Innovation Heroes

TRANSCRIPT - Vaccines were just the beginning: Intel’s Stacey Shulman talks AI healthcare revolution

June 8, 2021

Peter 

This episode of Innovation Heroes is brought to you by Intel 11th Gen vPro-based client devices, delivering the highest performance and most comprehensive hardware-based security.

[music plays]

 

Stacey 

Do I want to share my information so somebody can target me with a better advertisement? No, thank you. But if I knew that I could get customized and personalized medical treatment, yeah, I'll sign up for that.

[music plays]

 

Peter 

Welcome to SHI's Innovation Heroes, a podcast exploring the people and businesses giving us hope in our drastically disrupted world. I'm your host, Peter Bean.

[music plays]

 

Peter 

When we look at the way the tech landscape has changed lately, one industry that stands out is healthcare. Just think of the early response to get the ventilators out. [ventilator whooshing] The incredible coordination that went into rapid wide-scale testing, and of course, the vaccines, multiple, viable, miraculous vaccines in less than a year. We've also seen some major changes in how the healthcare industry collects and stores data. The days of your medical records being filed away in your doctor's office collecting dust are numbered. Most of your static paper-based health data is now being stored digitally, on local servers within hospitals and clinics, and sometimes even in the cloud. And the potential here is huge. We're talking about a major game change at all levels from the doctor, to the patient, to the large healthcare organizations, which begs the question, what if we wanted to do more with that data? What if there was a way to retain privacy and security, while also making our data more valuable? Scratch that. What if we needed to do more with it? What if the answer to a lot of the problems our overburdened healthcare system has seen in this past year could be fixed by a solution that's already years in the making?

[music plays]

 

With this digitization of data comes a new series of opportunities, most notably artificial intelligence and machine learning. Before COVID hit, about 45% of healthcare leaders were already using or thinking about using AI. Since the pandemic, that number has nearly doubled to 84%. So, here's what I want to know-- how is AI going to use our data, our personal healthcare data, for the greater good? Who's using it, and what are they doing with it?

[music plays]

 

To help me have this conversation, we are once again joined by Stacey Shulman, an Intel executive with the best title in the industry. Stacey is serving as Intel's Vice President of the Internet of Things Group and General Manager of Health, Life Sciences, and Emerging Technologies. Nailed it. In recent months, Stacey has been evangelizing the need for our strained healthcare system to transition into a new era of AI-driven patient outcomes. And I can't wait to learn more about it. Stacey, thank you for joining us. It's our pleasure to have you back as our first return guest on Innovation Heroes.

 

Stacey 

Thanks, Peter. It's great being back.

[music fades]

 

Peter 

It seems like you have a job now that is directly linked to the struggles we're all going through. What has the last year of your professional life been like?

 

Stacey 

Yeah, it's been wild. So, um, when I remember back in February of 2020, the things that I thought I was going to be talking about through 2020 was consumer experience and digitization of the experience. About two weeks after I attended the opening of the Sinclair Hotel, which we've talked about, I received a call from friends at XPRIZE, and XPRIZE is an organization that does big innovation challenges. And my organization had been working with them on setting up infrastructure for data scientists, and what we were focused on at that time was, we wanted to track the destruction of our rainforests. And so, we were setting up a multi-year project, and the intent was, let's put an environment in place for data scientists so they can log in and immediately get to the work, instead of spending most of their time trying to fix the data and get data ready. So, anyways, COVID hits, and this organization calls me and says, "I know we were gonna do that for rainforests, but can we use that instead for COVID?" "Because we-- we have health companies coming in with viral records and wanting to do something with that immediately." And, of course, we said, "Yes, absolutely." And within less than a month, we had millions of viral records in that database, and we had data scientists working on looking at that data to track COVID trends. When I look back, the last year has been absolutely consumed with how do we take the collective intelligence of our community and get them, you know, focused on what matters most right now? And that is stopping this pandemic, saving lives, and making sure that we're preparing for whatever comes after this, so that we don't have to go through this again.

 

Peter 

Yeah, I feel you on that one. I don't ever want to experience this again, and I'll tell you, if we do go through something like this again and have the same challenges that we faced the first time through, it will be wildly disappointing. But I don't feel like it will be. As we get into AI and healthcare, specifically federated learning, can you explain that for the layman, exactly what's going on, and what Intel is trying to build with it? I feel like everyone's going to need that as a basis point for what we're going to talk about.

 

Stacey 

Typical US hospital in the United States generates about 50 petabytes of data a year, and that's just a lot of zeros, just know that, and I don't even know how many zeros. I'll have to Google it. But 90% of that stays in the hospital, so what that means is there's lots of information sitting there that you can't access, and it's-- it's impossible to move it and centralize it to the cloud. It's just too much data. And so, what that-- you know, what ends up happening is, you've got loads of really valuable information sitting in places that you can't access it. So, one of the things that we started looking at is, instead of moving the data to the algorithm, which is a centralized server, let's move the algorithm to the data. Let's move that algorithm and start training computer vision and AI models on that data so that we can then start seeing the correlations in the data. Had we had this earlier, we would have seen correlations in pneumonia data linked to a new type of flu, which is the way it was being reported at the time for COVID. So, if you take that algorithm, you move it to the data, then what you do is you move the insights that that algorithm generates, and you bring them together in a federation. And so essentially, each hospital would take their own data, and without-- without having to share that data with somebody else, they can analyze that information based on an agreed upon set of rules with other hospital systems of what they're trying to look for. They look at that data, and then they get the insights, the correlations, and then every hospital then shares the correlations, and they come up with a really robust model and set of information that can then be shared across all the hospital systems. That in a nutshell is what federated learning is.

 

Peter 

Okay, so I'm going to read a quote from you, and then I have a very specific question about it related to this. You said in an article that I read, "One of the things that's currently holding the healthcare industry back is standardizing medical records and data sharing across organizations." So why is this so relevant today, with everything that we're facing?

 

Stacey 

Well, I mean, I think it's always been relevant, but what I'm seeing now is what COVID has done is it has unlocked the medical industry's ability to collaborate, and it's driven the need for it in a way that we've never seen before. And I'm seeing collaborations that we've just never seen in the industry. So, there's, one, we've unlocked the ability to collaborate, two, we have technology that's ready today, whereas before, you know, had this happened 20 years ago, we wouldn't have been able to talk about artificial intelligence, and the maturity of artificial intelligence just wasn't quite there. So now today, we have maturity of artificial intelligence, we have maturity of technical infrastructures, and we have this new, reinvigorated way of collaborating. And so, with that, it's unlocked a way of, like, working together. And so, it's important in that if we want to stop these types of things from happening, if we want to recover faster from this and get people back to work safely, we have to work together, and we have to use the information that's there, so that we can create the right treatments for people.

 

Peter 

Can you give us a couple of examples of-- of just really simple examples that people might understand, of what kind of treatments might be created, or what kind of-- what kind of wins we can unlock in the near future?

 

Stacey 

Yeah, so one of the things that we saw come out of this, if you look even at the vaccines, so you had Moderna, I'll use them as an example, who had been researching mRNA approaches for delivering treatments to people, and here's what happened with COVID. For the first time, you know, those-- those vaccines usually take years to deliver, so from the time that the sequenced DNA was delivered to Moderna, from that time to the time they had a vaccine ready to start testing, was about a week. And that's just incredible. That's why we need this kind of data, is that approach can be used for way more than things like COVID. It can be used for other illnesses, now that they've proven out that platform, and they've accelerated-- that was all accelerated through collaboration, and through, really, I'd say, focusing in on a real need at the moment. But now we can take that learning and, with other data, combine it and start creating therapies for things like Alzheimer's and all of these diseases that we've been really trying hard to solve for, we think are kind of ready to crack open now.

 

Peter 

Wow, wow. [Peter chuckles] Okay, that's the good. Let's take a step back and talk about the fear that comes along with all of this, right? Privacy is obviously going to be the biggest hot-button issue in this topic. Can you, again, in a layman's level, explain the challenges there, and are these concerns really rooted in reality that I'm reading about, or are they just simply born out of fear of the unknown?

 

Stacey 

I think there's both. I think that every time we're concerned about privacy, it's a valid concern. It should be top of mind, especially for technologists. We have to be very careful with understanding that the data that we're being trusted with is sensitive, and it needs to be handled with care. And so, I would say, yeah, there's some horror stories out there. There is some fear that isn't, I'll say, as real. But truthfully, like, you know, if we think of what we would want with our information, I wouldn't want my doctor emailing my information to some researcher. You know, I'd want to know that all of my identifiable information was stripped out of that, yeah. You know, I'm okay with my data being used in combination with other people's information so that we can create the right therapies for people, but I wouldn't be okay with that, you know, finding its way into the Internet, and published, you know, publicly. And so those are the types of things that we have to be really careful of, and thankfully, the medical system has-- has that in their DNA. That's part of the challenge that we're dealing with is, there is so much regulation on sharing information that we have to be extra careful. And so that's good and bad. I think we have some great solutions for how to-- how to fix that, and that's one of the reasons for federated learning, is you keep the data where it's at, you keep it protected, and you share the anonymized insights from the learnings.

 

Peter 

Sounds like the benefits could be so overwhelming that people could very quickly get behind this concept and behind this idea of their data being shared in this fashion for-- for these outcomes. Am I getting this right? I mean, is that-- is that the way that you see it?

 

Stacey 

People are willing to share information when they understand what it's for. That's true. Do I want to share my information so somebody can target me with a better advertisement? No, thank you. But if I knew that I could get customized and personalized medical treatment, and/or I could know that what I shared helped save somebody's life, yeah, I'll sign up for that, if I can also be assured that my data is being treated with confidentiality and professionalism.

 

Peter 

This episode of Innovation Heroes is brought to you by Intel 11th Gen vPro-based client devices. [music plays]

 

Now, more than ever, the business is counting on IT to save the day, and if keeping everyone in the office connected and hyper-secure wasn't hard enough before the pandemic now that office is turned into thousands of little offices. Say hello to the remote everything world. The new 11th Gen Intel vPro platform helps IT manage it all. It's the unrivaled business PC platform with the world's best mobile processor. That makes collaboration more productive, with up to 41% faster productivity during video calls, and it also boasts the world's most comprehensive hardware-based security for business. That's helping protect your team against emerging threats, even when software may fall short. And now it's available on the Intel Evo vPro platform, providing the best thin and light laptop experience, including features like instant wake and amazing responsiveness. The Intel vPro platform is built for business. It's how IT heroes get it done. Want to get started? SHI can help. Visit shi.com/intel for more information. [music fades]

 

I think you're on to the right point here, right, in how we create buy-in for this. It really is around education, around those benefits. There's a lot of fear around AI in general, right, so I'm wondering if you can give me a couple of real-life, even just one real-life example today that's already in production, where this technology or similar technology is saving lives that weren't being saved recently, or is making a positive difference in access to care or support, something that people could really understand and get behind so they can see that this has already been done, and that these positive changes already exist in our society?

 

Stacey 

Well, I would say AI is all around us, you know, and it's pretty prevalent. We give it different names. Sometimes we call it chatbots, and sometimes we call it an Alexa speaker, and we call it voice recognition. There's a lot of AI in our lives today. And so, it's already part of our daily life. I think the question is, how in depth will it get into our personal information, and what is it going to do with that personal information? When I think about AI, I'm a huge proponent of explainable AI. I think we have to be able to understand when models are being built, when correlations are being made, about information about us. There should be a way to explain it, and we should have the option to opt out at any point. We should be able to go back and say, "How is my data being used, and do I agree with that?" And I want you to stop using my data in that way." All of that is our right. And I think as we do that, we start becoming okay with-- with the use of our dat. In the medical space, even things like x-rays and CT scans, AI is being used so that you can have less exposure to that radiation. So, what will happen is AI starts making some assumptions about the fact that, okay, there, there might be a bone there, and it can, you know, remove that out of the way so that you can get a better picture. So, things like that where understanding, okay, your leg isn't being placed in the right position for the-- for the X-ray to happen, so reposition the leg. You can use AI to determine is it positioned correctly or not before you start with the X-ray. That would be one example. We've also seen other examples where ultrasounds are being done. It is absolutely mind blowing what you can see now. It is really like seeing the baby, you know, in full view. And, you know, that gives the clinician a lot more that they can do, and they can see the health of the baby a lot-- a lot clearer. So little things like that, it adds up. It makes a big difference. Those are some of the ways that AI is being used today.

 

Peter 

I think that's a perfect example. I mean, it touches on something that matters so much to so many of us, right, the health of our children, of our offspring, and something I think that probably creates the most stress for any of us in our lives. So, I think that's a perfect example. That's what I was looking for. I want people to understand the good that can come of all of this, right? Circling back to what you said at the beginning, it just really got me, thinking about how what you've done here with this data from COVID could unlock new opportunities to cure other diseases, like Alzheimer's, for example. You know, if this helped us get a vaccine done in that quick of amount of time, I can't stop thinking about all the other things that we have been unable to solve, the problems we have not had, the processing power, knowledge-- pick your-- pick your term-- to tackle, feels like we can now. You know, you mentioned people can opt out of this, right? That's a necessity. It's important. I think it's just as important for people to understand why they shouldn't, right, and the education behind that, because of the potential for what it can unlock.

 

Stacey 

There's nothing like an emergency that gets people collaborating and seeing the power of collaboration. And I believe that's part of it, is that, you know, you had doctor-- you have a lot of hospital systems, you have medical community of scientists, that generally compete with one another instead of collaborating with one another. And when you have a time of human lives are on the line, it's amazing how people kind of drop their need for acknowledgement in the work that they were doing, and instead of saying, "Okay, how do I get credit for this?" they focus completely on, "What can I do, and how can I use my expertise to better the situation?" And so, that's what I saw, is we saw organizations that historically were competitive with one another. I mean, again, you see it in the vaccine world, where you have one competitive company helping the other competitor create their vaccines, so they can up their manufacturing. You never would have seen that if human lives weren't on the line, in this way, in this focused way. Now, when we think about things like Alzheimer's, the question I have is, had we focused like that on Alzheimer's, and on cancer, and other diseases, could we have-- could we have had the same outcomes? I personally believe we would have, and we can. And so, that's the-- that's the question, I think, that we all have to ask ourselves, is-- not just even in the medical space, but when we collaborate with one another, instead of competing on things that matter like this. Can we do better? And again, what I'm seeing is, yes, we can.

 

Peter 

I absolutely agree with you. I think the million-dollar question here is, after all of this, right, when this pressure to collaborate is gone, and the, hopefully, the danger is gone, how do we keep this up? How do we continue to promote this type of collaboration? The reality is lives are still on the line. It's just not right in our face like it is now. How do you think we do that? And it's okay if you don't know. I don't.

 

Stacey 

Well, I have an-- I have an opinion. But, you know, I wouldn't claim to be all knowing on this. In all the experts that I've talked to you about this pandemic, and I meet with experts on a regular basis on this, every single expert refers to this as a practice pandemic. That way of thinking is, what should we be looking at to put in place long term? What's that biosafety infrastructure that we need to put in place so that we're prepared for the next one? As companies, how should this change our working environment? Should we have open offices? What should our coming-to-work environment be like going forward? These are the things that-- that we're all starting to think about, and I'm seeing a lot of companies have these types of discussions. I don't think we're going to go back to what we were. I think the world has changed. We are more collaborative, and that was happening anyways. You know, the next generation, the Zoomers are a more collaborative group of people than we historically have been, and they grew up collaborating, and so my kind of thought experiment on this says that we won't go back to a world where we start competing and not collaborating again. Will we keep top of mind the severity of, or the what's possible if we collaborate? I don't know. But I don't think we're going to go back to where we were before. This, if this really is a practice pandemic, then what comes next is ensuring that we absolutely take the learnings of today and not stop applying them, and not stop collaborating, and not stop working towards the betterments that we're working for today. So

 

Peter 

When this does happen again, and I can't believe I just use the word when, but when it happens again, we'll be ready.

 

Stacey 

That's right. And in between those times, we're going to use that biosafety infrastructure, we're going to use the digitization of the healthcare industry, to better our lives today. You know, my team, I'll say we are 100% dedicated to the digitization of the healthcare industry, and I'll tell you, we're all doing it as their lives depend on it, because we wholeheartedly believe they do. And I wish you and your team all the best with it.

 

Peter 

It is the most noble of causes. Just moving, to be honest with you. It's an incredible weight, but you carry it well, so thank you for being on the show, again. I loved having you back. I love our conversations. This one was absolutely fascinating. What you're working on is absolutely fantastic. It has the opportunity to change the world and make things better. So, thank you so much.

 

Stacey 

Thank you for having me on your show.

[music plays]

 

Peter 

One of the most fascinating things that emerged from my interview with Stacey was something she said about what allowed these AI innovations to happen. It wasn't that we suddenly invented a bunch of new technologies. AI, machine learning, all that stuff has been around forever now. It's that we didn't have the will, the teamwork, the collective kick in the ass to get everyone in the world to work together, to cut the red tape, and do something incredible, together. COVID did that. We still have a long way to go, but we're seeing the first step of what AI technology is going to look like in the months and years ahead, thanks to the work of people just like Stacey and her team. I truly hope we can look back in a few years and feel thankful for all the good that emerged as a result. Diseases cured, new treatments uncovered, and a whole new way for the world's healthcare experts and scientists to work together, without barriers.

 

Peter 

Innovation Heroes is an SHI podcast, with new episodes streaming every second Thursday on Apple, Spotify, Google, and everywhere else. If you liked this episode, and you want to be our hero, leave us a five-star review on your podcast listening app of choice. Next episode, we'll be talking to Omar Dajani, Head of Partner Solution Engineering for North America at Google Cloud, about the democratization of the superpowers brought to us by the public cloud. There's a ton of interesting new developments here, and you're going to love that discussion. Until next time, I'm Peter Bean, reminding you that wherever there's a challenge, there's an innovation hero waiting to be born.

[music fades]

 

Peter 

This episode of Innovation Heroes is brought to you by Intel 11th Gen vPro-based client devices. Visit shi.com/intel to learn more.

Podbean App

Play this podcast on Podbean App