Inside the IBM Watson AI Lab for Cancer | Thomas Anglero (1 of 4)

I'm glad you like the view.

This is your office.

This is the Watson Innovation Lab. This is my office. And I'll tell you the whole story behind it. So please relax, take a seat, take your coat off. It gets very warm with the sun.

So we stay here?

We'll stay here, yeah.

So, uh, are we supposed to sit on the ground?

You can sit on a chair and I'll get more chairs. Okay. You can slide the sofa. Yeah, yeah, yeah, yeah. Wherever you want, doesn't matter. Thank you. I'll get more chairs.

Just grab chairs.

Muppets. Let's go that direction. Blue ocean. More than enough.

It's just enough. It's perfect. And you keep on standing.

Hello.

I hope you don't mind if I take my jacket off.

Is it cool?

Okay. Well, welcome to the IBM— we call this the Watson Innovation Lab. And you are the first ever meeting in this room. So congratulations. Just got the furniture for this room. Just got the furniture for this room last Thursday. And we installed our smartboard late last night. So I mean, this is brand, brand new. I have a PowerPoint presentation I can show you and stuff, but I think one of the things you told me about, you want to talk about innovation and the whole story behind how we did this and all that.

And also I think maybe other people, um, what Watson and all the data and, you know, the artificial intelligence can do. And I mean not only purely for health care, but also maybe for welfare, healthy living, beyond that, because we are more group representing also a lot of people as municipalities, long-term care facilities, etc.

There's no doctors here. One doctor?

No, no, no, no. We had an appointment who was the delegation leader, but it's very difficult when the red is here.

Okay, so let me introduce myself and we'll take it from there. The best way for the— I don't need PowerPoint. I hate PowerPoint. I could talk all day. But the best thing is you guys interrupt me, please, and ask questions, because I'm here to give you information and you're here to get it. Don't wait, just ask. All right? So you get as much out of this as possible. Okay?

Okay.

Does anybody want to see the PowerPoint?

Could be a lead. I'm like a living PowerPoint. Yeah.

Okay. Next slide, please.

I challenge you. What joke? I don't want to know that joke. I said next slide, please. Next slide. Is that type of group? Yeah. Friday afternoon. Okay. Fine. Is this good? Anyway, my name is Thomas Anglero. I am the innovation director for IBM. I started working at IBM one year ago. And before I started working at IBM, I was innovation director for the Norwegian tax authorities for 5 years. So I've started up 3 companies in my lifetime in 3 different countries around the world. I am born and raised in Manhattan, New York City, and 20 years ago I fell in love with a 6-foot blonde Norwegian who you women in the room probably know really well. She said to me, I called her on the phone one day, I was so— I got into my engineering degree, And after I got my engineering degree, I was working and I was going to night school to become a doctor. I finished my pre-med. I called her on the phone. She was living in Norway. I was living in Brooklyn. And I said, I just got a letter in the mail. I got a full scholarship to become a doctor, but I have to move to California. And as a woman, she said— what did she say? What do women love doing? She threw the ultimatum at me.

Yes. I knew she was going to do that.

We are the joke. Worse. Either you move to Norway or you can take your ass to California. You know how women do it? How do you do it in Dutch? Is it all the body language too? Yeah.

Yeah.

It's been not that bad. No, we work more. Ladies, ladies, ladies, quiet. Men answer the question. How bad are they? Very. When you hear men say nothing, it's really bad. We don't dare. You don't dare.

I did it better than you did. You know, all these dark guys fell in love with Norwegian girls. I also did when I was 18. And I told her, I see you again within 35 years. And I did yesterday night.

So now I sit down and listen to that. That's the next life. That's the next life. So I bought a one-way ticket and 3 weeks later I moved to Norway and I've been here for 20 years. But I've been starting companies. I am— I have a patent. I'm one of these people who if I have a passion and an idea there will be nothing to stop me. So this room that you're in was an idea I got back in November. I had a meeting with the Oslo Cancer Cluster and During the meeting I said, "You know what? One of my passions in life is to help people, like most people do, right? But in my position I'm responsible for Watson." And I'll explain to you what Watson is. Watson has taken in all the information in the world on cancer and we put it all into software to help doctors. And that's my responsibility for Norway. So I have responsibility for this incredible software that could help save lives. So when I was in that meeting I said, "You know what? Watson, IBM, we're going to be here in the Oslo Cancer Cluster, into this room. I went back to my boss, he told me, "I don't get it, I'm not giving any money, it's a stupid idea." I told him, "Watch me." So I came back here, I signed the contract, and I had 4 months to pay a hell of a lot of money. With only 3 days left before the end of the 4 months, I finally got the money. All the furniture you see in this room is used furniture that I got. This room has cost almost zero. The beautiful smartboard you see over there is sponsored by Cisco. That's a donation for free, and they're giving me two more on this wall in the next few weeks. My point is this, and this is part of the innovation discussion and stuff, right? Is that if you have a passion and a vision, it's not about money. You know, just because I'm IBM and we have billions of dollars doesn't mean that Were you guys expecting a room with thick carpets and real expensive wood and stuff because you went to the IBM room? That's the old days, right? So now Watson, which I'm going to explain to you what Watson is. Watson is now down the hall from Oslo's number one cancer hospital full of cancer patients. We're here to help. That's the driving force. The driving force wasn't to spend IBM money and make the thing very pretty. 'Oh, look at this beautiful view, look at our wonderful furniture, aren't you impressed with IBM?' I could care less. The point of the story is that we're here to save lives and that's why we're here. Do you get the difference? So she asked me to talk— she asked me to talk about innovation. I actually want to tell you the real story of how I got these things done. And I'm not— I say I because there's been nobody else behind me, but then last night I took pictures of everything. And I have all my colleagues saying, "When can I come? When can I bring all my customers?" But for the last 4 or 5 months as I've been building this place, not one of them have come here to help me. It's fine. But the point is to save lives. The point isn't anything else. Do you get it? So that's really important that you understand what blood, sweat, and tears that you're sitting in this room. It's not here to impress you. It's actually here to save lives. What is Watson? Watson is IBM. You guys know IBM, right? We— everybody knows IBM. So IBM, we've been making the world's fastest supercomputers and mainframes for many, many years. But we also have 14,000 researchers, PhDs, and 5— a little bit over 5 years ago, we told them to make the human brain, the cognitive brain, the decision part of the brain, make it into software. They did. So we married our world's fastest supercomputer with the cognitive brain, we put it together and it gave birth to Watson. So Watson— we had this supercomputer, this super brain, and we did this 5 years ago, right? Watson reads 10 million pages in 1 second. When I say by read, if I tell you I love the Big Apple, Everybody knows I'm talking about New York City, right? The Big Apple. Well, Watson understands in that sentence that it's talking about New York. It doesn't translate it as, you know, I'm talking about an apple. So Watson reads 10 million pages every second. So Watson is able to read— so what we did was 5 years ago, we said, what is the biggest problem that the planet faces today? And that's cancer. So 5 years ago, we partnered with Memorial Sloan Kettering in the United States. We took 2 of their oncologists off of rotation. We sat them in front of a laptop. The laptop had Watson. And on one side we said, Watson, here is all the information in the world on cancer. Here's all the pharmaceutical drugs. Here are all the research papers. We fed Watson everything we could feed into it on all this data. Everybody heard of big data, right? The whole big data movement. So we fed Watson tremendous and tremendous amounts of data. And on this side we had 2 oncologists who were like Papa and Mama. Any of you guys have kids in kindergarten? Anybody? No? Long time ago? Does anybody remember the days of kindergarten? Yeah, of course. Okay, do you remember when the kid would come home from kindergarten and they'd come up to you and say, Papa, Papa, Papa, he did this, she did this, and all that stuff? Do you guys remember that? Right? So that day, what was that? The kid during the day was taking in all this information and then they came home and they said, Help me. And what was your job? To explain to the kid that this happened because of this, and that happened because of that. Yes? You follow me? That's Watson. Watson takes in all this data, and then the two oncologists, or you, you sit down and you train Watson. So you can ask Watson, Watson, in this situation with this patient that has this blood pressure, the red blood cells are misformed, the white blood cells don't have enough platelets, What would you do? And Watson, being a kid in kindergarten, would say, oh, I would do this, this, and this. And then what do you do? No, Watson. Bad Watson. You should do this, this, this. Okay, Papa. So you train Watson. And what we did was we had these two oncologists, Morris and Kettering, they've been training Watson. So after 3 years of training Watson, Watson now treats breast cancer, colon cancer, and colorectal cancer. And every single day it reads all the information in the world on cancer in those 3 areas. Every single week, 5,000 research publications are published in the area of cancer, each around 200 to 500 pages. No oncologist has the time to read all that. Watson reads all of that in half a second. Do you understand what Watson is now? Now that's Watson in health. And for oncology. But with that amount of power, Watson can be used in any situation. You understand? So here in Norway, we're talking to 14 different industries are looking at using Watson for all sorts of different things. So one of the things that I love about this beautiful view is people say, well, where's Watson? We're here at the Watson Innovation Lab. Where's Watson? There. Everything you see out here— you see the trees? We're talking to forestry companies who are using Watson to understand the different temperature, humidity, atmospheric pressure, the moisture of the soil in different parts of Norway. As you go up in altitude in Norway, as the terrain changes, they want to understand what is happening not just to the forest, but they want to know what's happening to each individual tree. Watson can tell them. 'Cause there's not a lot of data for Watson, it's nothing. It's very precise. Everything you see is involved with Watson today. Where before, you know how there's a big talk about smart cities, how the cities themselves are gonna be connected in IoT, you guys are familiar with IoT, the Internet of Things, putting sensors on everything? Well, let's talk about sensors on human body. Mind if I keep on picking on you? So, you know, we're talking about wellness and people and taking care of the elderly at home, people who are sick at home and things like that, right? So we have a project with a company called Under Armour. They're a bit like Nike making sports clothing. And they have sports clothing where it has a couple of hundred sensors on the shirt, couple hundred sensors on the pants, and they put it in the athletes. So the athlete during their competition or training, the coach actually knows what is going on in that athlete for every millisecond that they're moving. Every muscle, whether it's in contraction or pulling, whether it's hot or cold or everything, all that data is fed into Watson. So we actually have projects now when people are at home, she's resting at home watching TV, the doctor or the hospital, Watson's monitoring her entire body. We also work with athletes like a cyclist. I think this guy's really cool. I can show you the video, but I could just talk about it. There's a race across America, the fastest person to go from the East Coast of the United States to the West Coast. You guys ever heard of RAN? You guys want to try that race?

No.

It's a lot of riding. This guy has won it 4 or 5 times. It takes him 7 days to bike from the East Coast to the West Coast. That's insane, right? So the year that we were working with him, he swallows a pill, and that pill gives Trace— it has all the data that's happening inside of his body. He was going up a mountain in Colorado, and Watson sensed that in one organ in his body, the temperature started going up. Just at that moment that this— the temperature started going up. So Watson then told the van— because you know when you're on the bike, there's always a support van behind you— told the van, pull him over, get him to a hospital. They went— his name is Dave. Dave! They put over the stops the car and says, "Hey Dave, stop, you're sick." And Dave's riding the bike going, "I feel great!" Because Watson just sensed it 30 seconds ago that this guy's temperature and organs started going up. You don't know that the second it starts, right? It takes a while. So Dave and the band started getting into an argument, "Shut up!" "No Dave, pull over!" "Shut up!" Eventually Dave stopped, they took him to the hospital. They found that he had just started in the last 20 minutes an infection. So they hit him with some antibiotics, 6 hours later, he got back on his bike and he finished the race and he came in second place. My point is this, is that today we have the ability with all these sensors and by taking these pills to sense what's happening inside the body. We can not only monitor what's happening with athletes, but we also can monitor people who are sitting at home and how they're doing. It's just not monitoring, but we can monitor individual organs or individual limbs, individual tissue. Do you understand? And when we say monitor, a human being doesn't have to monitor it. Watson can monitor it. Another thing we did with Watson— I'll jump around from health to not health, okay? And then please ask me questions. I could talk all day, and that's not fun for you. Question. Question. Go. You tell us that technically everything is possible. Not everything. Almost. Where are the limits? Moral limits? What are— Fantastic. Fantastic. Hold on to that. I'm going to get to that in 2 minutes. Remember to ask me moral limits, right? And then I finish with this one, then I do that one. Because that's probably the most important one. Because that is the ethical That's a critical question.

It's huge.

To finish the picture on how Watson works in regards to body, another thing we did was we acquired a company that had over 200 million pictures of X-rays, MRIs, and other images. With each image, it has a description of what does that person have, what was the treatment, and whether or not it was successful.

And 200—

over 200 million different images. The reason why we did that is for Watson, data is food, right? You understand that concept, right? So we said, Watson, we want Watson to see. So what we do now is if you utilize Watson for radiology, a radiologist can take in— mind if I pick on you now? We take an X-ray of her chest. And here, at least here in Norway, if you want a radiologist to see your X-ray, you have to wait 3 weeks 2 months. The backlog is ridiculous in Norway. Yeah. So, well, with Watson, you take an X-ray of her chest, and the X-ray today is a digital image, right? And you feed— upload that digital image to Watson. Watson takes the image of her chest and compares it to 200 million X-rays of other people's chests. And then Watson comes back and says with 89% certainty, or 92.5% 4% certainty. Here are all the examples that are just like her, and therefore with 92% certainty we think that she has this condition, and then with 95% certainty we recommend the following treatment based on all these other millions of people who have been treated who were successful. And all this happens in under 10 minutes. And it all happens while she's still in the room. She doesn't have to get an X-ray and then she goes home for 3 weeks or 3 months and then she finds out. She gets the X-ray. As she's putting her clothes on, Watson's doing all this. By the time she's done and she's ready to sit back in the room, the radiologist has— well, not even the radiologist, but the radiologist or the doctor has the final information and we can have a discussion. Do you understand how Watson works? Watson already also that smart that it can also measure which body can cure itself without medical intervention. Because that's what we talked about. A lot of what is done in medicine is done unnecessarily because the body could cure itself. And I think if you can measure all these kinds of types of processes, one day you have all— we all take a pill and everybody can be followed for each first cancer cell, for each first whatever. But there's something different. And then, Curie. You guys have asked the two first questions, the two best questions I've gotten. You're Dutch. You're Dutch. Next slide, please. That's like the saying for the weekend. I'll say that to my wife tonight. I'll get myself in trouble. So let me go to Robert. I'll turn my tone down because your question is so heavy. And I still owe you a moral, I haven't forgotten. So this is— now I'm going to get humble on this whole thing. And this is one of my personal frustrations. So now I'm asking for a bench from you because I'm responsible for Watson. One of the frustrating things I see from a corporation point of view that we've done with Watson is that we've taken Watson to prove improve or to reinforce the existing paradigm of medicine. You guys all see that? Yeah, yeah, yeah. So now, this is why I'm not speaking for IBM, now I'm speaking as myself, right? Okay. So, and I believe from a corporate point of view, we're doing that because this is a product that we are selling and the customer is want to reinforce their existing paradigm. Got it? Now step over here. This same technology absolutely can be used to do what you're saying, right? With all this data and all this information, each one of you being unique human beings, and each one of you, if you have a certain condition and you have a certain condition, the reason why you both have condition is completely completely different, right? You can't give you both the same medication. You are so different at a molecular level. And also, when you breathe in, the oxygen reacts in your body differently than it actually— I mean, we're so unique, right? So Watson can treat, I believe, and we will get there, where each individual person could have a, say, a computer system monitoring everything about him, what's happening with him, not just monitoring, but maybe the environment around him. Taking into account everything around him, taking into account all his cells, what's happening to him on the cellular level, and then making recommendations immediately, right? It could be you're low on iron, you should take this or take that, or, uh, you have a toxic exposure, you should maybe do this or do that. Absolutely, that's what the technology is there for. But right now we are reinforcing the existing thing because we're trying to make money on this product. But the This product is available as an API. You guys are IT. This product, the technology is available on the internet in the cloud so that any startup can do exactly this, what you're saying. Absolutely. But IBM, we put together a finished product to reinforce this one and to make money. You asked a damn good— you're the first person to ask me that question because this is something I think about all the time. I believe what we're doing is to make the money But where society needs is the innovation people who put together this stuff and say, "Yo, you're unique, you're unique, you need this, you need that." You know, that's the way to do medicine. Fantastic.

So the future—

Thomas, did I answer your question? Did I avoid it? I wasn't trying to. Yeah, it does.

Okay, then I have another thought.

So if you didn't give the food, Watson can't think about it. What you said is, well, we do surgery for a certain condition now, but we don't know what should be when we did not do the surgery. Maybe that was the best condition. And if you don't feed also with that condition, you will never find out. You will never be able to prove that. That's what I mean. So you're absolutely right. But you're absolutely right. If you don't feed Watson the data on the history of that condition, then Watson's like a child who just came home from kindergarten, right? But what's amazing about Watson is, let's say, uh, I'll be Watson right now. Each one of you are a source of data, okay? If I only have these two sources of data how many conclusions or what type of conclusions I can make. My perspective is this much, but as I tap into all of you, my perspective gets bigger, right? What becomes amazing about Watson is the things that it finds and connects that we never thought of. Because there's so much different input that it says, you know what, this is related to that, and we never saw that before. So that we see all with Watson is that it— what I mean is in the software we call it anomalies. And I love anomalies because I always get myself in a lot of trouble with anomalies. There was an electrical company in northern Norway, the CEO gave us 2 years' worth of data and he said, "This is my electrical grid in Norway, it's the world's best electrical grid. I want Watson to look at it and tell me what it finds." So we presented what we found and then we got to this one There was one anomaly that Watson found. And we were on this conference call, so they were up in northern Norway, we were here in Oslo, and we got to that talk about the anomaly, everybody in the room said, Thomas, you got to talk, you got to talk, we're not going to bring it up. So I said, Mr. CEO, we found something. And I said, Watson found within 2 meters, here's the exact coordinates, in your network this is happening. The phone went dead silent. After about 30 seconds, we started sweating on the IBM side thinking, okay, we just got fired. And me, because I had a little bit of sense of humor, I put on like a little feminine voice and said, hello. And then coming back was this, Thomas, two things are going to happen. The first That never happened. The second, my lawyer is calling you to sign that you know that that never happened. I have spent the last 50 years spending hundreds of millions of dollars saying that that cannot happen in my network, and you found it happened in my network. My lawyer's calling you right now because that cannot go public, because then we are sued. We will be bankrupt together. Because when you connect all these different data sources that have never been connected before, something in that connection, there's something we never thought of, right? How many possibilities can you think of? You can think of this one. We're human. We're limited. Computers like Watson work on a quantum level. You, any of you gotta study quantum physics and things like that, right? In a quantum state, an object, a molecule, an atom can exist in two different parts of the universe. We know that in a quantum level. Well, that's what Watson thinks. We think 1 1 2. Watson is thinking on a quantum level, so it's connecting dots that we can't possibly connect. So one of my favorite parts about Watson is always the anomalous. I love asking, hey Watson, what did you find that I should— I don't even know how to ask the question. What should I have been asking? You know, and Watson goes, da da da, and we go with some project. What the hell is this thing telling us? Because it's phrasing something that we never thought of. Then you have a room full of Smart people, we spend a hell of a lot of time to understand what they found because we would never think that way. And that's the beauty of the anomaly part of Watson. But still the question I think is— Yes. That it's based on combining all type of data. But can Watson create new data which was not delivered into the computer? Yes, through the anomalies. So Watson will identify— Well, you guys have some great questions. Watson will identify a new area and then start creating his own data because that data has not been collected there before, but it needs to collect that data so it can come with new conclusions. Absolutely, it will do that. But you have to tell it to do that. Remember, it's still software. You still got to tell it. It's not self-learning. Self-creating. Okay, so stop. So, no, no, you got that. God bless the duck. When we created Watson, there was that first moment that we created Watson, we went into that thing. Should Watson be self-learning or should Watson be trained by people? We said humanity needs us to train it, that way Watson and people, that's the future of humanity. Self-learning software is fantastic because it exists in the quantum level, it exists in multidimensional, it has no limitations. That's a fantastic thing, but you don't control that. So the example I always give is self-learning software. If you ask self-learning software, what is the best way to save the environment on the planet Earth, the answer is always going be get rid of you. So be careful of self-learning software, right? But self-learning software, because it doesn't— is not limited, it will come to conclusions that we need. Absolutely. But be careful of self-learning software. All right, so the answer to your question is we have Watson Watson that is self-learning, but we do not make it available. The Watson that says, "Train me, Papa, teach me, Papa," is the type of Watson we make available to everybody. Because self— and by the way, Google has self-learning, right? And they bought a company called DeepMind in the UK, and that's completely self-learning. It's fantastic. I think self-learning is amazing, but if you can't control it, be very careful. It doesn't have limitations.

Right?

But it's inclusive. It will probably find cures for things that we never thought of, but at the same time, it also keeps on going to other places. So be very careful.

Thomas, what in your opinion is going to be the role of the GP in the future?

Right now, the way we position Watson is Watson— the way I position Watson is Watson is the GP's best— is the ultimate version of the GP. So when I'm in front of a bunch of GPs, I like joking with them because most of them have no sense of humor because they all think they are so, you know, it's really funny. So I go, hey, do you guys remember back when you were in university? And they're all sitting there, yeah, I was the number one in my class, all the ego stuff. And I go, do you remember on that Friday night when there was a bunch of male GPs that say, she came into the room with a really short skirt, she said, do you want to go out? Or I told the female GP, I go, He came in with his muscles and said, "Hey, you want to go out?" What did you do? You went out. Do you guys went out when you were in university and somebody cute knocked on the door? I'm asking you, yes or no? Yes? I still do. Okay. Well, Watson is that version of you that never went out and kept on unstudied. That's what Watson is. So today Watson is the assistant to the GP, right? Watson's read all this material while you were sleeping last night. There was a publication from some research journal that was published in Australia that by coincidence is, uh, their conclusion is similar to the patient that you have this morning. There's no way in hell you were sleeping, but Watson has it here. So then when you walk into the room, the Yeah, I'll show you. This is actually, this is a real scenario. Anybody have an iPad? Everybody has an iPad. So how big is this one? Oh my God. Okay, so this is, this is actually what happens today. And we have Watson in over 24 hospitals around the world, uh, that we can talk about. We have Watson in more hospitals, but hospitals don't want to talk about it, and I'll get to that in a second, right? So this is what happens in a real world. So you guys are the patient, this is the patient room. This is like, damn, this thing is big. Oh crap. So the doctor walks in, right? When the doctor walks in, who's going to be my patient now? She looks like she wants to be a patient. So when the doctor walks in, what Watson has done, by the time Watson has come in, Watson has looked at your entire electronic medical record. Watson has looked at the entire history history of cancer up until the time the doctor has walked in, including the 2 o'clock in the morning Australia thing, and Watson has made a recommendation saying with 92%— it has 3 lines of information. It says first line is the highest recommendation, so it says 92% certainty that you should be treated such and such a way. Then it has another one with, uh, 70% certainty and another one with 10% certainty, right? So then the doctor comes over and goes, can I touch you? I'm not gonna shake the doctor's hand. How you doing? Yeah! The doctor comes over and goes, "How are you today?"