AI Today Podcast: AI and Healthcare, Interview with Gil Alterovitz, Director of AI at the Department of Veterans Affairs (VA)
The AI Today podcast produced by cosmetica cuts through the hype and noise to identify what is really happening now in the world of artificial intelligence learn about emerging Trends Technologies and use cases from Cardinal Indica analysts and get experts. Hello and welcome to the AI Today podcast. I'm your host Kathleen Mulch. And I'm your host Wireless melzer. Our Guest today is Gil. Robots who is the director of artificial intelligence at the Department of Veterans Affairs and with the national AI institut. Hi Gail. Thank you so much for joining us on a i today. It's great to be here. Thanks for having me on the show. Yeah, and thanks for joining us. We'd like to start by having you introduce yourself to our listeners. Tell them a little bit about your background and your current role at the VA. I certainly so that's my background is had a doctoral degree from this joint program at Harvard in Massachusetts Institute of Technology MIT that kind of focused on AI applications in medicine and then became a Harvard Medical School faculty focusing on that integration and my dad actually had worked at Nasa. So I've been really interested in EM Mission related activities and the VA really has a great Mission and I you know, I guess I really had a chance to see that mission over when I was a White House presidential Innovation fellow program being a fellow in that program. I had a chance to work with the VA and really learn about the ignition. And so now working on AI and building that research and development capacity at that VA at the VA and through the new National artificial intelligence Institute. They're really start a few months ago has been really, you know, really inspiring I think because it's really nice. It really makes chemical very big difference for the veterans by supporting the veteran priorities and there's a lot of aspects of working together birth. Across the different parts of the VA that are really fulfilling. Yeah, I think that's great and thing and and you know, maybe our listeners are new to the fact that there is a National Institute and and may may not even be aware of the fact that exist. In fact that it's actually headquartered of the Department of the Veterans Affairs. So maybe you can talk a little bit about the national artificial intelligence Institute at the VA and what it is and of course kind of why it's important with it. Yes, certainly. Yeah, so it really what do you think about the VA is really it has this unique opportunity in the country for a research to be developed and and translated into clinical care of quickly to help the special population of veterans. It's it's got the largest Integrated Medical record system in in the country. It's and it's got the juice largest genomic data collection and the patient's really asked us to deal with their needs and to to prioritize on that work. And so what I'm doing with the national artificial intelligence Institute at the VA is working to build that capacity that capacity to do research and development for these parties of the idea of working on different Partnerships Partnerships, which may involve other organizations outside the VA as well, and yep. About collaborations and really enable leveraging of cutting-edge AI artificial intelligence in order to do that as well as working on different policies and off other areas that can help Advance this this field so we can think about the V really can be essentially a go to place for veterans through a I'm a research and development. We can really help Pioneer this approach where we can anticipate the needs of the veterans through the use of artificial intelligence and deal with those health and well-being issues to improve the lives of of the of the veterans and which has applications across the country. Yeah, you know, thank you so much for that overview. I know that a lot of people probably interested in digging deeper into that and just learning more and following that as well. We were fortunate enough to have you deliver a keynote during our data for AI week virtual off. For instance took place that in September of 2020 for our listeners that were not able to maybe attend that live you're thinking and maybe didn't get to attend your keynote package for those listeners that may not yet have been able to walk that and as a reminder it is available for replay so you can go to a data a icon, that's data wage and watch any of the sessions that were presented during the conference either live or on-demand or free that'll be up through the end of the year. So we encourage you to go back but for our listeners that haven't been able to watch that yet. Could you share some insights into the you know, what was delivered maybe some high-level high-level overview of what you were test. Oh certainly super-smelly was a great conference a really encourage everyone to see a number of the talks. So we're at that conference and was a it was a pleasure beyond that so often in my particular keynote what we talked about was the national artificial intelligence Institute as essentially as a use case and how it builds on some of the principles of the American voter initiatives and the national artificial intelligence research and development strategic plan of 2019 for the United States. And what we do is we we outlined how we're working with an AI passports within the VA on developing a strategy and getting that that actual veteran input through a veteran engagement board. And so using those elements combined with analyzing technical feasibility. We've also been able to outline these five fundamental Key areas of AI in applications in health and well-being. And so in that talk I go a little bit more into those five areas just to kind of tease them out there Thursday around deep learning trustworthy artificial intelligence privacy-preserving artificial intelligence explainable artificial intelligence and final multi-scale artificial intelligence analysis. These are five areas that we explore in the in the talk and these are five areas of life. We have found potential to be a trends that will the trends are leading these areas to potentially have a religion Major Impact by further developing these areas and applying them around the priorities that we're identifying for for the population and Thursday. Actually, we also talked a little bit about the new high-tech Sprint, which is starting soon. And the applications are actually now open for that. We ever very successful Sprint last year and that's been around setting up these collaborations and potential future Partnerships with industry non-profits Academia and others around data in research and development that's going on at the VA. We also work with other agencies and other data sources in the future in the in the past around different themes last year. The theme was around clinical trial search and matching in this year. The theme around interventions for veterans that are not currently being served by the VA those that maybe have not been engaged in the healthcare system and that they're dead. Maybe ways of helping them around areas of Health well-being and or are related areas. And so that those are kind of the overall topics and Thursday if you'd like to learn more, of course, I believe the keynote is still up there so you can listen and Yeah, it's definitely interesting and we do encourage our listeners listening. And of course the thing about all of our online conferences, of course all conferences are online these days is that it's free to attend. So, you know, and with the sessions are usually up for multiple months after the conferences are over cuz we know that you know, people are busy and working from home and you know life is different these days and you want to give people the opportunity to check them out. So we do encourage you to check it out against data AI Khan dat aai conf, and look for a girl alternative. It says keynote presentation off our sessions and you can definitely take a look at it and even some ways to interact with some information about the speakers are are there so you can reach out if you have any questions so really really thanks so much for giving us a feedback. Sorry a preview of the what you had talked about for those who haven't had a chance to check it out. So, you know, our listeners are around the world and their wage From kind to know folks who are perhaps in different Industries. Some of them are a lot of them say the private sector and the public sector and Academia. They're they're across the board and they may not be familiar with home of the scope of what happens at the Department of Veterans Affairs and the VA and just in general and maybe they're not even aware of sort of like the scale and size of the VA as a health care institution. So and and make sure that that will provide a little of insight because there's some interesting things that that you're doing with a i in the VA. So maybe you can share with a some ways in which you're leveraging AI or or maybe even looking to change to benefit veterans the VA and just Healthcare in general for those that are not familiar with the VA as an institution. Certainly. So there's a number of areas that is being leveraged, you know a few months ago. I was reviewing some of the different projects and there are literally hundreds of projects that in some way are leveraging artificial intelligence and it's not trying to think of what are some of the you know are their their civic types of examples and you know, so let me just kind of share with you a couple of that come to mind. There's you know, I think these are kind of interesting in a couple of months. So the first one is about you know, there's work that's on going to create the screen work whereby artificial intelligence research and development can be integrated into modules that would work in a clinical system. So as I mentioned before the VA being the largest integrated healthcare system in the country has hundreds of medical centers wage. Thousands of different facilities across the country. How do you any uses a similar medical record system across different sites, but you can develop if you have something with artificial intelligence that can help veterans or you know population in one site. It can quickly be deployed in many sites. And so we're working on developing a framework, whereby if people have a particular method that can work in in one in one type of application. It can be used quickly and piloted across a number of different locations for that for health and well-being in this particular case. This was around covid-19. If someone has a positive genomic tests for covid-19, what is their prognosis and so a dashboard report has been developed around that and that. Now being looked at being tested and getting user feedback and then piloting and so in developing that you got was or particular application. We developed in such a way that that framework could be used and replicated for other types of AI modules in the feature that researchers are people are developing different approaches could quickly plug in that other type of algorithm algorithm and and try that for a different type of application. So we're home that day. I talked to go another example mentioned the AI the AI Tech screen little while ago that we had one last year was about to start a new one right now where the applications are currently open. And and in the last year there was actually a tool created by this non-profit organization that basically seems to impart veterans to search page. Clinical trials, you know actually including covert related ones and what what it does is it leverages deep learning one of those five areas that I mentioned that we go over a little more detail in the keynote and it uses that to analyze the text of all the clinical trial eligibility criteria to know which trials to be a seat for that bedroom given their clinical information and elaborate ages. These means that were developed by God by the VA to actually access to enable a veteran to access their own clinical information. And so that is then leveraged in order to on of course consent of ordering to find the trials that match the clinical eligibility criteria and their medical records information so that they can get the trial that's best for them so long the past, you know, traditionally move happens, you know, the person would go in and they would have to either call in or kind of rely on their physician and caregivers and and this isn't to take away from that but really to augment that to empower individuals to look for trials and on their own see what's out there and then really essential is a Showcase of how you can leverage deep learning and AI to combine different piece information information coming from the VA around that clinical information right with the veteran log to get their clinical information and then information on clinical trials, which actually also comes from the Federal Government Federal Government from other other departments within the federal government, whether it be the NIH National Institutes of Health or other parts of the government that have information about the Dead Trials that are ongoing into given a point in time. So those are two examples of you know, one is around as mentioned the spring work enabling I research and development to be integrated to modules that I work well with a clinical system so you could go quickly from developing something to actually getting it to helping the the patients that Thursday we're here to serve and then the second around how we're enabling through Partnerships other organizations turn power to empower the popular package to to use artificial intelligence in a way that also promotes health and well-being through use of artificial intelligence. That was great. I know, you know Willy always like to hear how may I is used within different agencies and different departments. And I think that people can learn from that, you know, whether or not we're in the government or not. You can learn you know, this can be applied to Healthcare in general. And so we always like to hear when when you know different people are fearing these exams as well so much so that brown and I put a year-and-a-half have been running the AI in government event. So as always that's a free to attend event. It's now gone virtual like everything else. So if anybody would like to check that out go to in government, we have a different speaker every month who will promote or you know, and share and showcase how they are using artificial intelligence. So we encourage you all to check that out. We also have a bonus episode coming up with Gayle as well. So we're very excited for that. I'll make sure to link to that in the show notes so that everybody can log Listen to that will be digging a little bit deeper in there talking about some challenges with adopting Ai and Healthcare and then what the VA is doing to help develop an AI ready Workforce. I know that some of these are really hot topics right now and we're excited to dig deeper with that in our bonus episode. But before we wrap up this main podcast, I always like to end with this question for all of our guests as a final note. What do you believe the future of AI is in general and its application to organizations governments and Beyond. Well, I think let me start by first thing that the future of AI is definitely bright some of the the first areas where I think we'll be able to see benefits are around increasing productivity and efficiency to really augment humans in their decision-making processes. So not you know, not replacing but more often engine adding additional information so that people can take and analyze information that otherwise they may not have been able to you know, no person can read through the clinical notes of you know, millions of patients to be able to understand how their patient is different from all the other patients or how similar but a computer a computer can take that summarize that information make recommendations that then the physician or other caregiver they decide to use or maybe decide not to use so they're they're really a lot of dead. Crossed keys that I think can benefit from AI some from basic areas like a forms processing. There's some of that going on at the VA as well and notating wage images these maybe images such as a clinical images whether it be a radiology pathology a number of different areas there as often should process in clinical notes also promising area because it's really there's just so much information there to analyze there's certainly some technical challenges ahead in terms of computing resources that are needed for scaling for analyzing all that information. And also I think importantly there's this area of ensuring seamless integration into existing workflows. I think really hard not to I mean really can't emphasize that enough that ensuring that Integrations existing workflows is a way that you age Make sure that it's adopted that that partition zones is is used in adopted if it's not integrated into the workflows then positions or caregivers are are less likely to adopt and leverage that new technology. They're also issues around ethics in artificial intelligence that that are being looked at proactively so important to take a look at those areas and to really understand them before we take and develop mechanisms that made me involved different complex issues that wage is there and then some of the issues that are unique to artificial intelligence and some that, you know, we can see in general for any new technology but there are some that are unique to artificial intelligence but in any case, um, you know AI is most likely it's more, you know, it's most likely to be adopted birth. We're ready to be adopted. Once it is encapsulated in a in a tool or a programmer or some means we're expertise actually not required to use it. So essentially making it user-friendly is very important. So just as today, we actually are starting to see artificial intelligence used across the number of different means whether it be social media applications, you know, whenever you get a friend suggestion that can leverage I mean artificial intelligence or whether it be mapping software or maybe predicting the best route to a different location best road to take under the given traffic conditions and other tools those essentially encapsulate the office underneath and so they're friendly easy to use and you don't need the expertise to use it but to do that you need to have kind of a couple of different layers and that wage What is I see now being worked on within this field? So not only developing the actual best methods that you know, basically the ones that image of the mapping software give you the fastest route to your destination, but also one that are friendly and easy to use so that you can you know, listen to it and and decide what action to take off on that. So in summary, there's a lot of potential applications and I see a number of organizations looking at leveraging it and really the key off to take the next jump in the next week or word is all around adoption and to do that. It's important to proactively be thinking about ethics be thinking about issues wage. You're talking with the keynote around explainable or official intelligence so that the user understands why the algorithm is making a prediction certain way so they can decide if they would like to age Use that or not and be comfortable in the decision that they make and all of these in able to take the next step. So I think that's kind of a quick summary. Okay, great. Well, we definitely are enthusiastic about adoption to I mean that's actually much of the reason why we spend our time on these podcasts as well as Rai and government events, which for those of us who are listeners to this podcast may or may not know we also run a regular monthly online event series called AI in government where we spend time talking with folks in the public sector not just here in the US but across the world we've had Lord Tim Clement Jones from the UK House of Lords. We've had many folks, of course from US federal agencies. We also have state and local and you know just different folks and and and it's interesting to hear how AI is impacting the public, you know, the public sector that you know governments that you deal with, you know, every part of birth. I feel and and really contributing to it. I think it's an interesting to have this perspective. Of course, we spend a lot of our time talking to folks in Industry as well, you know banking ensuring my finance Healthcare retail automotive and pharmaceutical and all those other interests Industries in between right? But but then of course, you know the motives a lot of the motives for folks are putting in a i in to say a business are different than the motives for government, you know governments about serving the constituents and and making things more efficient and more effective and not necessarily about the bottom line, right? So it's always interesting to have that other perspective and to see how AI is being adopted even when things like profit is not the motive, you know, not about getting rid of people for example, people always think about that long and that's of course not the motivation from for most governments. We've talked to folks even governments out in in Norway. We've had a city manager for as low talking to us and you know clearly they're not Scandinavian. Since aren't motivated at all even in the government sectors, they don't treat it like business. So it's really really very great insights. And so Gil I want to thank you so much for joining us here on this podcast and sharing your insights. And of course, we'll have you on for our bonus episodes as well. So for those of you who are listening, please go to AI today. Live so you can see the bonus episodes that are here for the pockets in the meantime. Thank you so much for sharing us here on the AI Today podcast and thank you very much for helping me on it's been a pleasure and looking forward to seeing all the various other podcasts that are coming out great. Thanks so much Gil and listeners. If you've enjoyed listening to this podcast, please make sure to write us on iTunes Google Spotify or your favorite podcast platform as always will post any artistic Concepts discussed in the show notes including the link to Gil's keynote at the data for a i comp. Thanks for listening and we'll catch you at the next podcast and that's a wrap for today to download this app. Episode find additional episodes and transcripts subscribe to our newsletter and more please visit our website at Coggin join the discussion in between podcasts on the Facebook group and make sure to join the cognetic a Facebook page for updates on this and future podcasts also subscribe to our podcast in iTunes Google Play and elsewhere to get notified of future episodes want to support this podcast and get your message out to our listeners, then become a sponsor we offer significant benefits for AI today's sponsors including promotion in the podcast and landing page an opportunity to be a game on The Today Show for more information on sponsorship visit the Cog political website and click on the podcast link this sound recording and its content is copyright by cognate. Litica All Rights Reserved music by Matsu Carrabba's as always. Thanks for listening to a guy today and we'll catch you at the next podcast.