Carahsoft and FedInsider partnered to bring together a panel of Chief Data Officers and industry leaders for a dynamic discussion around the evolving challenges surrounding Data as a Strategic Asset. Our CDO Panel discussed the burgeoning use of artificial intelligence (AI) and how it is spurring invention and innovation across many sectors, including:
CDO Roundtable: The AI Era in Government
Intro 00:14
On behalf of Cloudera, NVIDIA and Carahsoft, we would like to welcome you to today's podcast: the Chief Data Officers Roundtable, the AI Era in Government. This podcast features government and industry thought leaders who are leading the way in the AI enabled era in government.
Mike Adams 00:31
So good afternoon, everyone. This is Mike Adams. I'm a senior director at Carahsoft, and I manage our AI solutions team. Thank you all so much for joining us today for our first Chief Data Officers Roundtable Webinar Series. This is the first event in a three part series on intelligence, artificial intelligence and government and how the Chief Data Officer is helping implement a data strategy to help bring AI programs within their agencies from idea to production. Data is our most critical asset, data quality, governance and security are key to the data strategy. And data is the key to AI. AI is not just a buzzword. It's the most important technology revolution ever, our nation must be the leader for the world in developing responsible AI systems. Our adversaries are spending billions of dollars and do not play by the same rules. Over the course of the series, you're going to hear from the Chief Data Officers themselves on how they're using AI within federal, state, local government and higher education for things such as defense, health care, Smart Cities, education, and many other use cases. Government agencies who can efficiently and securely manage and leverage data have a distinct advantage. You're also going to hear from some of our technology partners in the industry on how they're working with agencies to help them meet their mission requirements. Carahsoft is helping government agencies Connect technology and industry partners with best of breed artificial intelligence, machine learning, and high performance computing capabilities. We hope you enjoy the presentation today. Please, Nick Bhatia said share your questions, be happy to answer those, you know, on the call or afterwards, we'd like today's event to be as interactive as possible. And we look forward to hearing from you guys.
Jane Norris 02:08
Thanks very much, Michael. This is Jane and I appreciate that handoff. Today, it's my honor, really, to be moderating today's lively and interesting discussion on data as a strategic asset in the AI era in government. As new data sources, data sets, Software as a Service applications, and cloud data platforms are added to today's complex and hybrid data architectures, data governance must also adapt. These adaptations and practice improvements are leading to a new form of data governance, that creates a holistic approach, where organizations are governed with consistent standards for all governance policies and rules. So we'll explore current and future data governance strategies and best practices for implementing modern data management programs. This is a fascinating and critically important topic for government as it continues to organize its data to drive and support new efficiencies and advanced technologies for government. We have six amazingly talented guests with us today to help break it all down. So let me introduce them so we can get started. First, I'd like to introduce Dr. Clark Cully. Dr. Cully is the Deputy Chief Data Officer at DOJ.
Dr. Clark Cully 03:23
Thanks, Jane, happy to be here.
Jane Norris 03:25
Hey, it's good to be here too. It's good to be with you. Also with us as John Correllus. JOHN is the deputy state's Chief Information Officer and Chief Data Officer for North Carolina. Hi, john.
John Correllus 03:36
Hey, good afternoon. Glad to be here.
Jane Norris 03:39
It's an honor. And joining us today is Tony LaVoi. Chief Data Officer for NOAA, the National Oceanic and Atmospheric Administration. Hi, Tony.
Tony LaVoi 03:48
Good afternoon, everybody.
Jane Norris 03:49
Also, please meet Brandon Pustejovsky. Brandon is the Chief Data Officer at USAID. Hi, Brandon.
Brandon Pustejovsky 03:57
Hi Jane, greetings, everyone. Delighted to be here. Thank you.
Jane Norris 04:00
Thank you. And let's also welcome Anthony Robbins. Anthony is the Vice President of Public Sector at NVIDIA. Hi, Anthony.
Anthony Robbins 04:08
Hi, Jane, a most important topic come in their way.
Jane Norris 04:12
Absolutely. And please also welcome Henry Sowell, and he is the CIO at Cloudera. Hi, Henry.
Henry Sowell 04:18
Hi, Jane. Thank you for having me. Pleasure to be here.
Jane Norris 04:20
Great, wonderful. We have some great questions coming up about government data management and data governance and how data is being used to accelerate more advanced technology and government services. But first, we'll take a moment and introduce each of our talented guests today, and learn about what they do and their areas of expertise. So let's start with Dr. Cully. Dr. Cully, as you know, as the Deputy Chief Data Officer with DoD de he also serves as the Senior Advisor to the DOJ Chief Information Officer. He led the development of DoD's data strategy and the stand up of the DoD Chief Data Officer Organization and the integration of data into the department's Digital modernization program. He also advises the DoD CIO on innovation and technology transformation issues. Prior to joining DoD CIO, Dr. Cully served as an advisor in the Office of the Undersecretary of Defense for Policy. Dr. Cully authored portions of many of the department's key strategic documents, including the 2018 National Defense Strategy, the 2018 DoD Cyber Strategy, the 2018, DoD Cyber Posture Review, and he also managed DoD programs on 5g technology, innovation, professional education, nuclear weapons policy, and command and control modernization. Such an impressive background, please tell us about your role with the Defense Department. Dr. Cully.
Dr. Clark Cully 05:47
Thanks. Our focus over the past year has been standing up a new CDO organization under the Chief Information Officer. We rapidly grown to stand up the core team and begin putting in place the data strategy which was issued last October, the implementation plan the governance structures and the key relationships needed to implement data management across the organization. So we're still in the messy startup phase. But we've we've put a lot of initial points on the board that are carrying us forward. And most recently, the deputy secretary issued the data decrees that helped provide guidance to the department on moving towards an open data standard architecture that will carry us into the future. So if you've not had a chance to take a look at those data decrees, they just came out a couple days ago, and we're proud of that work as well. So looking forward to the conversation. And thanks for having me this afternoon.
Jane Norris 06:43
And we'll talk a little bit more about that, the strategy that was just announced a little later in the program. So Also joining us today we'd like to welcome our next guest is John Correllus. He's the deputy state Chief Information Officer and Chief Data Officer for the state of North Carolina. Since April 2016. John has led the data division which is focused on leveraging the state's data assets to support informed decision making and expanding and integrating analytics into agency business strategies and business intelligence initiatives. In addition, he served as the director of North Carolina's government data analytics center within the Department of Information Technology. John, please share information about your incredible work with our audience.
John Correllus 07:27
Great. First off, I'm glad to be here today to participate on the panel to be able to share state's perspective and quite honestly, selfishly looking to learn from the other panelists. Again, I do run a division called the Data Division within the Department of Information Technology Services. The Data Division serves as the Chief Data Organization for the state of the Chief Data Office. And within that the biggest group we have is called the Government Data Analytics center. So everything about my title, everything that we do is focused on data. I'm lucky to to work with a team of about 60 professionals that are extremely dynamic, and have a passion for data and serving the state. Our vision is to be a trusted partner and leader for enabling the sharing of data assets within North Carolina. And really our mission is to transform that data into information to facilitate decision support really improve outcomes for the citizens of North Carolina. And to create efficiencies. My group is focused on pretty much all the domains in government. And we've been obviously heavily focused on COVID-19 response and activities. But we're focused on healthcare, fraud and compliance, criminal justice, longitudinal, educational, safety, performance, and the list goes on and on at any given point in time. We have 40 plus data programs and projects. And the one thing that I'm uniquely positioned because I have great support of the governor, the Secretary of VIT, and the legislature, they all recognize that whole leadership core recognizes the importance of data to improve North Carolina. So thank you, I'm glad to be here.
Jane Norris 09:16
It's great to have a representative from the state. So thanks for being here with us as well. All right, next up, Tony LaVoi. Mr. LaVoi is the Chief Data Officer for NOAA, and has been with NOAA for almost 20 years as the NOAA CDO, he's responsible for the agency's data strategy, and all aspects of its implementation. He also served as the US Department of Commerce senior agency official for geospatial information. Prior roles in NOAA includes serving as the NOAA Geospatial Information Officer, as well as a National Ocean Service Information Services Director. Tony is a member of the Department of the Interior's National Geospatial Advisory Committee, the Federal Emergency Management Agency's Technical Mapping Advisory Council, the Federal CDO Council, the Cooperative Institute for Satellite Earth System Studies Executive Council, and is active in the United Nations Global Geospatial Information Management Working Group. Tony, it's a pleasure to talk with you today. Please tell us about your incredible work at NOAA.
Tony LaVoi 10:21
Great, thank you, Jane, for the introduction and the opportunity to participate in webinar today. As you mentioned, I've had the opportunity to work in NOAA for over two decades primarily in the geospatial and data areas. I should mention, I've been in the CDO role now officially for two weeks. So I think that likely makes me the newest member of the Federal CDO community. I would say that, from my perspective, it's a it's a really exciting time to be working in the federal data sector. With all the recent data legislation and policies. Raising attention on data is critical role in government technology innovations, I think we're going to talk about a little bit today. And then specifically the opportunities for NOAA as the science agency with the new priorities from the Biden-Harris administration. Similar to what Dr. Cully talked about, we're in a similar situation in terms of having finalized our data strategy. We're currently in the process of finalizing our implementation plan and governance approach. So very much looking forward to discussions.
Jane Norris 11:32
Yeah, I'm, we're looking forward to hearing from you too. So thanks very much. Our next guest, and this is just a plethora of great, you know, informed and experienced guests with us today, s Brandon Pustejovsky. Brandon is the Chief Data Officer for USAID. He leads a broad portfolio of data services at USAID to include enterprise analytics, Master Data Management, data science, and repository curation, privacy and risk management, systems integration, data literacy and capacity building change, management, and mainstreaming best practices and data management across the agency. In his work, he directs all aspects of USAID's implementation of executive orders, that make us government data public and machine readable, while ensuring rigorous protections for privacy and security were required. He also served as USAID's liaison on data related issues related to interagency collaboration. Brandon, please tell us about your fascinating work at USAID.
Brandon Pustejovsky 12:39
Thanks, Jane. It's really great to be here. And you know, juxtapose with Tony, I think I think I'm maybe, if not the the oldest serving or longest serving Federal CDO are close to it. So and probably because I couldn't get that job as a professional fisherman. But, you know, ultimately, it's a real privilege to work for an organization tasked with improving the lives of some of the world's most vulnerable people. And certainly, we're feeling that as a global population right now faced with COVID, obviously, but everything that we do focuses on ensuring the equitable treatment of individuals who've been marginalized in some cases for years or for a lifetime. And that includes as an agency, ensuring the availability of basic necessities based on local commerce. And building on local capacity is not simply bringing external resources, external resources, or external help, quote, unquote, but that taking local talent, local capacity, and and, and helping them realize their own potential. And so it's just an incredible honor to be able to do that and bring democratic opportunities for self determination to the societies in which we work. But, you know, I think the the important perspective, from our side is that fulfilling this mission can't be about guesswork. The situations in which we work are too precarious. And the last things that a human being wants in a precarious position is another shaky place to stand. So, you know, as an agency, our data roadmap that we are now implementing, or 2021 focuses on how we really owe it to these individuals to our partner nations to bring our best data, our best evidence into the picture green times of need, and provide a solid data informed foundation on which to stand and that's what we're doing as we provide foreign assistance around the globe.
Jane Norris 14:32
Thanks, Brandon. That's great to hear. I really appreciate your your background. Let's also welcome Anthony Robbins. Today Anthony is the Vice President of Public Sector at NVIDIA. He's responsible for building and leading the federal business for NVIDIA. His team helps the government leverage GPUs and the CUDA programming model to usher in an era of artificial intelligence, working with cloud service providers federal systems integrators, OEMs, universities, Value-Added dealers and resellers and the startup community, to service the largest it transformation in the world. He's a vocal supporter for government's use of AI to address key national issues. And NVIDIA has reinforced its commitment to AI deployment with platforms for collaboration, such as GPU Technology Conference. Hi, Anthony, please tell us about your wonderful work with NVIDIA.
Anthony Robbins 15:25
Thanks, Jane. I always I always comment about the size of this technology transformation that the government's undertaken, inspired by artificial intelligence. And it is in fact, the largest technology transformation our federal government has ever undertaken. And therefore, it's, it's the most important one they've ever done. And it's a giant team sport. And so, you know, we came off mainframes and went to client servers. And that was about it, then, that was about a 10-year technology wave. We went from there to we work on mobility and government. And that took us some 10 years. And no sooner did we not really not even complete the mobility wave. But here comes cloud. And you know, across the federal government, we're still working on cloud implementation, right duties is a notable one. And then of course, here comes artificial intelligence, which is studied by many around the world. And it's an it's, it's considered the largest technology transformation, ever, and the largest versus the previous three waves that came before. So this is really--this is a really important one. And so NVIDIA has been around for some 28 years, just a couple of weeks ago, you know, we passed our 28 year anniversary. Today, people don't really know. But we're about $17 billion in revenue, more than 19,000 employees. And we spent $3 billion in Research, Development and Engineering for a set of product suites, most notably GPUs that are playing a profound role in the world's progress and artificial intelligence. In my role, I'm I'm trying to help and be a good business partner, to the federal government, the most over regulated complex business in the world. And one bit since that's in need of strategic partnership. To help it with this large technology transformation. We spent time on the hill to the National Security Commission on AI and with most people that study AI as it relates to the government. We try to bring our technology and our partners and ecosystems to this federal or public sector marketplace. Through a platform we have called GTC, which is a big it's a big conference whereby we meet the AI developers and public sector and there's more than 2 million people that participate in that program. So I'm honored to spend time with these Chief Data Officers, and then to collectively add some value to the audience with respect to the learnings that they're after and the work that they must do, relative to the government's data and how they make it work for their respective missions. Thank you, Jane.
Jane Norris 18:08
Thanks, Anthony. And then also with us Henry Sowell. Henry is the CIO at Cloudera. In his role, Mr. Sol engages with stakeholders and technology partners to develop an implement complex technology solutions that help the company's clients transform and optimize their strategic missions. Priorities appointment at Cloudera, and we serve as a Senior Advisor for the FBI for nine years. He transformed mission critical operations by delivering Big Data industry and mission-specific consultation. He also tapped into his knowledge of mission and agile methodologies to identify and prioritize product backlog by creating mission critical data applications. And we also served in the US Marine Corps, where he was awarded the Bronze Star for valor. Henry, thank you for your service. And please tell us about your job the Cloudera.
Henry Sowell 19:00
Thank you, Jane. At Cloudera, we provide did a strategy, partnership and data management software solutions across more than 130 US government agencies, and support government agencies and more than 40 countries. We support a variety of different missions, including cancer research at the state, local and higher education level to supporting federal civilian missions, the Department of Defence and very complex missions within our intelligence agencies. We offer solutions that help mission critical operations across the entire data lifecycle, and that includes collecting and managing data at the edge supporting data engineering, data warehousing, and machine learning. All of these technologies that we develop and partner with our government clients help support many mission critical innovations.
Jane Norris 19:48
Thank you all for letting us know about your work and that your backgrounds. Thank you and appreciate all the insights that will receive for from you today. So let's get started with that. Let's begin our conversation and talk first about how government is managing its data. Tony, I'm going to ask you with sustainability issues top of mind for the new administration, which you mentioned, how has that impacted your data management strategy at NOAA? And given that your mission is to understand and predict changes in climate, weather, oceans and coasts, and share that knowledge and information with others?
Tony LaVoi 20:23
Sure. Great, great question. So I'll always make right up front, that I have a bias as the CDO, that we definitely consider NOAA to be primarily a data driven science agency. It's important to keep in mind that every day, we collect observations range from the surface of the sun to the bottom of the oceans, from coast to coast, we observe, we measure we monitor we collect. We process data, we use a full range of platforms from satellites and ships to buoys, planes, drones, sensors, we have a high performance computing infrastructure. What that means is, every single day, NOAA as an Earth Science organization, is creating 10s of terabytes of data per day. Certainly the CDO perspective, what's really rewarding about this is that all of these datasets, all of these Earth systems, datasets, specifically tied to our crucial, crucial missions, and tie them together. So when you look at weather, climate, fisheries, ocean, and coastal health, and communities, those were our primary mission areas, they're all underpinned by this observational data that we collect every day. And then we take that data, and we couple it with no as science, our service and our stewardship roles. And ultimately, what we're trying to do is we're seeking to understand, predict and provide tailored information based services about climate driven changes in the Earth's environment. So that's, in essence, the frame for the CDO role within NOAA. So specifically with our data strategy, and how it impacts our ability to share climate data, we believe that we're in a pretty good position, our data strategy was just published last fall, since it's a relatively recent strategy, we're currently working on our implementation plan. And we believe that this position just really well to support the new administration's priority initiatives, especially in the areas of climate and racial equity. Within our strategy, there are five goals. Within this particular topic area, I'd say that the goals that are most important are our data governance goal, our open data goal and partnerships. So how does this help us assist us and sharing knowledge with others. So our first is the full commitment that we have as an organization to open data. Again, I think we might be a little bit unique. NOAA is essentially an open by default agency. When it comes to our data. There is very little mission data that is not publicly available. So that really kind of infuses itself throughout the organization. We are all about high quality data and getting it out and getting used. We have a full commitment to the open data requirements. They're called out of the evidence act in the federal data strategy, the geospatial data act, those are all embedded in our data strategy and implementation plan. And then ultimately, this public publicly accessible data feeds not only into Noah's climate and weather models, but also other private sector, academic sector users of NOAA as data to produce the daily, the weekly the monthly weather forecast, when you think about it, most people in the United States are dealing with NOAA data on a daily basis, because everybody thinks attention to the web. The other thing that I would mention is that we're working very hard to leverage partnerships across all sectors. We recognize that many other federal agencies, state agencies, local governments, commercial sector rely on NOAA science and data. So ultimately, we are we are absolutely committed to supporting the sustainability and the other priorities, the Biden-Harris administration and believe that our biggest strategy position is just well to be successful.
Jane Norris 24:09
It just reminds all of us about the magnitude of your mission, and you know, all the elements that are needed to support current and future goals. Thank you for that answer. Brandon, you work at an international level with USAID. How is your agency handling its agency wide data strat strategy to support your mission to equip staff to improve development and humanitarian assistance outcomes? How do those factors enter into your digital strategy and your data strategy?
Brandon Pustejovsky 24:46
Yeah, it's interesting. You know, it is uniquely challenging when you're in, you know, over 80 countries around the world in very austere environments, that are challenged in terms of infrastructure connectivity, and even Data Acquisition. But at the same time, we are passionate about meeting the frontline needs both of our staff that are working in those environments in the hundreds of partners, that US government funds to create a better life, a more sustainable way of living in a democratic environment for the countries in which we work. And that, that presents a unique challenge as well in terms of being able to bring that data into the organization that is incredibly heterogeneous, often using different standards, different data collection methods and being able to tell a cohesive story from strategy to results in a way that continues to advance the agency's mission. I haven't see a question, I think that was down there in the chat in terms of, you know, how do CDOs, you know, what, guides CDO's data strategy, and for us, your question's very timely, because it's those frontline needs those priorities and pain points that have been driving our data strategy and roadmap. In fact, just last month, our data governance body, the USAID data board, established what we call a Mission Advisory Committee. And that's to ensure that our priorities, you know, over the next 12 months are responsive to those frontline needs, whether it's related to data literacy, of putting data visualization tools, in the hands of our staff, even starting to look at advanced analytic methods that pull heterogeneous data from multiple places and give data scientists access to the raw data and the algorithms to generate their own predictive models. It really runs the gamut from from the basic to the advanced. But we'll soon even be piloting a mission support package aimed at providing a standard menu of data management support options to our mission staff around the world. I was thinking just before this, this teleconference that my last international trip actually was in late 2019, to the Philippines, which I think was fortuitous in itself in light of what was to come in 2020. But during that trip, we prioritize ensuring that our cloud based infrastructure was responsive to the data processing needs of our staff and the Pacific region. And in hindsight, this actually seems like foresight as our prioritization of cloud technologies at that time made the agency's transition to full time telework and 2020, almost seamless. So we continue to look forward to supporting our frontline workers or implementing partners with a data informed approach to how we operate.
Jane Norris 27:42
Thank you very much, Brandon. That's great information. And that like to turn to Dr. Cully. Dr. Cully earlier, you mentioned the memo that was issued just I guess this week or early last week by the Deputy Secretary Kathleen Hicks, that discusses how the US can gain competitive advantage by turning DoD into a data-centric organization. So tell us how will we ensure that there are sufficient authorities roles, structures, policies and resources in place to support the use of strategic data assets?
Dr. Clark Cully 28:15
Thanks, Jane. The first step was getting in place a strategy that provided a common vision for the department at the enterprise level. So we seek to become a data centric organization, and to use data at speed and at scale, both for operational advantage for the warfighter and the people in field as well as increasing the efficiency and performance of our business processes. Now to do that, the strategy also laid out some specific guiding principles such as making data ethics of priority, ensuring data for purpose, and a number of other enablers to treat data as a strategic asset. And also goals for what right looks like data being visible, accessible, understandable, linked, trusted, interoperable and secure. So that kind of high level guidance then led us to begin building the relationships needed to execute that that guidance and implement this vision across the department. And everyone has a role in data leaders and all components are going to touch data in some form, no matter what mission they're in, as both a producer and consumer and steward of data in their space. And so we had to work with command commanders, we had to work with military department Chief Data Officers and component Chief Data Officers, we had to work with various functional governance bodies and begin giving data a seat at the table in decision making conversations going on throughout the department. And you really can only lead at the speed of trust and what wasn't needed was more top down guidance, of some one-size-fits-all solution or architecture. We really wanted to take as befits our organization, a more federated approach based on Open Data Standards and commercial standards wherever possible. And it also needed to be a two way conversation where data leaders were talking with the users and the people in the field affected by these policies and responsible for for implementing these policies. And so part of our vision is being close to the warfighter being close to the users who need these data driven tools, and spending time alongside them to make sure that we understand their needs and take an iterative and agile approach to policy development. And, you know, use flexible guidelines that can keep pace with the evolution and performance of technology. So that that type of collaborative model we think, has already borne fruit and just the first, you know, kind of six months of, of earnest implementation. And we're gaining speed. So we look forward to, you know, working more with with our interagency and international partners, as well as we kind of build outward from the DoD community to those in the broader community that we want to share and collaborate in our national security mission with. So, it's been a crawl-walk-run. But the the recent guidance from the Deputy Secretary again, helps to translate in simple terms, rather than, you know, detailed and cumbersome takedown policy, but very simple bullets, like what exactly right looks like. And so I commend the data decrease to folks if if they want to see how exactly we're advocating for this open data standard architecture. Thank you.
Jane Norris 31:47
It's an enormous responsibility. We thank you Dr. Cully, for that information. So let's take a look at the more commercial perspective. Anthony, you have more than three decades of experience delivering information technology solutions, can you tell us from a contractor government contractor perspective, some of the unique data management challenges that your clients are facing in defense and civilian organizations,
Anthony Robbins 32:12
I will attempt to let me first comment on something that Dr. Cully was talking about. You know, I think the audience should pay careful attention to the work that's going on in his office. And with base work, the CDO, the Department of Defense, there's some wonderful work that's been published relative to handling data across the Department of Defense. So again, just though I would just encourage, you know, our industry and the listeners on this call to pay attention there. So as far as the this big challenge of data management, let me put it in context. And so we're facing the largest technology transformation that federal government has ever undertaken. And I, I said that earlier. And I think that's fairly well established. And when you think about this large technology transformation, you know, in its essence, this is about leading change and transformation. And so leading leadership, leading change and transformation, in many cases is not, you know, as difficult, as the art of the possible with respect to technology.
Anthony Robbins 33:23
And so, you know, the Department of Defense has been working on artificial intelligence, I think on record going back to 1957. And we actually as, as a world came out of the frozen winter in 2012, with respect to artificial intelligence, and so the world's been making dramatic progress on a daily basis, on most things, relative to artificial intelligence, and today are, as all of us, you know, are impacted daily by the role of artificial intelligence and the tools that we use in the homes and the communities that we live in. And so, so this is a really big change in transformation effort. And so when I think about leading change and transformation, I think about the middle managers and government and I think Dr. Cully, you know, at one point and said, Hey, this isn't as much about top down anymore. And, I'd agree that, that middle managers, middle managers in the government are these career civil servants. They care deeply, and they built the systems that are in place today. They own the budget for the agencies for whom they work, and they outlast every SES-er, and General Officer who comes on some rotation. And so if we're going to talk about effecting change, and transformation, we have to focus on the middle managers. And that's why I think it's a giant team sport. There's four areas where I think we as an industry, have to do a really good job partnering with middle managers. I think we got to pay really close attention to the people that they have, you know, relative to data scientists, some agencies have incredible assets in their employees that are focused on data scientists, others do not. But every agency needs access to data scientists to do the work, this data work that is central to the progress that they must make in artificial intelligence. So there's a big people equation of this change and transformation effort that relates to data. And of course, there's the data itself, but we're going to, we're going to talk for the next hour about that. So I'm gonna come back to that in a moment. And then there's this infrastructure question, you know, the the infrastructure that you would desire to do world class, artificial intelligence work tomorrow is not the infrastructure that we have in place today, it just is not. And however, we're not in a new environment, you know, funding does not allow, nor are we an environment or kind of rip and replace. So we in industry have to do a really good job of partnering with the middle managers in government, helping them understand you know, where their their data analysts are, and how they may get access to them, do the important data work, some of which has already been blogged about, and then contemplate use of infrastructure that they have today. And that takes various forms, whether it's, you know, at the edge or on prem or in the cloud. And then finally, you know, we got to, we got to when we thought about changing transformation, we have to pick on we have to start now, move fast, not be not be scared of failing. But we do want to model behavior that tries to, to return some organizational wins. And so we want to pick good use cases. And there are many of them that we'll talk about them I know kind of as we go forward. So you know, partnerships with industry, matter a lot. Relationships with universities, I believe matters the most, you know, in this change in transformation versus others. And then relationships with a startup companies, many of whom will not understand the complexity of government. But there's programs, you know Carasoft has a big program that helps bring startups to government, and so does NVIDIA. We have an inception program where we have, you know, where we try to help at the moment, I think it's like 7,500 inception partners. So this is a really complicated puzzle that we're putting together data is one portion of it, but it's not the only one.
Jane Norris 37:23
Thanks very much, Anthony. Appreciate that answer. So John, please tell us how state local governments are organizing their data, and what outcomes that states hope to achieve. We've been talking a lot about the federal government, but you have different goals and different data management practices. So what is it that you hope to accomplish with your your data management strategies?
John Correllus 37:44
Sure, that's a great question. And what just want to say I'm impressed with the federal partners and the data strategies that they've been rolling out and interoperability and those types of things, which are really important and do drill down into kind of how we look at data from a state perspective as well, I definitely want to give kudos to our federal partners. But, you know, we like to think of North Carolina, and it's really not like to think about it as but it really is a data driven state. So, you know, what does a data driven state really mean, for us data is part of all of the conversations in how we can improve services, and how we can actually improve outcomes. So, you know, data is itself is a strategy, but it's also part of every business strategy in North Carolina. So when you really start to embed the data strategy with the business strategy, you're now I think, prepared to start informing or getting new insight into some of the problems and issues of the state. So we look at we like to think about, you know, data embedded in a business strategy that you think about, what problems are we trying to solve? So what are the issues? What do we need to do to help solve it? Does the data actually exist? What data do we have? What tools and metrics should we apply? So the analytic? What do we learn? Right? So those are the insights when we start integrating data assets to solve problems in unique ways. And then, you know, what, how do we change which are really the actions based on what the data selling? And really some of the predictive analytics that we can maybe do? And then what are the outcomes right and continuing measurement using the data? So this is what we think of as a data driven state. These are the questions this is like, what we what we really engage with the business strategies on. So you think about is this the problem that involves criminal justice? Are you trying to help with reduce fraud and increase compliance? Are we trying to increase three safety? Are we trying to improve health outcomes? This is where data is part of all of these areas. When you start talking about data management practices, you know, we think of that as the foundation to be able to accomplish that. So really, the strategy around thinking of data as a service--API's, accessibility? How do you actually leverage that data? In areas where we're continuing to focus on inventory in the data, what do we have? How can we use it? What are the privacy concerns, making sure we address those upfront so we can get some type of outcome, usually utilizing the data. Again, I'm an enterprise organization. So I work with all the departments, but doing the entity resolution for an entire state with multiple departments that may have different identifiers doing that on the front end. So it's not Master Data Management, but it's entity resolution that gets you gets you the ability to integrate that data more efficiently. So you can get to those. Good till you get to those new insights. And obviously, documenting the business rules, the exploratory data analysis, focusing a lot on data quality and data quality programs. And keep in mind, government has a lot of administrative data that was collected for a certain purpose that maybe wasn't collected, or analytics or for whatever purpose may want to use that. So really, you know, data is very important across all domains. And the easy stuff is analytics. The hard stuff is building that underpinnings the data management practices, the data governance and whatnot. When we start talking about kind of AI, I would say that, you know, state governments are probably are not leading AI applications. But I think you'll see, you know, AI like applications, such as chat bots, robotic process automation, are heavily embedded in state and local governments. But also looking at things like machine learning how we can improve fraud models, and how we apply machine learning to fraud and, and be able to train those models through the machine learning tools and whatnot. So data is part everything, but at the underpinnings are really that data management. And that's what we continue to work on so we can continue to produce results.
Jane Norris 42:06
Thank you very much, John. Next, we turn to Henry. Henry, are there strategies and best practices and data management that the private sector is using that can be brought to bear and customized for government?
Henry Sowell 42:18
Absolutely, there. One, it's been great to hear of some of the successes here and John, what you all are doing in North Carolina? It's, it's been interesting to see the advances in the government's view of the importance of data strategies. You know, we've seen a lot of changes over the last 10 years, where just until recently, you would see focus on solutions versus overall strategy. And you'd hear things like, what's your cloud strategy, right, and that that would be a primary focus. And now now, we're seeing things that it is, especially over the last five years significantly, shifting to focus on data strategies, where building data, folks cultures is critical, you know, focusing on governing, managing and protecting your data, and also all to drive efficient, appropriate use of that data to support your mission. Right, that is a transformational shift that we see. And and what we're seeing in the private sector as well. And you know, once once you have that cloud and other technologies, then support that overall strategy. In their critical areas that need to be considered. Where's the data generated? How can I support your mission, one of the governance and security requirements within your organization. And once you start understanding those areas, you can start considering what solutions are the right fit. Doing this out of order, often is where we see a lot of challenges. You know, for example, we've seen police solutions deployed that are addressing a point in time mission issue. But those aren't then capable of meeting complex data governance and security requirements. Or they require significant developments to integrate. And these are extremely, extremely fragile, as government missions evolve, whereas developing a solid strategy allows you to consider these concerns up front. And as you've heard from NOAA, USAID in the state of North Carolina, many government organizations are doing a great job executing this.
Jane Norris 44:18
Thank you. Thank you very much. I appreciate answers from all of you. So let's move from you know your current state to how you identify data needs and maximize innovation in your organization. So start with you on this one, Brandon. How have you incorporated feedback from your internal stakeholders at USAID throughout the data lifecycle? And has it helped you to identify their needs and maximize entrepreneurship and innovation?
Brandon Pustejovsky 44:49
Yeah, thank you Jane. It's a fascinating question, because it presumes that there's an understanding of this thing called the data lifecycle already and so we're kind of back to the The question of culture change, right there can be so much focus on the end product. And that's important the analytics, the insights that can be driven by the data and the technology themselves. But a lot of our work, particularly within the past two years, I would say has been focusing on culture change, and building awareness around this thing called the data lifecycle that you're not going to get the product that you want, if you don't engage in sound data management planning off the bat, if you don't, you know, back to, you know, Clark's interoperability discussion, if you don't plan for interoperability, and start to establish data standards, metadata catalogs, you know, master reference data and that sort of thing, then your data collection practices themselves are still going to be disjointed, and not provide, you know, the optimal raw material for advanced analytics, you know, machine learning, etc, down the road. So there has been a lot of work that we've done on that. But back to your question in terms of, you know, responsiveness to internal stakeholders, you know, there there will continue to be, you know, probably this demand for basic data visualization, charts and graphs, etc. But it has been interesting to see the growth of demand around predictive analytics and pushing the flow, and the desire to take simple data insight from what is happening now, to start predicting what may happen in the future based on a variety of variables. We might be, you know, programming in the southeast region of a certain country now based on famine, but what are the crop trends look like right now and five years into the future juxtapose with rainfall rates juxtaposed with potential population movements, etc, that may dictate where we should be later on, and how to not simply be responsive or reactive. But to be proactive, and how we plan where will the next outbreak of a certain disease take place. And so that means that we are starting to integrate some of this heterogeneous data from across the organization into a single advanced solutions that can respond to threats related to famine, disease, natural disasters, and so on. And I see that somebody actually mentioned that the chat will, how do you break down these silos and, and part of the answer to that is, is for our discipline, showing that you're famine data that does relate directly to this health data set, because you know, nutritional status affects health and vice versa, and health, you know, the health of individuals affects their ability to continue farming or to continue being productive individuals within the economy. And that affects our focus in the economic growth sector. And as somebody said, previously, this is absolutely a team sport. And we have to continue to propagate that message across the organization. Finally, I'll say that, you know, for USAID, our data is gathered by so many partners around the world, that we are also looking at very important work on a contractual side that says, you know, if you're gathering data with taxpayer funds, you know, we ask that you gather it according to these standards, if you plan accordingly, and that you provide it back to USAID in specific digital formats. And that is part of the process we're going through right now to really make some very significant changes contractually, and moving through the federal rulemaking process as we do that, to ensure that the public and our partners in general, are participants in that process, and not just a top down, sort of here's how we're going to operate we very much respect and welcome their feedback in that regard. Over
Jane Norris 48:40
Anthony, as government continues preparing it stated incorporate AI and other emerging technologies. And as Brandon was just talking about, you view it as a team sport. And so my question is, how does that collaboration take place across industry, academia, and other entities, to make sure the government has the skills and talents to take on this immense opportunity?
Anthony Robbins 49:07
How does that happen? Well, I think we have to acknowledge, not easily, right, and not without extraordinary leadership, but it's something that has to be done. And I'm going to go through a couple of the questions I saw and I'd like to maybe take a shot at answering the questions and then Jane answering your question at the same time. So when the strategic capabilities office you know ultimately then, you know, stood up the effort of Pentagon, right, stuff, the effort that became the joint Artificial Intelligence Center, there was a really important study that was completed by the RAND Corporation. And it was titled I think, in part that the DoD readiness for artificial intelligence and or no--the DoD posture for artificial intelligence, assessment and recommendations. That's from memory, and I think it's pretty close. And so in there, in that study, just like so many studies that have been done, we have the answers, you know, we have the answers. Right, this thing doesn't need to be studied, and more it needs to be acted on. And one of the things that the RAND Corporation made recommendations for, was the government outreach responsibility relative to artificial intelligence. And kind of the importance there is noting that the innovation community, whether they be in Silicon Valley, or Austin, or Boston, or anywhere else in this country, and for that matter around the world, that that the government that the the in the government needs to do it do a better job in the case of this technology, transformation of communicating to industry. And so I just watched it, I watched with impressed excitement, the work that's happe ning at the Jake, you know, they're really active on LinkedIn. They have AI conferences with amazing speakers, their upcoming conference just announced four-star General Clarke, keynote, former Dep. Secretary of Defense, Bob Work, you know, and many others. And so they're, they're, they're doing conferences, which are technology symposiums and interchanges. They're adopting modern methods of communications to industry, and through contracts like otas. You know, industrial, commercial companies, industrial partners can kind of come in to the government perhaps a little easier than yesterday. Right. And then the other thing that has occurred has been a really good model is, you know, there's been several cases where the government has actually released sample data. And they've created some challenges for universities and industries, to work on sample data sets to do work on on behalf of the government. So I think there's, you know, part of this collaboration, and it always comes down to this is, you know, as improved communication. And so I think we've seen evidence of improved communication between government and industry, there's probably not going to be any really good enterprise wide adoption and deployment of artificial intelligence, that doesn't include the roles of universities. And, and so the young, in career talent that exists in universities, and some of the wonderful work that's happening in universities, is really important for the government to get access to, you know, you see the Airforce, doing that with MIT, you see the Army doing that with Carnegie Mellon. And I'm sure there's hundreds of other examples. So I, you know, I always kind of would encourage, you know, the government institutions, to have a program that reaches out to universities that don't align to their mission goals and responsibilities. And then industry partners. You know, we have a big time responsibility here. And I often say, you know, you listen, you guys know, on occasion, you know, people say the government moves too slow. And I think, you know, that's just kind of nonsensical. You know, I think that the fact is, the government is the most over regulated and complex business in the world. And if we care about being good servants, to the mission, that they're responsible for, that we will care about being excellent in our craft, and helping them drive this change in transformation effort. And so so we as industry, have to take, in some cases, a primary role for helping government move faster. And this case, and I think the way that that happens is, you know, you think big, you start small, you move fast, and you're not scared to fail. And then we share as a community, right? There's not going to be one partner, that's going to be the winner of the government's AI journey. Right? It's going to be a giant community effort. And it's going to be incumbent upon all of us, as taxpayers and as citizens to do good by our country in support of the largest technology transformation. So it's complex. Jane, hopefully, I gave you some insight and some insight to the audience. And I'll look for questions in case I missed anything over.
Jane Norris 54:36
Thanks, Anthony. I appreciate that. And Dr. Cully, this one is for you. The Department of Defense is such a vast organization, using its own data to increase accountability. So how is DoD using data to support informed decision making? And how is the process of data sharing progressing among the many Department of Defense entities?
Dr. Clark Cully 54:59
Thanks. The Department of Defense is undergoing a big transition starting with the leadership at the top level. The Deputy Secretary has made great strides in transforming our decision making processes to use data to answer key decisions about how we're managing our organization. And therefore, decisions decision, quality analytics, and metrics are displacing static and legacy approaches using, you know, PowerPoint slides and, you know, outdated ways of funneling information through a lot of filters before leaders can see it to a much more rapid, transparent and hopefully, objective basis for informing decisions about resources and other strategic alignments. So the executive analytics work has started us on a journey of saying what are the strategic questions we want our data to answer? And what authoritative data sources best address those questions? And then how do we improve the data quality of those authoritative data sources? And how do we make that data accessible to everyone? And so very quickly, leaders using data to make important decisions, gets the whole organization focused on solving data management challenges, and pulling things into a common platform and making the visualizations that tell the story that underlie the data. It's, it's been a powerful journey that we're on, and it's been great to see how as a result, we're able to notice some second order effects that might not be as apparent in the legacy process in order to more quickly discover how exactly some of the inputs that we're proposing in various policy decisions might affect the strategic metrics that show the effectiveness of those decisions. And, you know, we're excited that that this effort is, is gaining speed, and allowing us to ask better questions and kind of creating a virtuous feedback loop and more and more parts of the organization are now getting connected, getting on board wanting to use this tool to answer similar questions at their level of the organization. And so there's there's kind of a virtuous cycle of trickling down from leadership at the top to other component heads, who are making data driven decisions in place of just, you know, sitting around the table and talking to bullet points.
Jane Norris 57:26
It is amazing. I mean, an organization as big as the Department of Defense, that's a lot of connections to make. So thank you for your insights there. Henry, when you were with the FBI, you won an award for technical excellence, tell us about unique data challenges for the intelligence agencies, and how they are handling them.
Henry Sowell 57:46
Thanks Jane. So there's a lot of things that they share similar with the rest of the government, they have similar data challenges and strategy needs, they often are working extremely critical missions that in some cases are life and death. And these, they're working in environments that akre often disconnected. Because of the the complexity of their environments and their security needs, they're often working with a subset of the tools that are available to those in other environments. And the data that they're working with is highly sensitive, and they often have complex security and governance requirements. So I think someone in the chat asked about, you know, there are there are parts of my organization that are not allowed to see all of our data or a subset of our data. And so how do you go about addressing that, and you have to develop solutions that consider things like attribute based access control, have well defined governance and security policies to help mitigate and accommodate those requirements needs. So that, you know, in some cases, you have part of your employees that aren't even allowed to know the existence of other datasets. So you have to have solutions that are able to meet and address your data strategy within those environments. And then I think, you know, on the message about the, you know, subset of tooling, often this comes from, you know, the vendor and industry world that maybe doesn't have the cleared staff, that have developed in such a manner that can make it available to the intelligence agency. So a continuing partnership and close work with the industry is necessary to help bridge some of those gaps and make more of those tools available.
Jane Norris 59:36
It just shows you and points out all the different, you know, use cases that the government has and all of the different needs that they have to address, you know, in their data governance policies. Tony, same with you, I'm sure while the federal government leads in many instances in developing and providing data about the United States and the world. Is there an integrated approach to using data to deliver on mission serve the public And steward its resources. Does the federal government do a good job of providing its data to other entities in government and other outside agencies?
Tony LaVoi 1:00:08
Yeah, thanks, Jane. So I'd say from the NOAA perspective there while we really believe that each individual NOAA data set is valuable on its own, because it's, it's based on a specific mission requirement. We absolutely recognize that the power of our data is oftentimes only increased through the integration of NOAA data with other data sources to support a specific use case. I've mentioned that in our data strategy, supporting data integration is a top priority. And for us, what that looks like is first a commitment to open data access. Second, is the use of strategic partnerships. And third, and this has been somewhat of a lesson learned over the course the past five years is providing the subject matter expertise on a particular data set, we have people with no order to spent their entire careers on a particular fisheries data set, for example. And we have people outside of the organization that are interested in that data. But oftentimes, they need access to the neuroscientists to be able to ask questions, to understand better appropriate news, potential user communities, additional integration opportunities. So I would say that probably maybe the best way to talk about integration from a NOAA perspective is just to give you a couple of examples of how it's happening within our agency, probably the best known example of data integration is our long standing in extremely productive partnership with the private letter enterprise, which I think many of you know response is well over a $10 billion a year industry and only growing and is now starting to include insurance and the reinsurance industry. And that is really underpinned by a wealth of NOAA weather and climate data. In a new area that is coming to us our two way data sharing agreements. Some of you probably seen within the new administration, a renewed focus on offshore wind, energy development. And there was some recent crests that NOAA has signed a two way data sharing agreement with one of the major wind developers and we see this as is absolutely a win win, because we're sharing our data, our science, our expertise, with industry, who needs it, they in turn, are sharing data with us that then we can use within our mission. And then we can provide access to that data for others to take advantage of. So we're definitely interested in pursuing more of these types of agreements. And then the last thing I've mentioned, and some of you might be aware of NOAA's Big Data program that started as Cooperative Research and Development agreements about a half a dozen years ago. And the whole concept behind them is to provide public access to know as open data through the cloud service provider community. And it's evolved over time and graduated from craters into contracts with Amazon, Microsoft and Google, we now have about 11 petabytes of NOAA's environmental data spread across three CSP's. And that data then sits alongside other federal data, state data, private data, international data. And because it's in the cloud, what it's allowing for is the integration of NOAA data, with other data sets on these cloud platforms, people are bringing their compute, their models, their algorithms, instead of you know, the old way of pulling data down to your on premise infrastructure, they're doing the compute. We're seeing significant increases, the return on investment of the particular data sets, new jobs being created, improved analytics. So we're fully committed as the Open Data agency. We're doing everything that we can. And we are thrilled every time we hear a new story of somebody out there taking a NOAA data set and using it for some purpose that we really never imagined that it would be useful.
Jane Norris 1:04:15
That's great. And I'm sure you'll get a lot of takers on, you know, integrating with you or participating with you. Thanks very much for that, Tony. John, I'll go to you. Although these goals that we've been talking about don't specifically apply to state and local organizations, you contribute to the overall federal data collection process in your reporting for North Carolina. And you also have reporting mandates. How do these kinds of initiatives impact your state?
John Correllus 1:04:40
Yes, sure. I'm somewhat lucky. My organization doesn't have any federal reporting mandates, but of course, the rest of my state agency partners do, but, you know, I really feel like the goals are very similar to what you'd find in data strategies across the country and all The state, it's about data quality, it's about interoperability. It's about understanding the use cases, beyond the friend front end of how data can be used. And like Tony's talking about, you know, some integration opportunities that you may not realize when you start to see how people are using the data. But what we're really talking about is the ecosystem. You know, the need for North Carolina to even you know, receive data or use data doesn't stop at North Carolina's boundaries, and we're all contributing to this ecosystem. So I'll use an example. And this really goes back to the data quality. You know, if you think about like Nick's guns check, each day contributes those people that have disqualifying offenses to maybe purchase a firearm. Well, we're all contributing to that. And then we're all utilizing that to ensure that we aren't necessarily providing guns or providing approval for guns to those that really are not supposed to have them. Sex offender registry, all these different data sets were contributing to and we're trying to contribute to them with high quality data, because we're also the beneficiary. But again, these goals may not seem like they directly impact us, but they actually do.
Jane Norris 1:06:15
Thanks very much, john, I'm sure they do. And I'm sure that your data governance policies are growing and changing as each new federal government mandate comes down, and your member agencies are going through the process of trying to respond to those requests. Thank you for your insight there. So I'd like to talk next about the plans for the future, and how data management helps lay the foundation for more advanced technologies like AI machine learning other areas of advanced technology work. So let's start with you on this one, Henry. Data is a vital component in helping governments, healthcare organizations and other sectors move to more advanced technologies. But not all data is created equal. And datasets are often incomplete. I think I saw a question about this in the chat window as well. So how can organizations resolve the data alignment process when you're using disparate data sets?
Henry Sowell 1:07:15
Thanks, Jane. I think most folks will agree this is where effective data governance comes in. You have to understand what data sets you have, whether it's data cataloging, or some other mechanism where that came from, and what's the veracity. This all feeds into the data engineering, which produces effective data sets that you can then do machine learning and AI. without some of those foundational pieces, you can't be successful. And often you would be making decisions and gaining insights on your data that are not verifiable and not accurate. So that is actually very critical for government agencies to get done right as that foundation.
Jane Norris 1:08:01
And that foundation, I think, is probably something that's been discussed for quite a long time. So maybe you can tell us, Tony,about NOAA's AI and machine learning strategy, and the data strategy that supports it, which aims to dramatically accelerate these of data across your agency, and with other key partners. How does this impact your work?
Tony LaVoi 1:08:22
Sure. Thanks, Jane. So we've talked quite a bit about the data strategy that NOAA developed. But I would mention that data strategy was actually one of six science and technology strategies that were developed over the past year and a half. So in addition to our data and cloud strategy, we have a strategy on artificial intelligence, one on omics, which is basically genomics, one on uncrewed systems. So think of things like drones and gliders, and then one on citizen science. So from a CDO perspective, you know, we're already--it's tough for me to judge, we are definitely engaged in AI, how we compare ourselves to other agencies. In terms of adoption, I'm not exactly sure that really our focus is expanding and accelerating the use of AI across the agency. From a CDO perspective, you know, I look at the ever increasing volumes of data, and in looking at AI and machine learning is, is a helpful tool. In terms of the alignment between the AI and the data strategies, we have the best of the fact they were developed at the same time. So they're very, very tightly aligned and fully recognizing the AI needs, you know, feeds off data and the more data we got, the better beginning AI gets. So as we move more and more and more data in the cloud, and are able to take advantage of AI machine learning capabilities, we think we're gonna be able to get out in front of some of the massive volumes of data that we're working with. The other thing that I've mentioned, two other things, actually. One is actively supporting the development of AI ready data standards, a lot of interest in our community on what exactly constitutes AI data, the maturity of AI and machine learning data. So we're participating in a couple of different working groups one sponsored by OSTP. And last, I mentioned, is workforce development, incredible interest in AI and machine learning in the organization, and we are trying to do everything we can to enhance the skills of existing staff through training, professional development, as well as work to bring on new partners and new staff to help us continue to advance our AI work. So back to you, Jane.
Jane Norris 1:10:42
Thank you, Tony. I appreciate the the information because each agency has their own I'm sure, new project that they're developing and working on and working towards, including the Department of Defense. Dr. Cully, the Department of Defense has been talking about multi-domain operations for some time. So what advances are being made and need to be made in data management and computing systems to enable that kind of advanced capability?
Dr. Clark Cully 1:11:10
Sure. Our data strategy is fully integrated with the department's broader digital modernization strategy. And that strategy includes enterprise cloud capabilities, artificial intelligence, modernization of our command and control systems to include 5g capabilities, and cybersecurity. Our emphasis within multi domain warfare or joint warfighting more generally is ensuring interoperability primarily through a federated approach that uses an open data standard architecture to ensure that our sensors and shooters and command and control nodes can all communicate. We also believe that compared to some of the authoritarian regimes that we confront, we need to make sure that our data systems serve mission command and empowering those at the tactical edge with the ability to make decisions that have strategic impact. We think that trusting our people in harm's way is a way to enable decision support tools and other things to serve the human judgment of those who are forward and therefore have have shorter command and control loops in more disaggregated command control networks, then our adversaries can match, and that'll take things like having advanced storage and compute capabilities, forward and at the edge that are able to operate in contested and denied environments. But there's a robust series of demonstration and experimentation programs underway to develop these capabilities. All the services and military departments have promising programs to try and fuse together appropriately they disparate information technology systems across their components. And we're working closely with them to pull these together into scalable data systems and a scalable data architecture that can serve the needs of a Warfighter. So we're still learning about what systems are sufficiently proven that they can be rapidly fielded, but we're looking to deliver results in the next year or two, where the warfighters and the commanders that are forward have data enabled technologies that will really be critical to the future survivability and lethality on the battlefield of the future. So we want to treat our data more like ammunition as a critical mission enabler, and to give it the appropriate priority and attention that we do to any other elements of our weapon systems. And we're on that journey now.
Jane Norris 1:13:50
Fascinating work, we could spend an entire show talking about just this one aspect. Thank you very much for those insights. Anthony, we'll move to you. How can government solve some of the data challenges associated with moving to more complex systems like those that Dr. Cully was just talking about; With limited budgets and a limited number of data scientists, do those factors limit getting to the next level?
Anthony Robbins 1:14:13
Well, I'd like to share an example with you of what I think has been amazing work inside of a government agency. And so we recently published some news about some work that's been occurring at the United States Postal Service. And many of you guys will know intimately about their operations. Others may not but but they do process about 40% of the world's mail, they process about 129 billion pieces of mail a year and they do that across about, you know, something like 195 Mail Processing centers. So it's a very big and very complex organization, and it's full of the things that the speakers today have talked about, you know, They have they have an apparatus in place today and a system in place today that has been collecting data for years. Well, what is the condition of that data? How usable is it, you know, and is it labeled, you know, with respect to doing AI work as it relates to Mail Processing. And so they created some benchmarks to kind of explore the art of the possible, the benchmarks proved that they could do some mail sorting work quite well, especially related to packages, where they process 7.3 billion packages a year, 20 million packages a day, and 231 packages a second. And so they ran the benchmarks and they determined that they could do some processing at these Mail Processing centers, augmenting infrastructure that they have in place. So no rip and replace, using data that they have been acquiring through their processes that have been in place heretofore. And so then they put some new data center equipment in place, just two data centers nominal investment, they built and created seven AI machine learning algorithms, where they did the neural network training in the data center. And then they deployed those algorithms to the edge. At 195 centers. And they did this in a year, one year's time during COVID. And with without a massive investment, and they, you know, they work on some data infrastructure with with Dell, they worked on acquiring edge infrastructure with HP. They worked with Accenture, for some, you know, some work on data governance and infrastructure and the like. And I just thought it was a beautiful example of a government agency leveraging infrastructure that they had in place with leadership that was committed to driving innovation, and they just did something now. We've shared that information publicly. It's easily accessible. I think it's a wonderful story of innovation in government. And that's one example of how government agencies, you know, might be able to do this kind of work, Jane.
Jane Norris 1:17:14
Thank you. Thanks, Anthony. And, Brandon, we'll go to you next. What about at the international level? How is USAID administering innovation through its agency wide digital strategy?
Brandon Pustejovsky 1:17:25
Yeah, thanks Jane. You know, the first two years of our digital strategy are really focused on establishing the baseline and expanding the agency's Digital Development Foundation, through research of existing capabilities, but also, based on that research, creating tools and resources and knowledge projects to help execute on the strategies objectives. In year one alone, you know, we've created a digital ecosystem framework that includes an assessment tool for our missions around the world that we've piloted in Colombia, Kenya, Serbia, and Nepal, for example, to help them better understand their own digital ecosystem and how to how to best plug into that and how to address gaps. We've put digital development advisors in our missions as well to provide guidance on anticipating recognizing reacting to changes and opportunities in the country ecosystems of our, you know, 80 something missions, we have 16. Now, in the first year of the strategy that have those advisors, we've created knowledge projects, products on COVID-19 on digital development, closing the digital gender divide, for example, digital payments, and cyber security. But then your three through five are about feeling responsible investments, to support open, secure and inclusive partner country digital ecosystems. I'll say quickly as part of this strategy, we do continue to pursue our commitment to openness and data transparency, that includes releasing valuable data assets to the public on a variety of topics, including economic growth, early childhood education, climate change, maternal child health, food security, and so on. We are still doing that through the development data library that you can visit at Data.usa.gov over the past few years, we've seen that inventory grow by about 30% annually. And finally, we release data on US government-wide foreign assistance via our foreign aid, explorer website at Explorer.usa.gov. I'm pleased to note on this call that we have partnered with our colleagues at the State Department to merge this site, in fact, with Foreignassistance.gov, prior to October of this year, that we can provide the general public with a comprehensive and holistic view of how their foreign assistance dollars are being spent. Back to you, Jane.
Jane Norris 1:19:42
Thanks so much, Brandon. Appreciate that. And John, I guess we're going to give you the last opportunity here; Smart Cities. That's an initiative we've been hearing a lot about from a variety of different programs that cities for the ins for instance, the city of Charlotte in North Carolina, Microsoft have identified five focus areas for pilot programs that leverage technology and education to benefit residents. What are you doing in that area? And how has it impacted your data strategy.
John Correllus 1:20:15
So I would say it hasn't impacted my data strategy. And I don't want to misrepresent my knowledge of what Charlotte's doing both Charlotte's focusing on really those five areas, which are upward mobility for residents, which is really education, smart transit systems, public Wi-Fi connectivity, public safety, infrastructure, and safer neighborhoods. Those are kind of the goals to improve. You know, when you think about Smart Cities, there's other city, Raleigh, the mother or smaller towns, Durham, and all that in North Carolina, are really invested in Smart Cities. The government's also invested in smart government. And as we talked about--and smart states, excuse me; we're really talking about data problems, right? So it's all about the data, how do we improve. And then when you start really thinking about Smart Cities, it's really engaging in partnering, you know, citizens or residents with government and businesses, as really, everyone is supporting and trying to achieve these goals together, which is really important as a community thing. And that's where I think the state state government needs as a lot to learn from what cities like Charlotte and Raleigh and others are doing across the country. But when you think about the data problem, the Smart cities and things are focused on "data, data, data", it's, I think we heard it earlier in this conversation, the infrastructure may need change. So you're now being hit with so much different type of data--are you prepared for that? What do you do with, and what do you make of all this data that comes in? A lot of this data is real time sensor data with real time actions that need to occur. So when you think about AI, this is probably an application in government of AI, that really is going to be very important as a Smart Citie application. You also have to think about, you know, one of the big implications is around privacy concerns. You know, the the implications are you're trying to share, you're trying to inform, you've got to be careful to make sure the data governance is covered. So you know, you aren't exposing data you shouldn't to whether citizens or whoever. But this idea of real time intelligence, is really where AI is gonna fit in nicely. And again, I don't want to say I'm an expert and Smart Cities. But again, North Carolina and across the country, there are some really good examples of some things that really the state and federal government really needs to be paying attention to, on what these communities are doing, and how they're engaging with their residents and businesses to solve problems together.
Jane Norris 1:22:55
Thank you. That's great information. This has been such a fascinating hour, I'm--hour and a half--and I'm shocked that it's flown by so quickly, so thank you for all of your insights. Tony, john, Brandon, Dr. Cully, Anthony, and Henry, we really appreciate you talking with us today about the data as a strategic asset, and the Era of AI which is already here. I know I learned a lot today, and I'm sure our audience did too. So let me turn things over to FedInsider's Claudia Hosky for some final thoughts.
Claudia Hosky 1:23:28
Thanks, Jane. And thank you to our audience for all of your participation. I see a lot of questions and comments for us to follow up on and we appreciate them. Please remember to visit Fedinsider.com to see what's coming up, and to register for multiple webinars at once. You can also subscribe to our weekly emails to see what's coming up, and check out our podcast, “Feds At The Edge”. And with that, this is FedInsider, signing off.
Outro 1:23:56
Thank you for listening. If you'd like more information on Carahsoft's AI solutions, please visit www.carahsoft.com or call us at (888) 606-2770. Thanks again for listening, and have a great day.