Discussions with DTRA: Episode 8

Relevant Links

• Defense Acquisition University
• Digital University
• MIT Horizon

Transcript

Announcer:

DTRA Public Affairs presents discussions with DTRA, where the Defense Threat Reduction Agency brings together subject matter experts to discuss meeting today's challenges of WMD and emerging threats, increase awareness and deliver morale boosting information. Our moderator today is Dr. Michael Howard, a Senior Program Manager with DTRA Acquisition Management Department. And now today's show.

Dr. Howard:

Good afternoon, DTRA. Dr. Michael Howard here today with Dr. Diana Gehlhaus, Senior Talent Advisor for the Chief Digital Artificial Intelligence Office at the Pentagon. Welcome, Dr. Gehlhaus.

Dr. Diana Gehlhaus:

Thank you. Great to be here.

Dr. Howard:

Yeah, we're delighted to have you today, Dr. Gehlhaus, let's take just a few moments to share your background with our audience today.

Dr. Diana Gehlhaus:

Oh, sure. It has been a great journey so far, and I'm just getting started. I'm an economist by training and a policy researcher, and I'm here at CDAO as a Senior Advisor coming by way of two think tanks, one, the Center for Security and Emerging Technology or CSAT at Georgetown University where I designed and led a line of research on AI education in workforce. And also from the RAND Corporation, where I'm still an Adjunct Researcher and I worked on all things DOD, manpower and personnel issues.

I've also spent six years in government at the Export/Import Bank and the Bureau of Labor Statistics, as well as another think tank here in DC. So I've been thinking about AI education and workforce issues, DOD manpower and personnel issues, and I really find a lot of joy in all things tech and talent because the economist in me knows that that's what's driving innovation and growth.

Dr. Howard:

Well, we're delighted to have you again. And as I've been engaged at DTRA through the artificial intelligence machine learning data science working group, we've been engaged with your office over the past year initially with the Joint Artificial Intelligence Council, the JAIC, and now the CDAO.

As I said, we've been engaged and really seeking your guidance along the way and your artifacts. And as we go down this path of establishing our own artificial intelligence machine learning and data science journey at the Defense Threat Reduction Agency. So let's get into the questions because we've got people who are really interested in what you have to say.

So we're talking about growing in AI ready, DOD workforce. The National Defense Authorization Act of 2020, and the National Artificial Intelligence Research and Development Plan, established a US policy to prioritize building an AI capable workforce. These workforce development policies emphasize an AI education strategy for the DOD as an important step in ensuring that the military can win future conflicts against peer competitors. So Dr. Gehlhaus, what is the CDAO's role in developing a DOD AI ready workforce?

Dr. Diana Gehlhaus:

Thank you for teeing this up. Let's absolutely get after it. We have a really big role, bottom line. Let me level set a little bit as to why. The CDAO, as you noted, has been set up just over a year ago and really to get after advancing data and AI capabilities and their deployment, their responsible deployment across the department. We're calling it Digital Transformation. We know that this is what we need to fight and win tomorrow's fight. And DTRA has been a valuable partner.

But you can't really do this without talent. We know that people are greatest asset and we know that Secretary Austin's one of his three key priorities is taking care of our people, and we take that very seriously at CDAO. And so we've actually set up an entire group that is dedicated to Digital Talent Management. So what is our role?

It's a really big one, and it also spans the entire talent management lifecycle. We want to understand end to end what's going on with this workforce, what their pain points are, what they need to be empowered, enabled and equipped, and support that and advocate as we can, as is needed, to make sure that the right talent is at the right place at the right time with data analytics and AI skills. So that means a lot.

It's also a little bit bigger than education. So you noted in your question that we do have an education remit and we also have a remit for AI. It's actually even bigger than that. So we're working an entire range of initiatives. We've got five working bodies of effort here. I'll quickly just touch on each of those if that's okay. Please feel free to cut me off, but I could talk about this all day.

We've got some really audacious goals here and we want to do things a bit differently. How are we going to get after that? First, we set ourselves up by reestablishing ourselves. We had this in the Jake days, but reestablishing ourselves as the functional community manager at the OSD level for the data analytics and AI workforce. And I'll note that we're working in close partnership with our colleagues in research and engineering who oversee the software engineering workforce. We know there's a lot of overlap there.

Similarly, with the Cyber Workforce, we work very closely with our colleagues at the CIO to make sure that we're all working together and we are bringing into the fold the Chief Talent Management Officer and P&R, who oversee a lot of the issues that we are also working with. So the first thing is to make sure we're setting the foundation and we're engaging with the strategic partners we need to work together to get some of these issues over the finish line.

What else are we doing? So in addition to establishing ourselves and some governance around what we're trying to do, we're also building an analytic capability to help DTRA and others with better analytics around who's in this workforce, what is this workforce? What are trends, and demand, and supply? How can we make sure we're planning ahead to resource appropriately? What are the education and training gaps? All of that.

We have a whole development body of work. We're doing education and training, full stop, and that includes baseline education, it includes practitioner upskilling, it includes education for our senior leaders who need to be equipped to succeed in a data and AI enabled world.

We're also thinking about building our community of practice. Well, we will, but how do we want to do that? We're calling it Digital Commons, and this is really meant to be a marketplace, a resource an all stop, one stop that is offering upskilling, career opportunities, being a platform for Chat, access to resources and tools, all of that, taking the best of breed from across the enterprise.

And finally, we're thinking about cultivating tomorrow's defense, digital workforce, and what we need to do, and I think we'll talk a little bit more about this to make sure that we as a DOD and broader US government, US economy, are set up for success and making sure that we all have access to the right talent, at the right place, at the right time.

Dr. Howard:

It's quite a large scope there.

Dr. Diana Gehlhaus:

It is. I'll call it audacious.

Dr. Howard:

Let's talk about that workforce development piece. You mentioned cyber, and I was engaged in some of the critical analysis working groups over the past several weeks back. So what is the DOD Cyber Workforce Framework and how does the DOD Cyber Workforce Framework currently serve components?

Dr. Diana Gehlhaus:

This is a fantastic question because I think a lot of people given the name associate the DCWF with cyber, when in fact it's now been expanded to include data analytics, AI, and software engineering. So what is this framework? Think of it as infrastructure. We're going about how we define and identify people with cyber, data analytics, AI and software engineering skills in a different way. We're moving orthogonally, I'll call it, to occupational series, which are just not as granular as they need to be. They don't get us where we need to go in terms of being able to identify talent, define talent, hire talent.

We know that people who have these knowledge skills and abilities span a whole range of uniform career fields and occupational series. This makes workforce planning really difficult. So what we're doing, and what the DCWF does, is they have work roles. It started with cyber and that's been around for several years, and we've now moved into expanding that framework, and I think the name is actually going to get changed.

But what these are kind of functional areas. So each work role, for example, data steward, data officer, AI innovation leader, AIML specialist, data scientist, these are some of our work roles. They have a definition. They have a whole set of knowledge, skills and abilities and tasks that are associated with them, and they're ultimately coded to work your billet. And what that allows us to do is say, "okay, where in the force does this talent live? What are the needs relative to the supply? Where are the gaps? How can we get after better targeted education and training to meet those needs? How can we get after better workforce planning to fill those gaps? What kind of resourcing do we need to have in place?"

It really helps us in a data-driven way, understand who the cadre of talent is from a spaces and faces perspective. So what is a DCWF? It's a new construct for how we are defining and identifying our workforce.

Dr. Howard:

Okay. That's quite thorough. You may have touched on this through your response to the question I just asked, however inquiring minds want to know, because we've been engaged in this for over a year now, a year and a half probably. Would you explain the pivot from the 2020 DOD AI Education Strategy, which included various archetypes to the new DOD Cyber Workforce Framework?

Dr. Diana Gehlhaus:

Yeah, I actually get this question quite a lot, which is really fantastic to see that so many people in the community were following closely to the 2020 AI Education Strategy and really found value in those archetypes and the education and training programming that came out of that. And we've definitely carried forward a lot of what worked well in terms of a best practice that we're continuing to offer and scale.

So the move away came because we became a CDAO. The AI Education Strategy was really focused on education around six archetypes that were specific to AI, and it did not have the workforce analytics piece. So it was missing a little bit of the rigor behind what talent you have, what talent you need, what kind of education and training does that suggest should be designed and provided to the enterprise?

So the intention was great, but when CDAO was formed, it was a merger of CDO, of JAIC, of the Defense Digital Service, of Advana, of Elements of Project Maven. So there was several groups that came together to form CDAO. Our remit expanded and what we wanted to get after in terms of talent management also expanded to be much broader than just education. So that's why we shifted away from that.

However, the Senior Executive Education Series that we offer came out of the AI Education Strategy. Our work role for the AI Innovation Leader came out of the AI archetypes. So really tried to bring over with us what was best in breed from that effort and was resonating as a need with our community.

Dr. Howard:

Thank you. Thank you for that. Next, I wanted to talk about this in the 2024 NDAA section 222. It requires an update to plans and strategies for artificial intelligence. In regard to CDAO's role, can you share progress on the update to your strategic plan to address the current AI training challenges for the Department of Defense?

Dr. Diana Gehlhaus:

Yeah, so I have not fully read this language, so let me start with that. Having not fully read this language, I can tell you what I know and what we're doing. And I know that CDAO is working on a Data and AI Adoption Strategy writ large for the department. That is not part of directly what our team is doing. We are tracking on that, and I know that that is a priority to provide strategic guidance.

We're also, as a CDAO, putting out several playbooks and other reference materials that are going to support building out your data and AI capabilities, providing some governance, best practices, all of that good stuff. So I know that that is an ongoing and active area that CDAO is working on and very regularly engaging with component lead, CDAO's, CDO's, understanding what the needs are, what the levels of maturity are across the different organizations that we're working with.

Really doing our market research and knowing our customer and being able to provide support and resources accordingly. That's exactly what we're doing in Digital Talent Management from an education and training perspective. So as we are standing up what that means to be a functional community manager, as we are working through and maturing our education and training offerings, we may also choose to put out some strategic guidance.

But in the meantime, I can tell you that we are working carefully, thoughtfully, and aggressively to make sure that our community has the education and training it needs.

Dr. Howard:

Awesome. Great segue into my next question. What is the state of access, availability and the adequacy of basic artificial intelligence training and education curricula available to the broader department be in workforce and military personnel?

Dr. Diana Gehlhaus:

I am going to put my researcher hat on for the first part of this question. I think it is piecemeal when you talk about the state of data and AI education across the enterprise. Especially when you're talking at that foundational or baseline level.

So I think that you've got a difference in how the components are thinking about data and AI in terms of whether they call it digital or data, whether they are emphasizing AI over data, over software, whether they're talking about fluency, acumen, or literacy. And the amount of educational offering that civilians and uniform personnel have access to really does vary.

So I'm going to say that right now, I think there's a lot of great and awesome momentum to bring this type of education to the entire enterprise. But it has been a bit piecemeal. Understandably, this is all part of how education and training is funded and provided at an enterprise level currently.

So this is not a surprise to say that in fact, we do see different levels of maturity and different scales of availability of this type of education. So we are trying to make sure we understand what's already being offered, what's working and what's not in terms of the delivery method, the content format, all of those best practices and pedagogy. And making sure that we are using a common lexicon to talk about some of these issues.

And we are also doing our part to provide baseline education that is asynchronous, that people can access on demand from anytime, any place. So we're actually going to roll this out at the end of the month. We're really excited. Something called MIT Horizon, we're calling it Digital on Demand. And it's not just data and AI, but it's actually all emerging technologies.

It's providing an interactive primer on a range of issues that people can do as much or as little as they want. They can watch podcasts and videos. The content's going to be updated quarterly and it's going to be offered through Digital University as one of our partners. But this is also to say we appreciate that our senior leaders also need tailored education for their needs. So we have our Senior Executive Course, we're also thinking, and we have piloted an online version of this course. So we're thinking about making sure that different modes of delivery are available to our customers based on what they need and how they need it.

Dr. Howard:

Well, that's exciting. The MIT horizons. Yes. I think I remember taking the 101 course with them, but that's awesome. As part of the Defense Direct Reductions Agencies AI readiness strategy, we've engaged in a number of CDAO and DAU pilot training programs, that's Defense Acquisition University. Will the CDAO continue to leverage DAU as a primary source for AI training support as well as other civilian academic institutions?

Dr. Diana Gehlhaus:

The short answer is yes. We are engaged in conversations across a range of potential partners because we really don't want to reinvent the wheel. We want to leverage what works and what's already out there and where people already know they can access information. So DAU has a customer base, DU has a customer base, and we know that both platforms offer basic education in AI fundamentals.

And in fact, DAU has been doing a lot in that space to make it a public offering. So we are tracking what they're doing, we are talking with them regularly, and we are eagerly and actively looking for ways to partner with these institutions to make sure that if there's already content out there that people can have access to that's been vetted and invalidated, then we want to be able to share that with our community.

So this is an all hands on deck, let's not overlap and duplicate. We don't want to reinvent content that's already out there, that is effective, that is resonating with our customers. And so we are having active conversations with a whole range of partners.

Dr. Howard:

That's great news. Well, lastly, what are your thoughts on the challenge we face to attract and retain the talent the DOD needs to compete in the future in light of the lack of enough in-house AI skills?

Dr. Diana Gehlhaus:

I love that question. And my first thought is how much time do you have? Because I think about this question a lot. This is the bread and butter of what I do. And I think that this is not just a DOD problem, it's actually a larger US problem. So we need to make sure that we've got the right economic, education and workforce policies in place across the board to organically cultivate talent and to attract this talent.

Let me just quickly tackle both threads here. From a DOD perspective, I think it starts with doing a better job of defining and identifying what this talent is and where this talent is across the organization. So when we talk about the mix of MIL and SIV, and guard or active, or contractor staff, even in your question, there are some assumptions that we don't have the best data to address or to validate and address.

And so when people ask, "Okay, well we want to attract and recruit a certain amount of talent. Let's do that good workforce planning, let's be data-driven. Let's know who we're talking about, what we have. Do we need to offer upskilling? And if so, in what direction? How can we make our investments more effective strategically when it comes to our human capital?"

I think we're doing a lot there. We are moving out on these work roles that's going to help us get there. But it also includes making sure that the infrastructure is there, that when talent shows up, they have the tools and resources they need to succeed on day one. It also is about organizational and cultural conditions to make sure that data and AI are celebrated, elevated, and prioritized as a strategic priority in organizations.

And that is still a transition that we are trying to encourage, but we still have a little way to go. So I think on the DOD side, it starts with defining and identifying talent that we already have and then being able to plan for the talent we want to bring in. And we have to do those both together. And we are working on initiatives across the board to get after some of these challenges from the attraction and recruiting side all the way through the development side and the assignment, retention, promotion, all of those questions that we know exist in our community. That's why we really wanted to create an end-to-end agenda. So that's on the DOD side.

And I'll just quickly say, on the US side, we need to do more to cultivate AI talent. So if there's not a sufficient pipeline of talent that we're cultivating from the beginning, it's going to really limit how much the government is able to attract and recruit talent.

So you've got this idea of, "Let's get more AI education into classrooms. Let's get more people excited for a career in AI or an AI adjacent field. Let's make sure that we are encouraging DOD to be a place that people want to come and work. And also let's make sure that we're setting the conditions right to attract the best and brightest from across the world. And what does that look like if we want people to be able to work on national security problems in different settings and in different contexts."

So I think there's a bit more work to do. So I have a lot of thoughts there, but I am really excited about what we're doing at CDAO now and with our strategic partners to get after some of these challenges.

Dr. Howard:

That's great. I see infinite possibilities.

Dr. Diana Gehlhaus:

Me too.

Dr. Howard:

Well, Dr. Gehlhaus, it has been a pleasure. I believe the audience would find what you've shared today, very insightful and very informative. I look forward to working with your office in the future, and thank you again for joining us for this podcast.

Dr. Diana Gehlhaus:

Thank you so much. It's been a pleasure and really appreciate the opportunity to share what we're working on, and thanks for being a great partner.

Announcer:

Thanks for listening. To hear more podcasts, don't forget to subscribe on Google Play or wherever you listen. You can find out more about DITRA at ditra.mil.