Speaker 1 (00:04):
This is the data science conversations podcast with Damian Deighan and Dr. Phillip Diesinger. We feature cutting edge data science and AI research from the world’s leading academic minds and industry practitioners. So you can expand your knowledge and grow your career. This podcast is sponsored by Data Science Talent, the data science recruitment experts.
New Speaker (00:32):
Welcome to the data science conversations podcast. My name is Damien Deighan, and I’m here with my cohost Dr. Philip Diesinger. Today we are featuring the California based XPRIZE Foundation, and we’ll be discussing their amazing work in applying artificial intelligence to solving the world’s most pressing social problems. Our very special guest today is Neama Dadkhahnikoo, who is the Director of AI Operations at XPRIZE.
His undergraduate degree was undertaken at UC Berkeley where he studied Applied Maths and Computer Science with a Minor in Physics. He also holds a Masters in Project Management and an MBA. Neama began his career in 2005 as a Software Developer at Boeing. In more recent years, he has founded companies in consulting and in healthcare, including an NLP startup called Techspert, which focused in the area of mental health. Prior to his current role, he was the CTO at Caregivers Direct, a health analytics company specializing in the home care market. Then towards the end of 2018, he joined the XPRIZE Foundation. Neama, we’re so delighted to have you on the podcast, so thank you so much for joining us today.
Speaker 2 (01:44):
Thank you for having me and I’m very excited to be here and talk about myself and XPRIZE and our winners as well.
Speaker 1 (01:50):
So Naema, perhaps you can start with your personal journey. How did you end up in the AI arena?
Speaker 2 (01:57):
It’s interesting, because I never meant to be in AI. My early career was in software development at Boeing. I was there for about 7 years working on interesting stuff, but the problem at working at a large company like Boeing is that if you’re not in the exciting parts, like on the factory floor building those airplanes, which is fantastic, but I was in the tools group, the software tools group, essentially building software to support the internal parts of Boeing. So you’re off the beaten path, if you will. And the problem there is, when you’re in a big organisation, it sometimes feels like you don’t have that impact. So I was there for 7 years, software developer, and I was like, you know what, I want something different, so I went and got my MBA at UCLA and as part of that, I kind of repositioned my career.
I was like, I’m going to go into startups. Startups equals impact. You know, your day-to-day work actually really drives the failure or success of your organisation or your company or startup. And as part of that, I was kind of tackling different problems. I tried Cleantech, it took too long, like it’s a 5 – 7 year cycle and that didn’t appeal to me. I tried startups in healthcare, in apps and different things like that. On the side, as you mentioned, I was doing consulting, essentially startup consulting to pay the bills. So one of the startups that I founded that we were getting some good traction with was something called Techspert, and employs AI for mental health. But it didn’t start out as AI for mental health, but it actually started out as a service where people could get advice and, in particular, advice about relationships and dating and you know, what to text and what to do and so forth and so on.
So we have people giving advice to other people. And first you build the software and you build this technology and this platform, this solution, and you expect people to use it in a specific way. Lo and behold, people were using it in different way. We thought it was gonna be a lot of guys asking for relationship advice, turns out it was a lot of younger women asking for specific advice around like how to deal with their friends, their family, their relationships. It was interesting, so it was more about mental health than it was about relationships and we were like, okay, that’s interesting. But at the same time, because this was a user to user platform, a lot of the advice seekers were there, but the advice givers would get burned out because these questions were very similar and it was also like heavy topics that you’re dealing with.
But as we’re doing that, and this is around 2016/2017, it’s like we have a lot of data. We’re gathering all this data. Essentially, we’re building this data set of questions and answers. This is something that artificial intelligence could actually be trained on and it’s good at doing right. These kinds of repetitive actions and repetitive questions. If you have a good data set, you can actually use AI to solve that. So I had a background in math, of course, you know, I’d studied linear algebra and all that and never used it, but I was like, wait a minute. This is the perfect opportunity, so I kind of crash coursed natural language processing NLP. This was the second time it was ever taught at Stanford and all their courses were online. So I literally took the second NLP class ever taught at Stanford online.
So I essentially became an NLP expert, you know, over a few months. AI was so new and NLP was so new that I could talk with professors at UCLA and USC about NLP and these technologies that we’re building and researching and I could hold my own, which is kind of crazy when you think about it, right. If you’re talking about a physics professor or some other topic that’s been around for a hundred years, and you’re trying to talk about that, you know, with an expert, you’ll never keep up. Right. But the NLP was so novel, it was so new. These technologies were so recently invented. I’d be like, oh yeah, I saw that paper too and we actually just used that technology or this algorithm that they invented in our system to see how it works. So it was very interesting. So that’s kind of how I got into AI.
And then after Techspert, I of course worked at Caregivers Direct and this was a similar area where it was home care and we were trying to build this AI system to improve the quality of home care. The interesting thing there again was, and this is the common theme that we’ll see with all of this, data is king. And with that startup, what I realised was that the data that we needed to build these systems was years away, home care has no data. And this is an interesting problem if you ever come across it, and data would be so important to solving these problems. Like just as an example, your interactions with people is a better predictor of your mortality when you’re older than even sometimes your health record. So if you had that data that you could actually be like, wait a minute, this person might be in danger of, you know, passing away because they’re just so isolated, but that data doesn’t exist.
But the problem for me was that that data was years away from being collected. You have to build these data sets from scratch. So after Caregivers Direct, I was kind of looking for something different and that’s how I ended up at XPRIZE. They essentially called me up and they were like, Hey, we’re looking for someone who knows AI, but also knows business. Do you know anyone? And I was like Me, maybe? So that’s kind of how I ended up at XPRIZE in 2018, but you can kind of see this winding path to where I ended up, it wasn’t planned at all. But all good things come to those who experiment and innovate and are open to opportunities and are willing to help and then to learn.
Speaker 1 (06:48):
Neama, can you give us an overview of the history of XPRIZE and maybe of this type of competition generally? And then maybe we can look at the work The Foundation has been doing since its inception in 1994.
Speaker 2 (07:02):
So the concept of a challenge competition or a prize competition is not something that was invented recently. This is actually, historically, something that’s been around for many, many years. In fact, the most famous one is the Longitude Prize. So latitude, apparently it’s very easy to figure out through the stars or whatever, but longitude is much more difficult to figure out on ships and you need to know latitude and longitude to know where you’re going in the ocean. So the British Monarchy essentially put up this prize. I think it was £10,000 at the time, which was a lot more back then, you know, in 1600 and they basically challenged people to solve this problem. And of course, all the scientists are like, we got this, we got this, we’ll figure it out. It turns out it was a clockmaker. A clockmaker figured out how to determine longitude and he worked on this problem for 50 years and solved it. And some people are good and that’s why the British Navy was so successful in navigating the world is because they had this single proprietary technology of knowing exactly where they are at all times. But this is how this came about. But this concept of the challenge prize this has been around for many years, everything from like margarine to cars, to the Ortega Prize, which was around traveling across the Atlantic, right. That’s where Charles Lumberg became famous, winning this prize. A French businessman put up the prize money. They wanted to drive air travel, drive interest and innovation in air travel and Charles Lindbergh was the one who was successful to be able to claim that. That concept is what Peter Diamandis took with XPRIZE in 1994. Actually with Charles Lindbergh’s son or grandson, they came together under the arches in St. Louis and were like, look, we’re going to do for space travel what the Ortega Prize did for air travel. And this prize came about. Now, the crazy idea here was that Peter Diamandis didn’t have the money. And soon enough, he was able to get, the Ansari family to put up $10 million for the prize, that’s where it started. And in 2004, so 10 years later, the team that became Virgin Galactic was the team that won that prize. They went to space. They essentially were able to travel to space twice within a two week period, and they won the $10 million prize. And as part of that, you kind of had this proliferation in space travel and private space travel. That’s where you kind of get these Space X’s of the world and Virgin Galactic’s of the world because someone was like, look, you have to solve this. Here’s a bunch of money, go figure it out. And as part of trying to win that prize money, you create an industry. So that’s XPRIZE in a nutshell, and then where XPRIZE came from. Since then we’ve had over 25 prizes and awarded $260 million or so. But some of the more interesting prizes we’ve done over the past two plus decades, the Google Lunar Prize, which was about a Lander to be able to land on the moon. That’s an interesting prize because no one actually won it. Many teams came close, but I believe an Israeli team crash landed on the moon and we gave them $1 million because it was close enough. But landing on the moon, a very difficult problem, no one was able to solve it within the timeframe, but it created an industry that now has over $400 million of investment in it and there are multiple companies working on this lunar economy. We have prizes around mapping the ocean floor, educating children in Sub-Saharan Africa. And more recently the Elon Musk prize, which is our largest prize to date, $100 million to be able to capture carbon at the gigaton scale. The premise there, of course, being that look, you have to be able to fight climate change from multiple directions. It’s not just a stop producing carbon problem, but it’s, maybe we can capture carbon problem as well. But there are many prizes in many fields and it’s all essentially put a very big carrot at the end of the road to solve a grand challenge that’s facing humanity.
Speaker 3 (10:40):
Can you talk a little bit more about this aspect of stimulating investments? That’s something that surprised me a lot with these challenge prizes. So you put out a prize, say €10 million for something, but then in the process of teams pursuing this, they double or triple the investment, right? And that creates this kind of industry. Do you have expectations when you set the prize or when you set the challenge, like how do you go through that process?
Speaker 2 (11:05):
The concept of a challenge prize is philanthropy with capitalism thrown in essentially, which is probably why it appeals to all the billionaires. But the concept here is basically, if I put up $10 million or €10 million, in the process of solving this problem, the teams will attract investors because the teams will be building some technology that is investment worthy, of course. On average, that drives a 10 X investment versus the amount that the prize was put up. So if you put up a $10 million prize, it drives up to a $100 million of investor money into the startups working on that technology. So it’s an interesting concept for philanthropy, because again, there’s this multiplication aspect to it, kind of. The teams are aware of this and we tell them this from the start, this is not a grant process. There are many organisations that give grants so it’s essentially the kickstart model, right? Like I’ll give you the money to go solve a problem. We’re the carrot at the end of the road model and actually we work with organisations, you know, like the Gates Foundation or, you know, Google, you know, their philanthropy arm. Whereas some of our teams actually get money from those organisations as part of their investment strategy. But our model basically says, look, big money at the end of the road, we will support you with advice and partnerships and things like that. But the investment concept, you have to be an entrepreneur, you have to be a salesperson. You have to go out there and be able to raise those funds and draw that money in. And if you’re not able to do that, then how will you be able to solve this grand challenge that has vexed humanity for, you know, hundreds or thousands of years?
Speaker 1 (12:34):
You mentioned earlier the story of the clockmaker who came from nowhere to win, possibly the original challenge prize, is that something you’ve seen repeated in the modern day with XPRIZE with people from way left field winning your prizes or doing very well?
Speaker 2 (12:50):
So the concept of the XPRIZE is that anyone from anywhere can win this as long as they come up with a good idea and execute on it. And we’ve seen this happen occasionally. Now to be completely fair, you know, when you’re talking about very deep technical prizes, it’s unlikely that anyone from anywhere can solve it. You often times get the best and the brightest and people with, you know, graduates from MIT or, you know, they are the experts in the rainforest and so forth and so on. But on occasion we have had teams make it to the finals and even win money where you would not expect it. The best example of this is actually the Shell Ocean Discovery Prize. The challenge was to be able to map the ocean floor in a way that’s never been done before and with the speed and rapidity that’s never been done before. And you expect the teams to be able to kind of do this concept are the ocean explorers and engineers that are able to do this. And sure enough, the winning team was this joint venture between a Japanese and UK two startups. And they actually are innovators in the field and they’re best of the best. But one of the teams that made the top five was a high school team. So it was a high school teacher and he actually loved XPRIZE and got his class and to work on these mini submarines and the team was able to make it to the finals and they were able to, you know, compete. They didn’t win the competition, but the judges were so impressed with their work they gave them, I believe, a $1 million inspirational prize. They made a documentary about the team. And so all of these high schoolers were able to kind of go to the documentary premiere. It was surprising, but it wasn’t impossible. And this, this is the concept, right? If you, again, these are judged on results so if you can come together, anyone from anywhere can actually do the work, you will win the prize.
Speaker 3 (14:31):
So Neama, you talked about basically the challenges that you set up and what kind of effects they have. What is motivating the XPRIZE in the first place and then also, how do you decide on which topic to put out there?
Speaker 2 (14:43):
The ethos of XPRIZE is this kind of global positive future movement. What that means is basically there is a positive future that exists. We just have to bring that future forward. So a great example with our XPRIZE is that private space travel is possible. It’ll take 50 years to come to be, or if we incentivize people to bring that future forward, it’ll take 20 years. And we’ve seen the results of that now. So with XPRIZE, we are tackling the world’s grandest challenges, these problems that haven’t been solved, either through lack of technology or lack of market interest. And we essentially incentivize people to go solve them because that will benefit humanity. So whether that is mapping the rainforest so that we know what the biodiversity is in there, what the value is there, so we stop destroying the rainforest. Or whether that is being able to show that you can use technology to rapidly educate children, even in a village in Tanzania, because that technology can then be used to educate millions of children all around the world. Or whether that’s even shorter term problems like with the pandemic response challenge, where we want to show that you can use data to alleviate the effects of the pandemic and cities and governments can actually use that information and we show that again with that competition. The concept here is (1) incentivizing to solve a grand challenge and (2) show them that it’s possible; and then (3) show that technology and humanity can solve these grand challenges and can do it at a much more rapid pace than say taking 50 or 100 years to be able to solve those problems.
And then finally, in terms of how we choose our problems, there are many different ways of doing this. Most often they come through a concept called visioneering where the best and brightest of the world can come together and talk and debate and argue around what kind of problems that need to be solved. What are the biggest issues? They kind of pitch against one another and suggest these, and then we go and fundraise and get those. So some of the prizes in the pipeline right now that are looking to be funded are around wildfire suppression, that’s a personal favorite of mine. I live in Southern California and I’ve had multiple wildfires where I’ve sat at my door and liked watch the hills burning off in the distance. This is around the concept where you can detect those wildfires and put them out as quickly as possible. There is frontline health, that’s another prize that we’re about to get funded hopefully. That’s around using AI technologies to be able to help those people who bring healthcare in the communities. Again, think of a village in Africa, where they have these nurses who kind of go out there and then try and bring healthcare to the people that don’t have health care. How can we use technology and AI to rapidly innovate in that space and drive forward those outcomes. We have other prizes in the works as well and we’ve recently launched prizes around feed the next billion, which is around cell-based agriculture. The Carbon Prizes I mentioned, which is a $100 million Elon Musk prize, which is around carbon capture. So you can kind of see it’s a very diverse space of everything from space travel to the environment, education, healthcare, multiple domains that are being tackled across the globe.
Speaker 3 (17:37):
And so how do we, you talked about this, the pitching process basically where different teams come in and they basically talk about challenges and so on and they pitch against each other. How does it lead to the challenge that you choose in the end? Like what are the criteria and who decides on that?
Speaker 2 (17:51):
So we have what we like to call a conclave of experts. So we basically have these brain trusts if you will, in different domains, right? If you’re talking to health care, if we’re talking environment so forth and so on, and they kind of look at these problems. Not every problem can be solved with an XPRIZE. So an XPRIZE has to be something where it’s a grand challenge. It’s a serious problem. It’s something that can be solved through incentivization and prizes. So often times social problems can’t be solved in this concept, right? Like housing, that’s a really big issue and a problem, but you can’t put up like a $1million prize to solve the regulatory issues, right. That doesn’t make any sense. So it has to be a problem that’s XPRIZEsable, if you will. And then also, often times it is a problem where there is a market inefficiency or a market doesn’t exist.
So you don’t want to solve a problem where everyone’s already trying to solve it because people can make a ton of money. You want to put a highlight on something, for example, on wildfires. Like wildfires, if you actually had technology that could prevent wildfires, we would save billions of dollars in insurance money every year, right? And save of course, hundreds of lives and billions of dollars of pain and suffering. But the market isn’t addressing that right now. There is no incentive mechanism there for someone to go and invent this technology and build it. So we want to kind of put that flashlight on that problem to have it solved. There is also a voting process as well around that. Again, you have these experts kind of tell the audience, and this is the first time we’re doing it with this global visionary concept where you’re going to have people from all around the world, vote on what problems to solve next. So you’re going to have these experts kind of educate the crowd and then let the crowd decide what is the problem to be solved next. We also have prizes occasionally where someone just comes in and approaches us and they’re like, look, coral restoration. We have, you know, $20 million, this is a problem we want to solve. We go and look at the science and technology. And we’re like, yes, we agree. And then we just launch that prize. So it depends, different prizes come from different places.
Speaker 1 (19:38):
Is there anything happening Neama regarding, I guess one of our biggest problems is moving to a clean way of generating our energy away from fossil fuels, is anything happening there?
Speaker 2 (19:50):
It’s an interesting concept and we are exploring some energy prizes. I’m not personally involved in those so I wouldn’t be able to speak to them specifically, but I do know that the energy domain is something that we’re looking at. We’re looking at the downside of course, with the carbon capturing carbon removal prizes that we’ve talked about. So essentially like cleaning up for the energy, if you will. It is interesting though, because just off the top of my head, I think for example, solar power, right? That’s something that is already being solved by the market. So that’s something where you don’t really need to incentivize it, but there are possibly areas where maybe there are technologies that are not being explored. I know geothermal is definitely an interesting area or maybe space based initiatives around energy and things like that. The short answer is yes, we’re definitely looking at the area, but I don’t think we have anything too close in the pipeline at the moment in those spaces.
Speaker 1 (20:36):
And moving on to focus on AI are most of your prizes that are currently running, do they have an AI focus or is that not necessarily the case?
Speaker 2 (20:47):
Yep, so we got involved with AI for the first time through the AI XPRIZE. The IBM Watson AI XPRIZE, which was a $5 million challenge for teams to use AI for good. It was kicked off in 2016. So this was, again, this was early, this concept of AI for good didn’t exist back then and this prize was essentially a kickstart for that concept. Now, of course, AI is everywhere, right? Responsible AI, ethical AI, AI for good, and you have different companies, Google, Microsoft, Facebook, so forth and so on. All of those are tackling AI for good, in different ways. But again, this didn’t exist back then. So that was the kickoff. That was the impetus. IBM Watson stepped up and said, look, we want to start this process. And it led to some really great outcomes. You know, we had over 150 teams apply for this competition from, I believe it was 30 countries from around the world.
They tackle different problems in healthcare, in education, in the developing world, many different and diverse areas. Some even looked at like the ethics of AI and robots and things like that. So it was really cool and really interesting and lots of great outcomes. But as you mentioned, now that AI is mainstream, there is no need for an AI Prize, but AI and data is present in many of our prizes. The Rainforest Prize, which is going on now, there are going to be drones, there are going to be satellite imagery. There’s going to be data that is being driven there that has a huge AI component to it. You look at a prize like the Wildfire Prize again, how are you going to solve wildfires, probably going to be this concept of satellites and drones and some people were even talking about like using rockets to shoot the fire, but the point is, there’s going to be AI. There’s gonna be computer vision and there’s gonna be satellites. There’s going to be data. So you can kind of see here that we had an AI Prize to kind of kickstart this concept. And now AI is embedded in multiple prizes and will be embedded in multiple prizes moving forward.
Some of the more interesting prizes that I’m excited about are around space. So there is like for example, a Space Debris Prize that’s being discussed. And if you don’t know about this, space debris is one of the biggest problems we have and is only getting worse year over year, because there’s more debris being spread out into space every time we launch a satellite. They are currently being tracked manually, essentially they’re at NASA and European space agency and you know, the Russian space agency. They just track these pieces of trash essentially in space. But it’s getting to the point where it’s going to be impossible to track by hand, you’re going to need AI to be able to track them and then hopefully solve the problem maybe by having these automated robotic drones going up there and essentially vacuum cleaning space. So you can kind of see how AI is embedded in many of our prizes as we move forward and will be embedded as we move forward and tackle these different problems.
Speaker 3 (23:18):
Yeah. I think the ISS is constantly trying to avoid hitting the garbage there. So they constantly manoeuvre a lot which I found quite interesting to learn. Let’s say you put out this big challenge on a certain topic. You have a lot of interest as teams coming in, investors coming in to work on it and so on. You have a winner in the end, you solve the problem. How do you like follow up with that? Do you measure basically what the impact was like a year later, two years later? Do you have specific goals that you set that go beyond solving the actual problem, but maybe stimulating both in the sector or something like that?
Speaker 2 (23:54):
After prize impact is an important piece to any prize? Because I mean, historically we didn’t have after prize impact and it was like, congrats, you won the prize, now you’re on your own. That doesn’t really help drive innovation. It’s more productive if you encourage and incentivize and build the support ecosystem. So now we have an alumni network, we have after prize impact as a specific phase of the prize. So for example, the AIX Prize that was recently awarded, we’re doing things around research to kind of show what the improvements were and how we judged them and then what the results were. There is kind of like a road show, if you will, where we actually go to conferences with academic and business to showcase the work, but the teams kind of talk and show what they did. We even do investor summits and things like that, where it’s like, look, you know, this team won the prize, investor you probably want to invest in this organisation. So there’s a lot of after prize impact work that’s done in that way. But to your point around kind of how to measure impact, that’s a very important thing to do because as a sponsor, you want to be able to see, and as XPRIZE we want to kind of see how successful were we, where was the impact stronger? What did we do good and what did we do bad? Some of the ways we look at impact are in terms of dollars, for example, the Google Lunar Lander Prize, where we can actually look at the teams that were there and see how much money they’ve raised. And we can look at the industry because the industry did not exist, this private space travel to the moon and doing work on the moon. You can actually go and see the contracts that NASA has awarded to these companies, it’s about $400 million, so that’s a tangible way for you to be able to measure that. Other ways of doing that is looking at the outcomes. So for example, the Global Learning Prize, you can kind of go and see the outcomes of these village kids who are educated and see how that’s done. And then you can look at the uptake of the technology, so these apps that were created, this technology that was created. How many other organizations are using it? We can even do post prize content capture. For example, there was a Tricorder Prize, which was literally like invent this concept of a tricorder from Star Trek, but bring it to the real world. And, you know, a team won that. But actually one of the, I believe is the second or third place team, was able to take their technology and deploy it in Mozambique, actually. So we were able to go to Maputo and see this device that actually was able to hear you cough and diagnose your affliction. You know, whether it’s like, you know, tuberculosis or something similar to that. So this team, we were able to go there and capture that content and promote their work so that other organizations can bring in the supplies and bring in the support and help them proliferate their technology. So there are many different ways for us to be able to support these teams after the prizes as well.
Speaker 1 (26:31):
If we then turn to the IBM Watson XPRIZE, Neama, can you give us an overview of the competition and the three finalists in particular?
Speaker 2 (26:40):
Absolutely. So, as I mentioned, the IBM Watson AI XPRIZE was a $5 million challenge that asked teams from all around the world to pick a grand challenge, to pick a problem, and then to solve this with AI. The implicit concept here was AI for good. Back then in 2016, it was always AI for bad, right? It’s AI is coming to take our jobs, the robots are coming to kill us. We wanted to change the dialogue and change this concept of like, look, AI can be used for positive. It can be used in ways that we are not even thinking about. And sure enough, that was true. We had 150 teams from all around the world, 30 countries sign up in that first year. And you know, year after year, we had our judges kind of evaluate the teams as they moved forward. It was interesting because this is a unique XPRIZE, most XPRIZES, if not all of them, have a specific goal – map the ocean floor, land on the moon, map the biodiversity of the rainforest, educate children. This one was turned on its head. So it was basically like use AI, that’s your only criteria, now go solve a problem. That makes it a lot more difficult to judge so we actually had one of the largest judging panels. We had over 30 judges, and then we also had another 30 red judges as we call them. They were like mentors who actually work with the teams to evaluate them. And the reason was we had teams solving of diversity of problems. So everything from using AI to improve beekeeping, to detect opioid relapse, sex trafficking, malaria prevention, depression, education for refugee children, AI ethics, robotics and ethics, and things like that. We had teams tackling everything and anything, and it was really fun and really interesting to kind of see them go through these problems, solving problems around the world in different domains. It was very difficult to judge, but by the end of it, we have three teams.
Aifred Health from Montreal, they were using AI to improve the treatment of depression. So they were essentially medical doctors and AI engineers. And they had invented this technology based on medical records and the way we were treating depression to help the doctors to be able to improve their treatment. And this is important, this concept to understand,. It’s not that the AI is replacing the doctor. It is the AI improving and enhancing the doctor’s treatment. You don’t want the AI system kind of willy nilly deciding on its own, that’s where you get bad outcomes. What you want is this kind of human AI collaboration. AI is really good at noticing patterns at going through large pieces of data. The humans are really good at intuition and things like that. You merge them together and you get positive outcomes. Aifred Health, they successfully raised a series A – this goes back to this concept of raising money. I believe in Q3 of this year, they’re going to be able to do their clinical trials. They’re at the cusp of proving that they’ve done their preliminary studies and now they’re going to be doing their phase one and phase two for that technology.
Our second team is Marinus Analytics. They’re based in Pittsburgh. They are a female founded team that actually uses facial recognition technology and other AI driven concepts to improve policing, in particular around sex trafficking. They were founded in 2014 if I remember correctly. So they’ve been around for a few years and they have actually multiple paying customers in the nonprofit and the police space. Think about your local police department, how AI savvy are they? Probably not that well, but these researchers found, look, we can use facial recognition plus data, for example, from places like Backpage press and things like that, to actually identify people who are being sex trafficked, who are being abused and use that to essentially find those criminals and to save these people. And they have been able to use their technology in thousands of cases and be able to save, you know, hundreds of victims of sex trafficking. And now they’re actually expanding into different areas. They’re looking at child protective services. They are looking at other domains where there are people who are vulnerable, who can benefit from this usage of AI.
And then our final team is ZzappMalaria and they are based in Israel, Tel Aviv. They are using AI to improve the treatment of malaria. So malaria is an interesting thing. It’s solvable. We have eradicated malaria in the United States, in many parts of South America, but we haven’t solved this problem in Africa and it actually causes hundreds of thousands of deaths every year. And that’s kinda crazy to think about, right? It’s a solvable problem that’s causing untold human suffering. What ZzappMalaria does is they have an app that allows the field workers to kind of go into the field, map water bodies. Malaria comes from mosquitoes, mosquitoes breed in stagnant water. So if you know where the water is in these villages, and you can kind of track it and take the field worker to those puddles, you can drop treatments. There’s these kind of like pills, essentially that poison the water so the malaria can’t spread. Then you can also use things like spraying and nets to drive this kind of malaria prevention. Now, the important thing to understand again, for us in the Western world we don’t understand this, but in these places resources are limited. So if they had infinite dollars, they could be able to, you know, send infinite amount of workers out there, spray everywhere, treat every puddle of water, you know, put up mosquito nets everywhere. And they would be done with malaria, you know, in a few years, but they have limited resources and these governments, whether it’s in Ethiopia, whether it’s Sao Tome, whether it’s Mozambique, you know, all around Africa, they have to decide essentially, how do I allocate my resources? That’s kind of where this app really comes into play. You have this app, you have AI, you have weather data, you kind of map where you suspect the most bang for your buck is, and you can drive those workers to kind of go and do that. So that’s where ZzappMalaria comes into play and that’s where they’re doing, they have multiple projects across Africa, with many different countries. Their projects are affecting over 500,000 people currently. And right now their biggest project is in Sao Toma, which is this island off Africa. And they want to eradicate malaria from the island completely.
Speaker 3 (32:34):
You basically already said at the beginning, for this kind of challenge to flip the concept on its head .You say, use a tool which was AI to solve a problem. How did you define what you mean was AI there is like vastly different approaches basically to the center and it’s a very wide field?
Speaker 2 (32:53):
That’s a great question and it was a question that our judges asked as well, right? What is AI? And we took a maximalist approach. In general, XPRIZE is technology agnostic. We don’t care how you solve the problem, we just want you to solve the problem. In this specific case, we did care a little bit, right? It’s like it has to be AI, but our definition of AI was very broad. We had tech teams using way different technologies. Some are using the standard, you know, CNNs and RNNs. Again, neural networks, very, very classical AI. Some of our teams were using evolutionary algorithms, which are out of favor, but they had invented this concept of essentially in that case, they’re using it to determine diabetes and you can’t use normal neural networks to do that. so it was an interesting concept there. We also had teams that were using more basic AI, you know, linear regression and things like that and at some point the judges were like, you know, is this AI or is this just, you know, data science? And that was an interesting debate that the judges had to have. And they also had to weigh the advancement of the AI versus the outcomes. It was an interesting trade-off and hopefully we’ll be able to kind of publish some research on this soon. But teams that picked the hardest problems and showed the most advancement or showed the least advancement,I should say, were able to kind of like tackle hard problems, but they weren’t able to show their results. But then we had some teams that actually picked basic AI concepts and were able to deploy it in the real-world setting with novel data and drive home impact. And you’re like, well, your AI isn’t that impressive, but you just saved, you know, a hundred thousand lives.
So that was an interesting trade off that the judges had to do as well, where it was kind of like impact versus technology versus scalability, right? Like in the future, how can you actually go to market or expand your market or take this global? And then ethics and safety, that was the other criteria that we’re looking for as well. If you’re working with humans, you had to have an IRB. We want to look at these concepts around ethical AI, whether it’s around informed consent, whether it’s our human rights and things like that. So you can kind of see immediately, like how difficult the judging process was and how many different categories you had to kind of like evaluate across and we had some novel, interesting solutions to be able to kind of tackle those problems.
Speaker 3 (34:58):
And did you see, on the other side, did you see an impact on the outcomes of the challenge? Was it like a broader field or even more teams coming to meet the challenge?
Speaker 2 (35:07):
I think the AI XPRIZE was unique in that we probably had the most diverse crowd, if you will. And what I mean by diversity I mean across everything, right? I’m talking everything from gender and ethnicity, but also professions and backgrounds and where they’re coming from, geographies. But that was the nature of the beast, right, because it allowed the teams to pick their problem and then go solve it. When you’re talking about, for example, carbon capture, right? You’re going to probably draw a very specific crowd that is already working on that problem but if you allow them to pick their own problem, then they can diversify and go solve those. I’m not sure it’s recreatable honestly, I think the solution to that problem of like diverse in big prizes. For that, you have to basically be able to pick domains that are globally applicable, or you have to go the other way, where you have niche problems, right? It’s like, look, this problem is going to attack South America. This problem is gonna attack Africa. This problem’s gonna attack North America or whatever it is and then by having a diverse bundle of prizes, then you can kind of cover the whole world as well.
Speaker 1 (36:11):
So Neama look, a lot of people want to use AI for good and get involved in AI for good, but they don’t know how or where they can go to get started. What would you say to that question?
Speaker 2 (36:22):
The most surprising lesson from the AI XPRIZE for me was how AI is not actually being used to solve so many problems. Some of our best teams were the ones who basically took basic AI. So you don’t even need to be a PhD, you can just be someone who has a basic understanding of AI, who knows how to use a few, you know, neural networks, AI 101, if you will. But what they did was they went and found problems that were not being addressed. They looked at datasets and problems that no one else was thinking about. Maybe they’re too small. Maybe they’re off the beaten path. Maybe they’re in different countries and they would use AI to solve those problems. And they would have immediate impacts. That was the most interesting piece. If you go and find those problems because Google and Microsoft are too big to think about them or get around to them, you can drive innovation and impact at an immediate scale. You know, whether it’s in Sub-Saharan Africa, whether it’s even in your own community, right? There are technology deserts in the United States, right? There are parts of our country, if you’re based in the United States or in a Western country that are not getting those benefits and they’re not included in the datasets. That’s the important piece as well there right. They’re not represented in these data sets. So if you can go and find those communities, go get involved with them and work closely with them. That’s the other key, don’t be the outsider who’s like, I’m going to come save you, get involved, partner, work with those organizations and then volunteer. So if you’re like at a big company, a big organization, one of the coolest things I saw was at IBM, some of the researchers were allowed to essentially volunteer and help the teams out. And they had so much fun working with these teams because it was like, look, I’m getting to work on these really cool problems I never get to experience, but those teams exist. Those problems exist. Those nonprofits exist, who would love to have a data scientist take a look at it, right? Who would love to have an engineer, you know, help them out a little bit. And those are the problems that AI, in my opinion, over the next decade or so, will be able to really solve. Let Google tackle self-driving cars, right? Let Microsoft tackle these like big AI problems and let their researchers invent the most cool and novel algorithms. What you should do is take that research, which is out there that they’re publishing their papers, take those innovations, take that cool stuff and take it to the rest of the world. They’re the ones who will benefit the most. And it’s up to you to do that because again, Google will not get around to saving, you know, your local community, but you can. And that’s where I would say people should focus most of their energy on.
Speaker 1 (38:48):
A fantastic message. So where can people find out more about you and the work of XPRIZE Neama? Where should they go?
Speaker 2 (38:55):
So go to XPRIZE.org and XPRIZE on Twitter, on LinkedIn, on Facebook. You can learn more about the AI XPRIZE and many of our other prizes as well and upcoming prizes that will be happening. And you can always find me on LinkedIn and, you know, message me. I’m active on social media and always happy to help someone out who’s willing and interested to use AI for good.
Speaker 1 (39:20):
Please make sure that you do join us for the next episode, because we talked about the AI XPRIZE finalists. We will actually be speaking to the winners and doing a deep dive into how they used AI to solve the winning problem. And they’ve been doing some really amazing work. So for today, I’d just really like to thank Philip and our very special guest Neama. You’ve been amazing, shared some really great insights. It’s been marvellous having you on thank you so much. And we look forward to having you with us on the next podcast. Thank you.