Interview with Oki Mek, Special Advisor, HHS


Hello and welcome to the AI. Today podcast I'm your host Kathleen Walsh and I'm your host Schmeltzer. Our guest today is okay. Mack who is the senior advisor to the chief information officer at the US Department of Health and Human Services? Hhs So okay thank you so much for joining us on Ai Today. Thank you for having caffeine wrong. Yeah thank you so much for joining us. You'd like to start by having you introduce yourself to our listeners. And tell them a little bit about your background and your current role at HHS. I am a. I am the senior adviser to Sasebo. And I'm also the technical integration. Lead but edgy reimagined but my background is really cybersecurity but my principal around cybersecurity. That just knowing cyber alone is not enough. You really need to know and understand the business I've been with. Hr FINANCE BUDGET GRANTS ACQUISITIONS. I was leaving sponsored to help onboard new employees for personnel security and badges and laptops. So I find that to be very critical and crucial in terms of modern nine station in the government just knowing to three sixty just understanding what you know what budget stream is coming through. What is the budget calendar Audi acquisition services and products and even just getting people on board with badges an laptop? Just getting that three sixty view and been very critical in terms of you know modernizing trying to innovate in trying to change the landscape Embracing Emerging Tech and dealing with big data paradigm shifts excellent. Yeah it's definitely the government's been very interesting. Adopter not just technology in general but especially with artificial intelligence. Because you know is is a transformative technology. Just like You know many of the other big transformative waves in the Internet and mobile and big data and the cloud and now five G. and then blockchain so all of these technology transformative right. They had changed a whole lot of things or multi. System the multidisciplinary and so obviously there are many ways that we see being applied to government agencies organizations. And we've been very impressed. One of the things that been talking about here a lot today is sort of the many ways in which is being adopted. Not only an industry as a whole but especially in government very impressive actually some of these ways that hopefully it's making the government more responsive more efficient more effective more. Give it more visibility into things. So can you tell us some of the ways that? Hhs is currently adopting AI and various solutions. And maybe how those ways are unique to. Hhs or perhaps similar to what other of different governmental agencies are doing in this space so we start of area to the CIO. And I started a program like celery leveraging ai machine learning and blockchain. It was the first watching network that had been authorized to operate in the federal government but in terms of a I think we are using similar to agency mainly we use it to clean and format the data and as you know you deal with a lot of data set the day. I need data. The biggest thing is cleaning the data. Ninety five I believe ninety percent of it is cleaning the data. People WanNa do they WANNA do machine learning they WANNA do. But they don't focus on they're gonna up a root of baking because Clinton. The data is the biggest component of that process. So we use a machine learning to basically normalize the data from different data sets using supervised and unsupervised learning to normalize the data. And then we also get into linear regression as well in terms of predictive analysis. One of the main thing that we'd do accelerate is looking at prices paid and just as so large we do about twenty five billion in spending on products and services just minding data and cleaning the data at it was a big call just to look at. Why are we buying things that different prices and a good example is? I'm just throwing the example while we buy that. Dobie pro at cms for aided aqua- eighty dollars per license and buying that CDC thirty dapper night or licensed an opportunity to mind data to come to the table to be able to negotiate a different price. You can't come to table to negotiate without having that insight so data claiming and looking at data mining and looking at predictive analysis the three main usage for a for us. That's great you know Melinda. We produced a report last year. And then we did a follow up report this year Dataprep prep and data labeling and I think that a lot of people underestimate how long it's going to take to actually get their data into a usable state so it's great that you pointed that out because I think that people underestimate the time and the difficulty that sometimes it can be actually get data in a usable state. Data's the heart of AI. So you need that. For these systems to learn as a government agency adoption of new technologies such as artificial intelligence can bring its own unique set of challenges that sometimes the private sector doesn't always run into. These can be issues around privacy data usage. What can and can't be used where it can be stored so can you tell us some of the challenges you've seen with Ai. Adoption in your agency and how you're overcoming them by the biggest mistake that people do is looking at the technology before they look at the business and the mission of agency. I always say that does no such thing at it. Project and the business project with it components. You have to start with the business. And what are you trying to saw? Most of these challenges that the EPA my experience I've been in a government close to twenty years half awaiting contractor in private industry and also have in federal government. The challenge is not the technology challenges to culture is leaning chain change management and. I don't think we do not a strategic planning. Just because we have one hundred idea doesn't mean we pursued ideas. We need to look at you know strategic planning comes into play. You may have one hundred ideas but you should only pursue one idea through strategic. Planning you know. The first thing is really feasibility studies. I didn't even feasible to at any regression issue. Anani downstream impact. Accountability is big. I think in terms of trying to modernize. Can you prove a concept just because you have this idea? People might think is a crazy idea but if you do it at a tangible way that you could prove that idea is feasible and it's scalable because you wanNA start small and scale that big as well. Sustainability is huge as well just because the project is a success. Doesn't mean that sustainable as your culture your agency your privacy mature enough to take on this new shining toy you know this new technology but the important part is marketability I we. I can't overemphasize that did not marketability is people. Don't think that the government market but you have to market you ideas. You have to win. The hearts and minds of the workforce the people that They coders you have to be able to market in different ways to senior leadership to middle managers workforce. I think culture is the biggest obstacle. But I think doing a thorough strategic planning analysis and putting emphasis on marketability analysis is key and the biggest part of marketability is human design is really engaging the workforce I'll model in terms of accelerated that. We're not building the solution. We are allowing the workforce to go to solution getting them engaged yet here at thoughts and pain point incorporating agile depth. Bob You know building something. Every two weeks and bring him back to them is a Saab issue and we cycled out every two or three weeks and that really is the key of getting people to lean in and to win the heart and mind of people.

Coming up next