wants to do this or maybe haven't even spoken about it to do that linkage. And I've known people that, I want to be a manager one day and it's like, well you got to get some budgeting experience. They may wait a year or two before an opportunity even pops up. There's other smaller opportunities that can always be done and AI is kind of helping to tap into that. So that, that person gets that experience a little bit quicker. So that's been a real boon and actually retaining employees, getting higher level of productivity, but also establishing a better succession path for all roles within the company.
Guy Nadivi: Other than ROI, is there a single metric that best captures the impacts of artificial intelligence on business and IT operations?
Neil Sahota: Huh, that's a interesting question.
And I'm going to give... I'll be honest. I know it's probably a bit of a loaded answer. But there's one metric, it's accuracy. And I say it's loaded because that actually obviously ties back to a bunch of other stuff like ROI and speed and all that. But accuracy is really important when it comes to AI. It goes back to that truth and trust in the technology we were talking about earlier. And of course we want the AI to be right. We should also manage our expectations that AI will never be a hundred percent correct. Remember as human teachers and we're not perfect either, but I think the accuracy is really the key thing. So going back to the 20% misdiagnosis in healthcare, we know that if we can reduce that number, right, that limits obviously repeat visits to the doctor, hopefully reduces the amount of stress and severity of illness because it gets caught earlier.
There's a whole bunch of things, obviously just tied by being more accurate. And I think that's the one thing you have to really think about, that we as people, how well of a job are we doing? And are there some tasks that we should be offloading to an AI system to actually improve that. One example I'll share is you think about actually IT security. So there's lots of threats, there's lots of bad actors out there, and there's always new threats emerging as a result. But think about how many threats IT security person actually has to try and monitor for.
I know there's tools and reports, but they're literally looking for over potentially two million things that could occur! Plus all these new things that no one has thought of before. That's a lot for a human being to juggle and the longer it takes us to catch it and correctly to identify the type of attack, well, the more damage that gets done. And today there's a lot actually AI systems now that are actually cyber security warriors, that they can actually track two million threats in real-time, all the time and do it with a much higher accuracy rate than we've seen people that's been benchmarked. And so you can just think of what's the advantages or benefits from actually having an improved quicker response time in IT security, you can start to see why it's such an important metric.
Guy Nadivi: PWC has estimated that by 2030, less than a decade away, China will be the biggest AI market accounting for 26% of global AI market share. What can and should the US do to ensure it doesn't fall so far behind China on AI, that it can never catch up?
Neil Sahota: Well, I'm going to be a bit controversial here, Guy. So I will actually say that I respectfully disagree with PWC, because I think China's already ahead. I've seen it myself, China has built something called AI Town. I was there when they broke ground on it back in 2014. They're already cranking out thousands of AI companies, they're actually already using a lot of these solutions. We've seen some of it, what they've done in the pandemic and with their smiles program, they're cranking out 600,000 data scientists a year from their universities. This is not lip service. China recognized the power of artificial intelligence probably shortly after we did the Watson Jeopardy Challenge and they made the investment and more importantly the commitment to it. So in all honesty, I think that China is ahead of us. In fact, I would suspect there might be a couple other countries ahead of us as well.
It doesn't mean that we as the United States can't catch up. And I think the two full thing that we need to do, one is obviously building the competencies and the skills. That doesn't mean we need to turn everyone into a machine learning programmer or data scientist, but it does mean that I think each person needs to know the foundational capabilities of AI. You don't need to know how the hammer gets built, but you need to know how to use the hammer so to speak. And unless we learn how to do that, we can never take advantage of the real capabilities and then actually unlock the real potential with artificial intelligence. So everyone, and I'm pretty confident when I say this in that, probably within the next 10 years, virtually everything we do will have some AI component to it. So having that skillset for everybody, including the business folks, absolutely critical.
The second is we have to make the commitment to make this happen. So that's enterprises, private government agencies, established companies, small companies, teachers, students, parents, you have to make a commitment to actually want to make this happen. China built literally a whole city, made the investment to do that. Saudi Arabia has something called project NEOM going on, where they're building their own future tech city and developing all these skills and competencies and capabilities. And they've already sank half a trillion dollars into it. Canada has done the same thing in Montreal and now Toronto. We have really yet to do something like that in the United States. It's really on us to make the investment and commitment to actually build this kind of AI ecosystem and development area so that we can actually unlock these innovations, and we actually take advantage and be the drivers of these solutions, rather just the consumers.
Guy Nadivi: Neil, since you're looking ahead down the road, what do you think will be some of the biggest disruptions we'll see in artificial intelligence over the next one to three years?
Neil Sahota: That's a great question, Guy. And I thank you for not saying 10 years, that's always a tough question to answer. But at least in the near future here, like one to three years. One big disruption is going to be artificial empathy. It's the fastest growing area of AI right now, and essentially, even though machines don't feel our emotions, they can actually recognize them in people. In fact, we've seen they're very good at understanding emotional state of a person and dynamically change how they interact with you based upon that. So you start thinking about your chat bots or your AI concierge on your mobile app, or even a tool to help with therapists or a psychologist. Having that kind of connection really enhances the experience. And I think we've gotten to a point now where at least the younger generations have comfort and prefer some of this interaction with a machine than a human being, that the more natural flow is going to become a must have rather than a nice to have.
So you're already seeing Citigroup is already incorporating artificial empathy, through neurolinguistics and psychographic profiling into their AI concierge in their mobile app. You're already seeing companies like, Cyrano.ai that's developed tool sets like chat bots to help people with mental health issues so that they feel like there's some sense of connectedness, not meaning to replace a therapist, but creating a safe space that if someone has an episode or something like that, feel like there's a safe space that can actually share what's going on and feel that connection like that person at least empathizes with them. So I think you're going to see a huge transformation and a real huge disruption with artificial empathy.
Guy Nadivi: Neil, for the CIOs, CTOs and other IT executives listening in, what is the one big must have piece of advice you'd like them to take away from our discussion with regards to artificial intelligence at their organizations.
Neil Sahota: That's a good question. I'll say that other than reading my book, which is definitely useful, one thing I would really advise them is think beyond automation. There is value in using AI to automate, but you're really only tapping into 20-30% of what you could do. You have to think beyond automation to get to the real innovation. We have a whole new tool set, we have new capabilities, we know that machines think differently. You got to use that, there are different ways to do the work that we're doing today, that might be far more effective, far more accurate than we are. And I've seen this in the work we've done with self-driving cars. You know the first-generation self-driving car cars relied on camera data because we drive with our eyes, right? And then, there's an episode, I won't name the company, but the guy had his car in self-drive mode and he was busy watching Harry Potter.
Didn't see this truck wiped out on the highway. The truck bed was blocking the road. And even though you could see it from about 300 yards away, the self-driving system never picked it up. Ran right through it, it ripped the top of the car off. Of course, everyone was like, how did this happen? Well, the truck bed was grayish white, it was a cloudy day, so from the camera's perspective, totally blended in the background. And of course, looking at this like, man, if we were using radar or even LIDAR, this never would have happened and it's like, wait a second. Why aren't we using that? Machines can process LIDAR information. They can process GPS information, right? Auditory, we know you can hear that little kid run across the street before you see the kid, right? IOT sensors in the car, other cars on the road.
I mean, that's what happens today now. These self-driving vehicles are processing over a thousand data points per second in real time, processing information that no human can actually process, which is why by working in the United Nations, we don't talk about when do we legalize self-driving vehicles. We actually talk about when do we ban human drivers? Because now, human drivers actually inject more variability into the system. So you talk about the accuracy rate. Well, hate to say it, machines are better drivers. I love to drive, but they're actually better drivers. And so don't just think about automation, right? If you're a IT leader, like I said, there's value in some of that, but the real value, the real innovation, the real disruption's going to rely on finding those different ways of doing the work, tap into that new capability that AI gives you to do that.
Guy Nadivi: All right. Looks like that's all the time we have for on this episode of Intelligent Automation Radio. Neil, I think at this point, everyone knows how profoundly disrupting AI's impact is going to be on our world, but this is the first time we've had a guest on with a framework for how to think about that disruption and harness it for good outcomes. Thanks for coming onto the show today and sharing your expertise with us.
Neil Sahota: Hey, thanks for having me on Guy. Had a blast!
Guy Nadivi: Neil Sahota, AI advisor to the United Nations, IBM Master Inventor, and Chief Innovation Officer at UC Irvine. Thank you for listening everyone. And remember don't hesitate, automate.