The center will focus on tactical autonomy, meaning systems that act independently under the bounded authority of human support. The systems support missions like situational awareness, force protection, cyber defense and logistics. The center will also focus on system collaboration and man-machine learning.

The collaboration will be the first academic research center affiliated with the Air Force and the 15th connected to the Defense Department.

“It’s a gap in our suite of research institutes right now that we don’t have one focused on this area of autonomy,” Air Force Secretary Frank Kendall said at the Pentagon Monday. “I’m very focused on the threat of Chinese military modernization and what that means in terms of our viability of our forces for the future. Part of the future of the military is going to be autonomy, there’s no doubt in my mind to that. We’re seeing increasing evidence, evidence for almost in every conflict that occurs.”

The Air Force is committing $12 million per year for the next five years to the collaboration. DoD will be adding an extra $2 million per year.

The Air Force and DoD are currently working with the 11 HCBUs that qualify as high-research activity schools as designated by the Carnegie Classification of Institutions of Higher Education to decide which school will lead the center. That school will be responsible for building a consortium of educational institutions to conduct research.

The collaboration is also working a secondary goal of increasing diversity and inclusion, a mission DoD has been working on since the national response to the murder of George Floyd.

“This is an opportunity to tap into universities that have enormous amount of capability in science and technology,” Kendall said. “The HBCUs put out about 30% of the scientists and engineers that are that are produced by that community.”

DoD will also work on growing the STEM community within HBCUs. The center will not work like other academic research consortia. The Pentagon wants the consortium to build capabilities, but also build its research abilities.

“We understand that there are historical inequities, and we want help them build capacity, as well as deliver results to us,” said Victoria Coleman, DoD’s chief scientist. “We want to advance their efforts to move at least one, hopefully more than one, institution from the Carnegie Foundation Research Classification from ‘R2,’ which means a high research institution, to an ‘R1,’ which is very high research.”

Current and former Hill staffers say technology can and has solved many common problems for members of Congress, but they want to see members tap into more commercial-friendly platforms and give centralized authority to bodies like the Bulk Data Task Force, or the House Digital Service.

Stephen Dwyer, senior adviser to House Majority Leader Steny Hoyer (D-Md.), pointed to solutions such as the Dome Watch and Dome Directory mobile apps, created by the office to help members of Congress, their staff and the public better track movements on the House floor. The 13-year-old private intranet DemCom for House Democratic staff was also redesigned last year with expanded access for Senate staff, mobile functionality and a bigger database of information.

But custom-built systems for “uniquely Congressional purposes,” as Dwyer said, are not all that’s recommended. He told the House Select Committee on Modernization last week, the programs are representative of what is possible when the legislative body coordinates its technology efforts, but that requires in-house digital staff for each office. He recommended hiring digital aides with programming and development skills for every member, in addition to more traditional political science and communications staffers.

“There’s just so much more that they need to do, even versus five, 10 years ago when I was in a congressional office,” Dwyer said. “A lot of that is in digital communications. Every office needs to not just take a bunch of pictures and post them on Twitter and Facebook but they have to do more technical Facebook Lives, they’ve got to take their boss live, there’s a lot of technical tasks that didn’t exist many years ago.”

But Dwyer said Congress needs to recognize the demand for these workers and compensate them appropriately. The House is raising its staff salary floor to $45,000, after essentially a decade-long pay freeze and record inflation made it difficult to attract and retain employees.

One of the Modernization committee’s recommendations last Congress was to create a common committee calendar portal to reduce scheduling conflicts. Vice Chairman William Timmons (R-S.C.) asked witnesses for suggestions to get the ball rolling on what he said could have a big impact on members and staff. Reynold Schweickhardt, Lincoln Network senior adviser, said an issue is that between the House clerk, the chief information officer and the committees themselves there is no clear button to push for technology needs. As such, another recommendation was to clearly focus the responsibility for legislative product.

“I think the other challenge that I alluded to is there’s no gatekeeper for scheduling projects. The CIO, they may be working on five to 10 projects, so they tell you they’re working on your project and they are, but they’re sort of shuffling things back and forth,” said Schweickhardt, who served at the Committee on House Administration for 13 years and the Government Publishing Office for eight years Versus a program-management kind of functionality that says, ‘What are the three things we want to accomplish in the next couple of months? Let’s knock ‘em out. Let’s figure out what the next set of important things are.’”

Dwyer said the foundation is laid. For several years, House rules have required all committees to post hearings and testimony in a central place, putting the body ahead of the Senate, but amplifying that with a more consumer-friendly version would help, he said.

Melissa Dargan, co-founder of AppMy LLC and a former Hill staffer, seconded the use of funds for a centralized scheduling platform. It’s a problem she tried to tackle when she launched the TourTrackr app to better manage constituent tour requests.

“From constituent tour requests to flown flag purchases, these important responsibilities were tracked using printouts, binders and excel spreadsheets. It was a fragmented, inefficient process. At the time, there were no digital alternatives that House offices were approved to use. So while these tasks seemed easy, they were tedious, repetitive and time consuming,” she said, adding that technological innovation does not need to come at the expense of security.

“I respect and understand that the House has high standards for new tech approval. Protecting security and personal identifiable information are critical to ensure the integrity of the institution. That said, upholding those priorities and creating a welcoming environment for new tech products can be done simultaneously,” Dargan said.

The Defense Counterintelligence and Security Agency is adding more data categories to its “continuous vetting” program, while more than three dozen non-defense agencies are using DCSA’s services as it moves ahead with security clearance reform efforts.

DCSA is now monitoring 50,000 cleared individuals from 38 non-defense agencies under its continuous vetting program, according to Heather Green, assistant director of vetting risk operations at DCSA. That’s on top of the 3.6 million Defense Department service members, civilians and contractors who were enrolled in continuous vetting by last October.

“We do anticipate this continuing to grow through this fiscal year and beyond as we add those additional CV services,” Green said on Inside the IC. “So as more services and capability comes online, we’re going to continue to grow our service to our federal agencies, as well as our DoD customers.”

DCSA is one of the lead agencies implementing the government-wide “Trusted Workforce 2.0” initiative. The effort aims to streamline the government’s personnel vetting process through automated record checks, simplified security standards and more information sharing across agencies.

Earlier this year, a White House official called 2022 the “most significant, most consequential” year for security clearance reform yet.

DCSA’s continuous vetting system is one of the centerpieces of the initiative, giving investigators the ability to receive automated alerts when a security clearance holder faces an issue that could put their clearance at risk.

DCSA’s continuous vetting system hit the “1.25” milestone last fall when the DoD cleared population was fully enrolled. But the system at that time was limited to three data categories: criminal activities, terrorism, and eligibility.

The agency is now adding alerts about suspicious financial activity, foreign travel, credit history and “public record information” to the continuous vetting system on the way to the “Trusted Workforce 1.5” milestone this fall, according to Green. Already, 2 million DoD clearance holders are  enrolled in that expanded system of alerts.

The continuous vetting system is replacing periodic reinvestigations, where investigators would do a background check on a clearance holder every five or 10 years. Instead of learning about potentially suspicious activity years after it took place, the system is intended to provide security offices with alerts about such activity. Investigators can then decide whether to follow up.

But Green says continuous vetting is not a “gotcha” program. Instead, she says it’s intended to improve security while also giving cleared personnel the chance to self-report and mitigate any potential issues.

“In the grand scheme of things, very few individuals actually receive an alert or require the additional investigative action to take place,” Green said. “But CV isn’t just about generating those alerts. It really is about self-reporting. There are self-reporting requirements for clearance holders, and it’s really supporting the goal of helping us identify potential issues before they fester into a larger insider threat concern.”

Security clearance reciprocity timelines down

DCSA has also made major strides in the time it takes for it to process and adjudicate a security clearance granted by another agency, a process referred to as “reciprocity.” The process affects personnel transferring from one agency to another, as well as contractors working on different projects for different agencies.

DCSA now takes an average of just one day to make a reciprocity decision, down from a peak of 65 days in mid-2020, according to Green.

She credited “business process engineering” leading to more efficient decision-making on reciprocity requests, as well as the merger of several organizations under DCSA, including the former National Background Investigations Bureau and the DoD Consolidated Adjudications Facility.

Last year, DCSA also completed the shift from using multiple personnel security databases to the Defense Information Security System (DISS).

“Having the ability to control the end-to-end process was certainly a part of that success,” Green said of reciprocity.

While other agencies, most notably in the intelligence community, take much longer to make reciprocity decisions, Green thinks the “transfer of trust” process, as it’s called under Trusted Workforce 2.0, will continue to improve with time.

Initial vetting hurdles

One of the next major hurdles for security clearance reform will be speeding up the time it takes to get an initial applicant, with no prior government background investigation, through the vetting process. The initial background investigations process often takes months and even years in some cases, making it harder for the federal government to hire new employees.

A quarterly update issued by the Security, Suitability, and Credentialing Performance Accountability Council earlier this year shows DCSA is expected to begin offering initial vetting products, using more automated processes and the new National Background Investigation Services (NBIS) IT system, starting next June.

“We are committed to being what I would call the ‘personal security provider of choice,'” Green said. “We’re working very hard to provide new and enhanced products and services to support that full TW 2.0 implementation to include initial vetting products. The actual implementation of the new standards will take some time and will be fully phased in as those products and services are available. But we are leaning forward, looking at how we can continue to evolve all our vetting products and services.”

The IRS announced Friday that individuals delinquent on their taxes, who receive a mailed notice from the agency, can call an artificial intelligence-powered bot and set up a payment without having to wait on the phone to speak with an IRS employee.

Taxpayers are eligible to set up a payment plan through the voice bot if they owe the IRS less than $25,000, which IRS officials said covers the vast majority of taxpayers with balances owed.

Taxpayers who call the Automated Collection System (ACS) and Accounts Management toll-free lines and want to discuss their payment plan options can verify their identities with a personal identification number on the notice they received in the mail.

Darren Guillot, IRS Deputy Commissioner of Small Business/Self Employed Collection & Operations Support, told reporters Friday that the agency’s expanded use of voice bots and chatbots will allow the IRS workforce to assist more taxpayers over the phone.

The IRS earlier this year answered about three out of every 10 calls from taxpayers.

“If you don’t have more people to answer phone calls, what are the types of taxpayer issues that are so straightforward that artificial intelligence could do it for us, to free up more of our human assisters to interact with taxpayers who need to talk to us about much more complex issues,” Guillot said.

IRS Commissioner Chuck Rettig said the automation initiative is part of a wider effort to improve taxpayer experience at the agency.

“We continue to look for ways to better assist taxpayers, and that includes helping people avoid waiting on hold or having to make a second phone call to get what they need,” Rettig said in a statement

Guillot said the IRS in December 2021 and January 2022 launched bots that could assist taxpayers with questions that don’t require authentication of the taxpayer’s identity or access to their private information.

These bots could answer basic questions like how to set up a one-time payment, and answered more than 3 million calls before the end of May.

But this week, the IRS expanded its capabilities and launched bots that can authenticate a taxpayer’s identity and set up a payment plan for individuals.

“It verifies you really are who you say you are, by asking for some basic information and a number that you will have on the notice you received. That gives you a phone number to call and speak with the bot,” Guillot said.

Guillot said taxpayers can name their own price for the payment plan, as long as a taxpayer pays their balance within the timeframe of the relevant collection statute or up to 72 months.

Once a payment plan is set up, the bot will close the taxpayer’s account without any further enforcement action from the IRS.

“Those taxpayers didn’t wait on hold for one second,” Guillot said.

Guillot said the IRS is ramping up its bot capability incrementally to ensure the automation can handle the volume of calls it receives. The bot, he added, is currently at about one-quarter of its full capability, and will reach 100% capacity by next week.

The bots are available 24/7 and can communicate with taxpayers in English and Spanish.

Later this year, Guillot said the bots will be able to provide taxpayers with a transcript of their accounts that includes the balance of their accounts.

Guillot said the IRS worked closely with National Taxpayer Advocate Erin Collins on the rollout of the voice bot.

“She raised legitimate concerns that some taxpayers, because they can name their price, may get themselves into a payment plan that’s more than they can afford,” he said.

The IRS is working to ensure that the bots ask some additional questions to ensure taxpayers are able to afford the payment plans they set for themselves.

Guillot said that this week’s rollout marks the first time in IRS history that the agency has been able to interact with the taxpayers using AI to access their accounts and resolve certain situations without having to wait on hold.

“I have friends and family that have to interact with the Internal Revenue Service, and when I hear them talk about how long they’re on hold that bugs me. It should bug all of us,” Guillot said.

Guillot said the IRS also added a “quick response” QR code to the mailed notices that went out to taxpayers. The QR code takes taxpayers to a page on IRS.gov showing them how to make a payment.

Guillot said the IRS originally expected to launch this capability by 2024, but was able to expedite the rollout given the perceived demand for this service.

The IRS in recent years has seen low levels of phone service that have decreased further since the start of the COVID-19 pandemic.

IRS is looking to further expand the range of services voice bots can provide, and is part of a broader effort to improve taxpayer service.

“We never lose sight of our first interaction with every single taxpayer is never enforcement. It’s a last resort. Our first effort is always around that word ‘service,’ and trying to help customers understand the tax law and almost always work out a resolution with them meaningfully,” Guillot said.

The Senate on Thursday passed the Sergeant First Class Heath Robinson Honoring our Promise to Address Comprehensive Toxics (Honoring Our PACT) Act. The bill now heads back to the House for final approval before heading to President Joe Biden’s desk.

The legislation, at its core, would expand disability compensation and health care benefits for veterans exposed to toxic substances during their military service.

VA Secretary Denis McDonough told the Senate VA Committee on Tuesday that the PACT Act, combined with the agency’s fiscal 2023 budget request, will give the VA the resources it needs to prepare its health care workforce to treat up to 3.5 million additional veterans.

“This is a very important piece of legislation. I think it’d be very difficult to implement, but oftentimes, the most important things are difficult, and I think we’re ready for it,” McDonough said.

The VA said in a statement last month that the PACT Act would be one of the largest substantive health and benefit expansions in VA’s history, comparable in scale and impact to the 1991 Agent Orange Act.

The VA’s budget request for fiscal 2023 includes $42.2 billion for medical service staffing, and provides for 22,789 full-time equivalent hires, an increase of 14,000 FTEs compared to last year.

McDonough said the PACT Act would give the VA much-needed authority to set higher pay caps for certain health care positions, and that the VA’s ability to get veterans into care more quickly is “obviously impacted by the tightness of the labor market.”

“The first thing we need to do is retain the docs that we have, and you’re giving us new authorities to do that. Pay is a big one, and you’re giving us enhanced recruiting authorities as well. We’re thinking very diligently about this and planning very diligently about making sure that we have the people in the spots, and that we have the buildings for the increased demand that we anticipate seeing,” he said.

White House Press Secretary Karine Jean-Pierre said in a statement last week that the PACT Act marks “one of the most significant and substantive expansions of benefits and services in the Department of Veterans Affairs history,” and that the Senate is taking steps to ensure the VA has the support it needs to effectively implement the legislation.

The Congressional Budget Office estimates the bill would increase spending subject to congressional appropriations by $147 billion through 2031. CBO estimates that 5.4 million veterans, nearly a third of the 19 million veterans in the U.S., will receive some disability compensation this fiscal year.

The PACT Act contains a slew of provisions meant to bolster the VA’s workforce, health care facilities and claims processing capabilities, and is the latest in a series of bills meant to recruit and retain in-demand health care workers and address the agency’s record-high rate of turnover.

Committee Chairman Jon Tester (D-Mont.) and Ranking Member Jerry Moran (R-Kan.) introduced the bill.

Congress in March also passed the RAISE Act, which raises the pay caps for certain VA nurses and physician assistants. McDonough said about 10,000 nurses will see their salary increase next month under that legislation — that’s about one out of every nine VA nurses.

McDonough said the Office of Personnel Management recently gave the VA another year to continue with its direct-hire authority for health care workers.

But even with that expedited hiring authority, Veterans Health Administration officials recently told the committee it’s taking 95 days on average to hire new employees.

“The direct-hire authority, of the many variables in the equation of bringing people on, it’s perhaps the most impactful. It accounts for probably a third of the savings we were able to get, in terms of time to hire. But the hiring and onboarding process is still so sclerotic, that we’re finding things that can change,” McDonough said.

Among the changes the VA is looking at, McDonough said nurses have to write an essay as part of their onboarding process.

“I think that’s antiquated, and we should get rid of that,” he said.

The PACT Act outlines many provisions meant to make the VA a more attractive employer for health care workers in a competitive labor market.

The bill would also give the VA up to $40 million a year to buy out the contracts of certain private-sector health care professionals in exchange for employment at rural VA facilities.

The bill also expands recruitment and retention bonuses for VA employees, including merit awards and pay incentives for employees that have a “high-demand skill or skill that is at a shortage.” The critical-skills pay incentive cannot exceed 25% of an employee’s base pay.

The bill also includes expedited hiring authority for college graduates into competitive service jobs.

The PACT Act also gives the VA 180 days to work with OPM to establish qualifications for each human resources position within the VA, and to establish standardized performance metrics for its human resources positions.

The bill gives the VA a year to submit to the House and Senate committees a plan on how it will recruit and retain HR employees.

The agency would also have 90 days to provide enhanced monitoring of the hiring and other human resources that happen at the local regional and national levels of the department. The agency must also provide at least annual training to human resources professionals in VHA.

VHA officials told the committee last month that non-standardized HR processes at the local level have led to hiring and onboarding inefficiencies across the agency. The legislation, if signed, gives the VA 18 months to develop a national rural recruitment and hiring strategy for VHA.

As part of this strategy, the VA must determine which clinics or centers have a staffing shortage of health care professionals, and develop best practices and techniques for recruiting health care professionals for such clinics and centers. The PACT Act requires the VA to provide the House and Senate VA committees with updates on its progress in implementing the rural recruitment and hiring strategy within 18 months of the bill going into effect and then annually.

The bill would waive pay caps for VHA impacted by the closure or realignment of their official duty stations, which may happen if the agency’s recommendations to the Asset and Infrastructure Review (AIR) Commission come into focus.

The PACT Act also waives pay caps for VHA employees providing care to veterans exposed to open burn pits.

While Senate Majority Leader Chuck Schumer and House Oversight and Reform Committee Chairwoman have opposed the planned closure of VA medical centers in New York, Sen. Richard Blumenthal (D-Conn.) said the VA can’t continue to deliver modern health care in outdated facilities.

“This strategy is untenable. In fact, it’s not a strategy,” Blumenthal said. “There is no way that the Veterans Administration can continue quality care with facilities of that age at a time when technology requires that the entire structure of a facility be designed and built to accommodate the most modern means of delivering care, of monitoring patient health.”

McDonough said that if the AIR Commission process doesn’t move forward, the MISSION Act still requires the VA to conduct four-year reviews of its real-estate needs in each of its regional health care markets.

“We’re watching to see what you all choose to do with the nominees for the AIR Commission. In all cases under the MISSION Act. We’re required to go back and look each four years at what the needs are in each of those markets across the country,” he said.

Meanwhile, the PACT Act authorizes 31 leases for new VA health care facilities across the country.

Bill mandates VA updates on claims automation

The PACT Act also gives the VA 180 days to submit a plan to Congress on the state of IT modernization at the Veterans Benefits Administration.

The report should identify any legacy systems the VA plans to retire or modernize, as well as update Congress on the progress the VA is making in automating claims processing decisions.

The bill states that automation “should be conducted in a manner that enhances the productivity of employees,” but keeps VA employees in charge of making the final decision of granting benefits to claimants.

The bill makes clear that the automation should “not be carried out in a manner that reduces or infringes upon the due process rights of applicants.”

McDonough told the committee that the current claims backlog is 188,000, which is down from 265,000 claims only a few months ago.

The VA announced in January that the automation pilot, through its newly created Office of Automated Benefit Delivery, is processing claims within a day or two, while the traditional method of processing these claims currently takes well over 100 days, on average.

The VA began the pilot by automating claims of service-related hypertension, and is adding three new diagnostic codes each quarter.

McDonough said that by the end of the year, the 12 most common claims will be automatable.

While the Biden administration and some lawmakers have pressed for federal employees to return to the office, McDonough said the VBA productivity increased during the pandemic.

“As we think about questions about do people come back in the office, or do they work virtually, we’re taking that into consideration,” he said.

VA is also in the process of hiring 2,000 additional claims personnel.

Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.

When a veteran passes away, it currently takes more than two months on average for the Department of Veterans Affairs to process their survivors’ claims for dependency and indemnity compensation. But automation is helping to drastically shorten that timeline. Assuming the application is complete on the front end, the department can now approve those claims within about four hours without a claims process or ever having to see it. Kevin Friel is deputy director of VA’s pension and fiduciary service. He joined the Federal Drive with Tom Temin to talk more about how they’ve been using automation he in his section of the department.

Interview transcript:

Jared Serbu: Kevin, thanks for doing this and wanted to start by getting a little bit of an update on where you are with automation. VBA automation has been a little bit in the news lately with the disability compensation part of the administration, Getting in on some pilots, but you all in the pension and fiduciary service have been working on automation for quite a bit longer. Tell us how robust, how mature all your automation efforts are, what kind of progress you’ve made so far.

Kevin Friel: Yeah, Jared, thanks for your question. Just to be clear, we are separate from compensation service. So our automation, as you said, is a lot different. We’ve actually been automating since about 2014. There was a public law 114-315 that Congress passed that basically told us, as it relates to surviving spouses, we should pay them the benefits they’re entitled to based on the evidence or record. So in the past, if a veteran passed away, the spouse would have to submit an application to us. And we would have to go through that process. When that law was passed, what we looked at is what we could do so the two things we made out that we could do is we could pay burial benefits.

So VA helps to offset burial benefits for veterans who pass away and are in receipt of VA benefits at the time of their death or have a claim pending where they would have been or if the remains are unclaimed, right, for unclaimed remains. So with that, if we have we have the surviving spouse on record, once we get that notification, and the spouse calls us, we can process the burial claims and burial awards and get them out to the spouse without even getting an application just because we know we have identified the spouse and linked them to the veteran. And then we have for, DIC, dependency indemnity compensation, right, typically that is meant for to be payable to the surviving spouse for the veteran’s death was related to a service-connected condition.

However, there’s a law that says if the veteran was rated 100% compensable, right, for service-connected for periods of time, that we can go ahead and pay that administratively. So we also look at that in the automation process. And if the veteran meets those criteria, we will go ahead and initiate DIC to the surviving spouse automatically, once again, without an application. We also do some automation, State Veterans cemeteries, who in turn into our veterans, right who are eligible for burial in the National Veterans Cemetery. They do that at no cost to the beneficiary, or the survivors. The VA will, however, reimburse them, the cost of plot and plot benefit, which is about $832 – I may be a little off on the money, but it’s over $800. We went through that process and looked at it to see how we can improve it. And we’ve automated that process, too. So now when the state submit us, we can run those through automation and get them done quickly. And then we went to our next level, which was how do we do something where we get a claim report from a survivor or for … a DIC. And over starting in 2019, we started to work on automation, we have been looking at it and trying to get it implemented over a period of time. In 2019, we finally got everything lined up. And we started to roll that out.

So today, we’re seeing an application for burial claim or for a DIC claim (dependency indemnity compensation claim), we extract the data from that, and then we run it through our automated processes. If automation can process that claim to completion, which is normally the award, right, always an award or benefits, it’ll run that through all the rules, and it’ll pay that out. With our current processes if the surviving spouse or gives us all the information we need, which is typically the form is completed the way that it needs to be completed, all the blocks are filled in that need to be filled in. And we have a copy of the veterans death certificate, we have been able to process those claims in as little as four hours. From when it comes in, it’s it’s automatically established, the data is automatically extracted from the form and then it it is run through our automation thing. And it’s processed out. If we don’t have that if the information isn’t clean, or if we don’t have the death certificate, it just delays the process.

So getting all the information up front is where we’re going to be. However, what we don’t want is somebody to hold off on applying. Because, we know we understand that there’s there’s many times delays in getting a death certificate, right? They’re not always readily available. Typically you have a year to apply for the benefit, which will go back to the date of the veteran’s death. So our thing is get the application in, if you don’t have the death certificate or you’re waiting for it, get that application in. At least we have your benefit, your claim established, as as you work to get that death certificate. Once we get the death certificate automation, we’ll go through and extract that just like it already has the form and if we have everything we need once the death certificate comes in, automation will pick that up and run it. So it’s not like a one-and-done, we look at these claims every evening. And if there’s a change in status, we will go ahead and work that out.

Jared Serbu: In a best-case scenario, if the claim has been submitted with everything that VBA needs to process it, does a human ever need to touch it , or is it automated end to end?

Kevin Friel: In the best case scenario, right now, no a human doesn’t need to touch it. We have done it, we have an automatic claims establishment, we have criteria around that. So if we the information is within our system, we could establish that claim right away based off of the form and the information we have on the form, then we do accept we do the automatic data extraction from it. And so if the data extraction, it comes in clean, we run those likely. And if if everything’s there, it’ll it’ll go right through.

Jared Serbu: And we were mentioning a little bit off the air that pension automation is coming a little bit later. Is that just a more complex process? And and tell us how that’s been going? If you could,

Kevin Friel: it is pensions a little more complex, because basically, we have to look at all the income that a beneficiary has, we have to look at the medical expenses they have. And in the pension program, we use medical expenses as a means to reduce income. It’s a needs-based program, right, so it’s an income-based program. And the income levels are are set by Congress, the thresholds, but we are allowed to use a medical expense to do reduction of the income to potentially get them below the threshold. So just to use round numbers, say that we say a veteran can have $14,000 a year in income. If they have $17,000 or $18,000 or $20,000, but they have, $7,000 or $8,000 or $9,000 [in] medical expenses that they have paid, we would, eventually we can reduce that income down based off of their medical expenses to get them below the $14,000 threshold and then be able to pay them benefits.

Jared Serbu: Similar objectives on the pension side to get to somewhere in the four-hour area, or at least inside a day?

Kevin Friel: There are. I mean, and I’d say I like the four hours, the best case scenario, we don’t get many of those. But we do get a few every every evening. Typically we have to send these out because we don’t get all the information. But yeah, the goal would be to get as many of these through as quickly as we can. Because when the beneficiary comes into us for what we see within the pension world, because we have responsibility for burial and DIC and pension. So as I said, pension’s needs-based. So when they come in to us, they’re really in need of this benefit, right? Financially, they’re struggling, and the faster we can get money into their hands, so that we can help them meet their day-to-day living needs and stuff like this is where we want to be. And with the survivors benefits the DIC and burial, typically, these are people who’ve been married to the veterans, veteran and spouse have been married for years. It’s typically not like, a short period of time. But you’re talking, 20, 30, 40 years, and so they’re going through the worst time of their life. So if I can do anything I can to get their, A) is to get them the money to help reimburse for the funeral and then B) is, if they’re eligible for DIC to get that flow where they can help offset the loss income, we want to do that as quickly as we possibly can. That’s why even like with the first automation, as I said, we do that without getting any application. We get a call, we send a letter out. And then six days later, we don’t have anything that says we shouldn’t pay. We pay that benefit.

Jared Serbu: Regardless of whether the application is being processed manually or going through automation, veterans still need to submit essentially the same thing, or veterans or survivors still need to submit essentially the same documentation.

Kevin Friel: Correct, yes.

Jared Serbu: Can you give us just what was some of the most important things that VBA needs to ensure that they’ve got the best chance of getting through that automation process as quickly as possible?

Kevin Friel: Sure. So for the veterans’ purposes, when we get a pension claim, one of the things that we – two things actually that kind of slow us down. One is the application not being completely filled out, right? So for the application to be eligible for pension, whether it be a survivor or veteran, we need to know their income. We need them to put down and tell us what income they have. Typically, it’s only for our population. It’s a small, it’s either social security and maybe retirement or annuity or things like that. But that all has to be listed. And we need to narrow history and we need the medical expenses. And if they’re in a nursing home, we need specific forms filled out. So filling out the application completely, right, as completely as possible will help us expedite it.

Also, if they’ve never filed a claim with VA before, submitting their proof of their service verification, right, so the 214 or the 215, whatever they have for service verification, because one of the requirements is you have to be a veteran, right? So we need to validate that they are a veteran. For the population that we look at, like when you go back to like Vietnam, Korea, World War II where we get the veterans and survivors within that population, they’re typically, we don’t have their electronic records. So we’re totally tied to paper on these people. So for all these veterans and their survivors of getting that form 214 in and we’ve actually recently in November of 2021. We changed our rule, right? Previously that 214 had to be certified, had to be stamped by a VSO (veteran service organization) or it had to come right from the Department of Defense or the branch of service.

We have now said, if we get one that comes in from the beneficiary, from the applicant, and there’s no indication that it’s been altered or adjusted or whatever, we will accept that. So we won’t delay the claim anymore. Because I’m sure we’ve all heard about the federal records delays that we’ve had because of COVID. Right now, with our new rule, if it comes in, and it says it looks good, we’re going to take that and move that forward. But having that as part of the package, too. And then on the survivor side, when we get a DIC claim, one of the big things is having the death certificate. To be eligible for DIC, [the] veteran had to have passed away from a service-connected condition, whether it be a primary or secondary cause of death. So, having the death certificate is the only way we can make that link.

Now, one of the things that I will tell everybody is, do not make the decision on your own whether or not you’re eligible for our benefit. If the veteran passed away, and you believe it’s service-connected, or you think it may be just send it in, let us make the determination, I can’t tell you that we’re going to grant it to everybody, but allowing us to make the decision, and allowing us to review the data is probably the best scenario for anyone, even if we deny it. And at least you know that VA has looked at it, and you haven’t been told by some third party “don’t apply because you’re not eligible,” right? Let us make the decisions and submit the applications to us.

Jared Serbu: And then the lesson in our last couple minutes here, Kevin, I do want to emphasize that there is third-party help out there for folks who want to use it. VSOs, attorneys, others, can you talk a little bit about what VBA generally recommends, if someone’s gonna go that route and some of the red flags for people who may be a little bit less honest about helping veterans?

Kevin Friel: Sure, thank you. That’s a great, great lead. And so I will tell you that for our purposes, the veterans service organizations are our biggest advocates, right? For VA, they are our frontline. They are actually out, working with different communities and being available to them, especially like rural areas where you have VFW’s and American Legion posts and things like that. That’s where you should be going if you need assistance,. Or you can call the 1-800-827-1000 number, right? Which is our VA call centers, and we have representatives there who will assist in filling out a form and providing guidance in there. And then if you’re close enough to a regional office, you can schedule an appointment and walk into a regional office and they will sit down and help you fill that all out.

And it’s really important that if you’re going to need to seek assistance, seek something from these organizations, these are these are all representatives. Veteran service organizations have all been that validated by the VA and they they’ve all been certified to do the work that they do, and they don’t charge. We even tell it on our forms. Individuals cannot be charged to submit an original claim. So a lawyer fee can’t be charged, shouldn’t be charged for submitting an original claim. Now, on a subsequent claim, or where they are appealing something and they want a legal representative, that’s a completely different story. But for their original application, there should be no fee. VSO will help fill that in help submit it. And the other thing with the VSOs, that’s beneficial to the beneficiary or the claimant is the VSOs have direct contact with the VA. They have lines that they can call, they can talk to us. So they have the ability to follow up on your claim, even if the circumstances changes.

So if someone is terminally ill, or if they’re about to become homeless … the VSOs can let us know that and we will work to expedite the claim, and make sure that we can help these veterans out and their survivors as quickly as possible. But you have third-party companies out there that are coming in and and saying we’ll help you with these medical expenses and we’ll help offset these, we’ll give you a loan and then you can repay it or we’ll take a percentage of whatever your payment is for this period of time. And that shouldn’t be happening. So our thing is to stay with the VSOs. Like I said they’re certified by the VA, right, and we work closely with them. So that’s probably the best benefit.

The Pentagon’s new chief digital and artificial intelligence office is quickly bringing together multiple tech specialists under one roof in the latest bid to scale a “different operating model” for delivering digital technologies across the Defense Department.

The CDAO hit full operational capability on June 1 and is hosting an online “DoD Digital and AI Symposium” this week. The office merges the Joint Artificial Intelligence Center, the Defense Digital Service, the chief data officer and the Advancing Analytics (Advana) platform that originated in the DoD comptroller’s office.

Margie Palmieri, the deputy chief digital and artificial intelligence officer, compared it to a major merger and acquisition activity in industry.

“Companies go into mergers and acquisitions to be competitive,” she said on day one of the symposium. “And that’s exactly what the Department of Defense is doing. We are increasing our competitive advantage by bringing these different groups together. And for the first time in my career of over 15 years in government at this point, all the right levers of change and influence are coming into play in the CDAO.”

While recognizing that DoD’s industrial age acquisition approaches are still appropriate for the thousands of aircraft, ships and vehicles the Pentagon buys every year, Palmieri said software, digital technologies and data analytics require an alternative approach to scale across DoD. The various organizations coming together under the CDAO all have experience in piloting alternative approaches to the Pentagon’s traditional development and buying processes.

“This team coming together to show what a different operating model looks like is one of our top priorities,” Palmieria said.

Who’s on the CDAO team

In late April, the Pentagon announced Craig Martell would be the first chief digital and AI officer. Martell was most recently head of machine learning at Lyft, previously led machine learning at Dropbox, and he led a number of AI initiatives at LinkedIn. He also was a tenured computer science professor at the Naval Postgraduate School.

“I’m doing it because of the mission,” Martell said during the symposium. “It’s extremely important that we get this right. And there are not a lot of folks who have the intersection of AI and a government background. And when the deputy secretary of defense calls you up and says, ‘We would like you to take this job,’ you have to think really hard about why you wouldn’t take the job, and not the other way around. And I think getting this mission right is extremely important.”

As the No. 2 at the CDAO, Palmieri has more than a decade of DoD experience, most recently as special assistant to the vice chief of naval operations. She was also founding director of the Navy Digital Warfare Office, a relatively new organization focused on data analytics and Project Overmatch, the sea service’s contribution to the military’s Joint All Domain Command and Control concept.

The CDAO also has filled out its ranks with officials from the JAIC, DDS, the CDO’s office, Advana and other DoD organizations:

• Clark Cully, previously DoD’s deputy chief data officer, is now the deputy CDAO for policy, strategy and governance.
• Sharothi Pikar is the deputy CDAO for acquisition and AI assurance. She joined the JAIC in the spring to lead acquisitions. She also previously held positions as acquisition executive at U.S. Cyber Command and as the associate director for cyber strategies in the office of the under secretary for research and engineering.
• William Streilein, who had a long career at MIT Lincoln Labs before joining the JAIC as chief technology officer this spring, is now CTO for the CDAO.
• Joe Larson, currently chief of Project Maven, is leaving the project to be Deputy CDAO for Algorithmic Warfare. Larson helped co-found Project Maven, the Pentagon’s AI pathfinder program, which is now transitioning to the National Geospatial-Intelligence Agency.
• Greg Little is the deputy CDAO for enterprise capabilities. He was previously deputy comptroller for enterprise data and business performance. Little was also program lead for Advana, a big data program that now falls under the purview of the CDAO.
• Katie Olsen, who was the deputy director of the Defense Digital Service, is now the deputy CDAO for digital services.

The DDS in particular excelled at bringing tech talent from outside the government into DoD for temporary tours of duty to work on high-profile problems. Recently, DDS led Project Rabbit, an effort to allow employers to verify their previous or current Afghan employees seeking asylum in the United States.

Olsen says she think the CDAO can help scale the DDS model.

“We’ve been a ragtag team of 50 to 60 people, give or take, for the past seven years,” Olsen said during the symposium. “I’m excited about being part of the CDAO because I think it’s an opportunity to replicate and scale that talent and the [idea that] when you give smart people agency to do things, what can happen.”

CDAO operating model

Palmieri said the groundwork laid by CDAO’s predecessor organizations have helped establish good practices for how to approach digital and AI technologies within DoD.

“The good news is everything that the department has to do to scale digital analytics and AI is relatively known at this point,” she said.

In particular, she said determining the right “feedback loop” between users and developers will be key as CDAO looks to tailor the budget, requirements and acquisition processes to its technology goals.

“We really see the opportunity for the CDAO to put together a different operating model for how do you actually deliver these types of capabilities in a meaningful way to users and create a more rapid feedback loop where requirements, acquisition and funding are all in response to that capability need, instead of driving the pace, as happens today.”

Former Google chief executive Eric Schmidt told Martell he should avoid a wholesale reliance on prime contracting. Schmidt has been highly influential in driving DoD’s technology organizations and strategies in recent years as chairman of the Defense Innovation Board and then co-chairman of the National Security Commission on AI.

“My guess is that at the end of the day, your success will be your own software people in the government, as well as essentially small firms where you’re essentially contracting with a firm, but it’s really one person. And then you’ll have some priming. If you start with a full private approach, you’ll never get there.”

Martell said it will likely be a combination of in-house coders and contracted work, suggesting a collaborative approach.

“We’re not going to grow the talent fast enough to have all of the coders in government to do what we need to do,” he said. “We’re not going to replicate these agile new AI companies. But if we have the authority to say, ‘You’re sitting with us, and you’re our agile team, and you’re going through the loop with us as we’re building it,’ I think we have a greater chance of success.”

For the last few years, the IRS has been changing its external reputation and internal culture of an agency that doesn’t take technology risks.

The Pilot IRS program may be one of the most well-known examples of this external evolution, reaching out to vendors to bring in innovation and new approaches to contracting.

Internally, the use of robotics process automation in the procurement and financial offices has been a strong influence on the workforce’s culture.

Shanna Webbers, the assistant deputy commissioner for operations support at the IRS, said a combination of short-term wins and agencywide collaboration helped drive two major changes.

“The technology enables our staff do other things or things that they may find more interesting, instead of just mundane, repetitive tasks. That’s really what we want to do. We want to create an environment where our staff want to come to work, where they’re excited to come to work, where they feel like they are making a tremendous amount of value to executing our mission at the IRS,” Webbers said on Ask the CIO. “How do we create that and beyond just upskilling and rescaling in areas with human resources-related to technology or robotic process automation? We really are looking at the whole person. How do we ensure that every individual in the organization has the right skill set, the right experience and the right knowledge to take on positions of greater responsibility?”

Webbers said that meant changing how they trained the workforce. Instead of focusing 80% of the training on the technical aspects of procurement, the use of RPA and automation has let the IRS refocus the training to 50% on technical and 50% on other skills like critical thinking, writing, leadership and collaboration.

Agency collaboration

The second initiative to drive the culture change is to create a partnership with the agency’s chief information officer’s office.

Webbers said the CIO’s office must give its final “blessing” before the bot can launch, the office has provided acquisition and financial with liaisons to help work through the documentation and security processes.

“We have a process in place where when we have ideas, we, through an intake form, submit them to the CIO and they get reviewed. There’s an IT advisory board that makes recommendations for how to move forward,” she said. “If we don’t have the capacity or the funding to move forward, the CIO’s office helps to prioritize the RPA investments.”

Across procurement and financial, the IRS has implemented a small amount of RPA bots so far, but expect to increase the number over the next year.

“We have other ones that are in the queue right now. In the next six to 12 months we are  focusing on RPA implementation that’s working around data reconciliation and management for our manual adjustments for refunds and deposits, that could save up to 35,000 hours per year,” said Teresa Hunter, the IRS’s CFO. “There’s a significant opportunity. We are a very paper-based organization. It’s just a matter of our IT organization having the capacity and the funding to keep up with the demand that is going to be coming their way. The CIO has a big job of making sure that we are secure and safe. It’s a balance between having a good relationship with your CIO office and understanding their perspective as well as the needs of the businesses.”

Hunter, like many CFOs, are seeing the value of bots for financial operations.

She said she is encouraging the staff to take a new way of looking at how they could do their work, which areas are repetitive and require mundane tasks that somebody had to do that.

“We know it does save time and effort on our staffs’ part,” Hunter said. “As we’re working on the automation, the innovation, the efficiency effort within CFO, we’re also looking at skill sets of our staff and how can we upskill or reskill them? What are the core areas that we want to focus on of making sure our staff are being trained to develop and grow in their role as we move forward with some of these shifts and changes? We aren’t looking to reduce full-time equivalents (FTE), but we’re looking at being able to be more analytical in our decision making and how we are approaching the work that we have to do so that we can be more successful and how we’re making decisions, how we’re coming to conclusions, and really getting ahead of any type of like audit issues or anything like that, where we’re really understanding our data and our workforce is really growing and developing along that path that would get us to the future of finance and the skill sets that are going to be required for people in a CFO organization.”

Relying on the innovation team

For both Hunter and Webbers, the continued move toward automation and using bots will be a balance of enthusiasm from the early adopters and managing those who remain cautious about it.

“One of the things that I believe helped us in that was just being open to hearing what they have to say about using bots. Every viewpoint was critically important to understanding the risk that may be associated with using a robotic process automation on a process that we had not proven out. So trying to take all of that input, letting them know that their input was important and mitigating the particular risks or accepting those risks, or coming up with a different approach to eliminate the risk, was our approach,” Webbers said. “At the end of the day, because I was in charge, I said, ‘Okay, let’s do it, we’re going to try and see.’ Fortunately, it worked out really well, and I think that those instances where we got those quick wins was important for people to gain confidence.”

She said the IRS looks back at every RPA implementation to create lessons learned and figure out where they can improve upon the process for next time.

Hunter added the CFO’s office created an innovation team to which employees can submit ideas for how to automate or improve a process.

“What I wanted to do was make folks a part of the process, where, I’ll steal a quote from procurement, can you tell me what you hate so that I can make you love it?” she said. “What are those opportunities that you just dislike doing every day? Let’s take a look at it because maybe there’s opportunity to do the work in a different way or automated it or whatever the solution may be, but there’s got to be an answer. We’ve focused on that as well as the change management portion of it as we’re thinking about how we’re looking at our work products. It’s really the mindset of how can a bot help me in my day-to-day life?”

For several years now, the Air Force, at its most senior levels, has recognized its policies and procedures have a strong tendency to force airmen to misspend their time on ancillary tasks instead of the ones they signed up for.

The service has had some success at reducing distractions like excessive computer-based training and performance evaluations, but it’s done less in the arena of bureaucratic administrivia. Things like paperwork.

Just like many other government agencies, the Air Force’s internal processes live and breathe via forms that have to be filled out manually, and each one takes time. How much? That was a difficult question to answer until the Virginia Air National Guard’s 192nd Wing saw a chance to do things differently.

“The information we were filling out on these forms was very repetitive: ‘Who’s your supervisor? What’s your social security number?’ And everything that they need to know on all these forms is already known about us. There’s technology out there that should be pre-populating this information,” said Chief Master Sgt. Joseph Young, who helped launch just such a technology solution while he served as the 192nd Operations Group’s senior enlisted leader. Young now serves as the senior enlisted leader for wing inspections, and has stayed involved in the process.

Early results from what the wing calls “HR Smart Weapon” are extremely promising. Its back-of-the-envelope math, based on a fair amount of real-world testing, suggests the Air Force could save about 156,000 person-hours per year if the approach they’re experimenting with was deployed across the whole service. And that’s for just one form.

The initial idea grew out of an Air Force initiative that tries to find “airmen-led” ideas that might cut costs, improve readiness, or give airmen back some of the time they waste on non-value-added tasks. Assisted by AFWERX, the service’s main innovation arm, the program allocates tens of millions of dollars each year to what the Air Force calls squadron innovation funds.

Young said the paperwork streamlining idea first came up during a training event in Las Vegas for how to spend those funds effectively. There, airmen were asked to pitch ideas, and the time suck surrounding manual form filling quickly emerged as the biggest pain point for the 192nd’s attendees.

“Even though they’re digital forms, it’s still paperwork,” he said during an interview for Federal News Network’s On DoD. “We have a particularly difficult time with the onboarding process. It can take upwards of three months from the time a person shows up for them to get their initial paycheck, in particular, for people who are coming from different organizations that are already in the military. No particular office is the problem. It’s the paperwork and the disparate databases that are involved, and they don’t necessarily talk quickly to one another. So we were trying to solve that problem.”

AFWERX suggested using a Small Business Innovative Research (SBIR) solicitation to ask vendors for help, and after whiteboarding the problem with several of them, the wing settled on Tackle AI, an Illinois-based firm that focuses its work on exactly this issue but had never done government work before.

“They specialize in extracting data from unstructured documents, so they could take all of our legacy forms, ingest them into their AI algorithm, and then create profiles for each individual,” Young said. “So the next time somebody would bring up a form, this profile information is automatically populated. And whenever the database didn’t know something about you from the ingest process, it would ask the user for that information. And once a user provided that information, you wouldn’t need to provide it again in the future.”

But it took a fair amount of back-and-forth between the 192nd’s testing team and TackleAI’s engineers to get the concept working for the Air Force’s needs. That’s partly because the wing wanted the technology solution to not just help fill out the forms, but also track their progress through each office who needed to view, approve, and send them up the paperwork chain.

They pulled that off too, Young said, and it was important that they did. After all, the inefficiency of the current, paper-centric system isn’t just a pain for the airman who first fills out a given form. Supervisors and various other approving officials need to see and sign off on them, and right now, that’s mostly done via PDFs attached to emails.

“Email is a terrible place to track things. Things get lost all the time, you have to sort through what’s important and what’s not, and it’s easy for things to fall through the cracks, especially when you’re trying to route and coordinate things. HR Smart Weapon makes it easy by showing you, ‘Hey, it’s got to go to these five offices, and this is the person who approved it at the previous office,” Young said. “So it helps you not only save time, but you can do analytics and see clearly whether a certain office is getting backlogged. Then, the commander can make decisions to shift resources to plus-up that office with additional personnel so they can be more efficient, or we can look into other ways of continually improving that process.”

The wing’s initial tests, or what it called a “bake-off” between the current, manual paperwork process, and the AI-assisted one, appeared to show concrete time savings: 35 minutes versus 20 minutes per form, on average.

And the forms were more accurate, which is key too. A simple typo in a manually-entered field can cause the whole form to get kicked back to the last person who approved it. Those kinds of errors happened 12% of the time under the manual process, but 0% of the time with HR Smart Weapon.

Granted, the tests were extremely small scale. Only five forms were tested for each method, and it’s extremely likely the broader Air Force would want to see much more testing before scaling it up. HR Smart Weapon would also need to make its way through various other approval gauntlets, like earning an authority to operate on DoD networks, no small feat.

Young said his team understands all that. It’s just a proof-of-concept for now, but one that’s worth the Air Force’s attention. And he acknowledged that even if the approach gets approved for operational use, there are going to be cultural adaptations needed too, partly around the question of whether commanders will be willing to trust algorithms to fill in boxes that used be completed by highly-trained human beings, no matter how mundane those box-filling exercises really are.

“I fully expect people to be hesitant, he said. “Even myself, it’s going to take me a little bit longer initially, because I’ll want to double check everything before I hit ‘submit.’ But if we have a level of confidence that this does what it says it does, and it’s accurate, I mean the sky’s the limit. As soon as we add more forms to it, we’re going to have that level of trust that it is filling out accurate and current information. That will help reduce time and give it back to our airmen and our commanders to do what matters most, and that’s to be in front of your airmen and not be encumbered with all of this paperwork. We can get after the mission by doing less paperwork.”

Patrice Tibbs, chief of community open source at the CIA, said open source has “proven itself over and over,” especially given current events like Russia’s invasion of Ukraine. OSINT is generally defined as unclassified information, often publicly available, like data gleaned from social media feeds.

“My five-year-old grandson understands the value of the iPhone and in communicating,” Tibbs said during a May 2 event hosted by the Intelligence and National Security Alliance. “If we can’t get on board and figure that kind of thing out now, and understand how that can be leveraged to make sure that we are clear in every country, every city, every home, in some cases, we will lose the lion’s share of any benefit we have in open source.”

Spy agencies have traditionally been organized around other forms of intelligence, like geospatial intelligence or signals intelligence. Agencies have struggled to define how OSINT fits into its broader tradecraft, but the array of public information about the Ukraine conflict has started to shift the conversation about OSINT intelligence circles.

The 2022 Intelligence Authorization Act is also pushing agencies to build more OSINT capabilities within the context of competition with China.

“The Intelligence Community must reorient to engage in a strategic competition with the PRC while countering China’s malign activities globally,” the Senate’s report on the Intelligence Authorization Act states. “To do so, it must continue to build open source intelligence capabilities and augment capacity; enhance sharing of intelligence capabilities; and strengthen the analytical and collection capabilities relating to non- military threats including technology competition.”

As head of the CIA, Director Bill Burns is considered the “community functional manager for open source,” meaning he reports to Director of National Intelligence Avril Haines on OSINT policy, requirements and funding, according to Tibbs.

The intelligence community also has a “National Open Source Committee,” which includes senior leaders from each of the 18 intelligence agencies. Within the committee, there are subcommittees specifically focused on issues like OSINT data, collection management, training and tradecraft, according Tibbs.

She said senior leaders are starting to take OSINT more seriously, providing a chance to set common standards around open source training and tradecraft

“The key for me is just understanding how we modify and change and adapt to the amount of data that’s available,” Tibbs said. “And because there’s not a consistency of how all of the different 18 organizations are utilizing, are capturing or are integrating open source into their workflows, there is inconsistency sometimes in how that is translated and shared and a variety of other things.”

Tibbs said the most difficult challenge for OSINT is “getting everybody to agree on the direction.” She also said ensuring the intelligence community has the technology infrastructure to support OSINT is important.

“Also, it’s just having the individuals with the right skill set and the motivation to come in and really take on these challenging roles, especially in the federal government realm,” Tibbs added. “This is this is not always the easiest place to work. It is not always the highest paid place to work.”

DIA open source center

The Defense Intelligence Agency is also looking to elevate OSINT through its Open Source Intelligence Integration Center. It was established in late 2019, and governs the military’s use of OSINT by leaning on standards, processes and tradecraft, according to Brad Ahlskog, chief of the center.

Compared to traditional forms of intelligence, Ahklskog said OSINT is now playing an “outsized role” in scenarios like exposing Russia’s build-up of military forces and subsequent invasion of Ukraine.

“I would argue that along with the amplified information warfare aspect of hybrid warfare, there’s a vital need for what I’ll call ‘hybrid intelligence,’” Ahlskog said during an April 26 presentation at the GEOINT conference in Denver. “It features a larger reliance on OSINT to both identify the threats, support deterrence operations, and provide the ground truth of the battlefield situation during the conflict.”

But he said the Defense Department will rely on contractors to help with a “data-centric approach” to OSINT. The approach will rely on artificial intelligence and automation to speed up the exploitation of such data, but it doesn’t come without some caveats.

“This translates to a need for datasets that are cataloged discoverable, or resident in government or commercial applications that are readily machine ingestible through application interfaces for transfer into other applications in formats that are not proprietary,” Ahlskog said. “Too often in the past, some open tools or applications were created with self contained datasets that were not easily blended with other data, or datasets from other sources that cannot be easily moved into a proprietary application.”

Furthermore, he said it will be crucial for analysts to understand how algorithms are analyzing the data and reaching specific conclusions.

“We want to produce actionable intelligence from them, and we must be able to easily understand and clearly explain how application ‘X’ produced information ‘Y’ that was used to inform decisions ‘Z,’” he said. “Without transparency into how OSINT data is obtained and processed, the intelligence community personnel will be very reluctant to rely on that PAI for intelligence purposes. Verification, validation and sourcing play integral roles for data in these cases.”

Beyond technology, Ahlskog said he thinks the culture around relying on OSINT is starting to change.

“I think more and more of our customers and personnel realize the value of open source,” he said. “They also get it first, frankly, in many cases. It’s very immediate. . . . It wasn’t possible 10 years ago, or 15 years ago, for information to be that widespread, immediately available to anyone who has a mobile device, or computer at their desk. So I think a lot of our customers, and our all source analysts and other collectors are getting much more comfortable with relying on open source early and often.”

The Biden administration has repeatedly called for equity considerations to be incorporated into new policies or programs, and more federal leaders are grasping what that means for their particular agencies.

For the Center for Medicare and Medicaid Innovation at the Department of Health and Human Services, equity is “the attainment of the highest level of health for all people,” according to its Healthy People 2030 strategic refresh. One of the refresh’s metrics for success is stronger data collection and intersectional analyses for populations defined by race, ethnicity, language, geography and disability, in order to identify gaps in care and develop interventions to address them. To do that, CMMI said it would require participants in all new models to collect and report data to identify and monitor impacts on health and the reduction of disparities, while existing models are incentivized to do the same.

All new models will also include patients from historically underserved populations and safety net providers, such as community health centers and disproportionate share hospitals — facilities that serve a significantly disproportionate number of low-income patients and receive payments from the Centers for Medicaid and Medicare Services to cover care to uninsured patients.

CMMI Deputy Director Arrah Tabe-Bedward said we’ve spent a lot of time really just trying to understand and start to collect data to understand what is the Center’s reach and how it can get more providers to participate.

“We think that there are incredible opportunities to do that in advanced primary care models and [accountable care organization] models. That sort of a structure, of course, requires that there is ample opportunity for information to be exchanged efficiently and effectively across providers and across care settings, in order to optimize the patient experience,” she said. “And so being able to get that right, as we are driving towards the very ambitious goal that we’ve set out for ourselves for 2030, to ensure that all of our Medicare beneficiaries, and the vast majority of those who are in Medicaid, are in those sorts of aligned care relationships with providers.”

Tabe-Bedward said CMMI wants to make sure there is technology to support those relationships and to ensure that the care being optimized.

Some communities lack experience with AI and machine learning, or actively distrust the technology, thus hampering its implementation. Susan Gregurick, associate director for Data Science at the National Institutes of Health, said this very problem was observed by NIH’s Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity (AIM-AHEAD) Program.

“In this case, federated learning is the right approach. And I think that’s one of my messages is that there’s no one hammer for all the many use cases … we really have to adopt and adapt our technologies for the communities and the research programs that we really want to address,” she said.

As the White House Presidential Innovation fellow at the Technology Transformation Services, Nina Walia is passionate about accessible data. For health care data in particular, the troves of PDFs and documents used by providers keep valuable information trapped, leading to a lot of redundant data entry.

“If we started to just mass and bulk and all adopting the idea of optical character recognition compared with computer vision, we could start to actually extract this data in an automated process so that this data can be machine readable,” she said.

The intelligence community is drafting a new data strategy for the first time since 2017, with a big focus on training a data savvy workforce well equipped to take advantage of an increasing deluge of information that intelligence agencies are both collecting and producing.

Intelligence agencies have made “great strides” since the first data strategy was published in 2017, according to Nancy Morgan, who just retired as chief data officer of the intelligence community. Her last day was April 29, and the office of the director of national intelligence has yet to select her replacement.

“We’ve made some significant improvements to what we’ve been doing with data lifecycle management since the first IC data strategy was published in 2017,” Morgan said in an April 28 interview on All About Data and Inside the IC. “We feel we’ve done a lot of work to enhance sharing and safeguarding, but there’s still more to do.”

Chief data officers across the 18 intelligence agencies are focused on using automation to do more data preparation, Morgan said. The goal is to give analysts more time to “do higher order tasks” rather than rudimentary jobs like data tagging.

“We’re collecting and producing more information than ever before, the IC is launching more collection capabilities than ever before at astounding volumes, certainly since I began my career 30-plus years ago,” Morgan said. “It’s just astounding how much information we’re gathering. So it creates a data volume challenge.”

ODNI is also updating the IC IT Enterprise, or “ICITE,” strategy, a major guiding document for how intelligence agencies will use computing in the years ahead. The work is being led by Adele Merritt, the chief information officer for the intelligence community.

The new IT strategy will be pivotal to “enhance the critical data management capabilities to achieve our goals,” Morgan said.

CDOs in the intelligence community are also looking to create more interoperability across intel agencies and the broader Defense Department. Morgan said leaders want to share successful approaches across organizations.

“How do we integrate and involve multidisciplinary approaches that solve the IC’s most challenging and emerging data issues?” she said. “We find new data challenges  every day in every domain area.”

But beyond technology, a major piece of the forthcoming data strategy is the workforce. Morgan said spy agencies aren’t just focused on bringing in highly sought-after data scientists, but also training the existing workforce to be more data savvy.

“How do we increase the data acumen and tradecraft, by not only attracting but developing, growing and resourcing the data savvy workforce?” she said. “So not just the talent we recruit, but the workforce we already have. How do we give people a chance to develop new skills and make them even more powerful and valuable to the community?”

Career pivots

IC data leaders are looking to create opportunities for intelligence professionals to start learning new skills related to digital technologies, data and cybersecurity, according to Morgan.

“It’s really very powerful when our domain experts learn some of the foundational skills for working with technology, working with automation, working with artificial intelligence, machine learning, being paired up with data scientists and data engineers,” she said.

The focus isn’t just on developing data professionals, but on building data aptitude across mission, business and policy areas, including acquisition, contracting, privacy and civil liberties, legal divisions and finance, according to Morgan.

“Frankly, it’s about supervisors, managers, leaders, senior executives at all levels of the organization,” she said. “Are we asking the right questions about data when it’s presented to us? Do we understand the data that’s driving our decision making and we say the words data driven decision making but how are we actually putting that into practice?”

Morgan noted the Fiscal Year 2022 National Defense Authorization Act requires the Office of Personnel Management to establish new occupational series for not just “data science,” but “data management” as well.

“I was really proud of helping influence some of the wording on that, because while I absolutely want to have a strong data science cadre, you need the full data management realm,” she said. “You need data managers, data policy experts, in addition to those data scientists and those data engineers.”

ODNI is also preparing to conduct the pilot phase of a new public-private talent exchange. It will allow intelligence officers to work temporarily in the private sector, and vice versa. The pilot phase will allow for six-month details, according to Morgan.

The pilot phase will include specific focus areas, including professionals working in data, as well as a category for artificial intelligence and machine learning, according to Morgan.

“Launching the pilot is a bit complicated, working through some of the security issues working through some of the acquisition and legal issues,” she said. “But our goal is really to help intelligence officers and private sector colleagues better understand each other’s mission, landscape, inject diverse thinking and gain new insights and really, hopefully create a more two way flow of talent skills and ideas.”

She also said it could help inculcate a culture where there’s more back-and-forth between the government and private sector.

“I don’t know that people will have the same sort of trajectory of a career that’s more only in the government or only in the private sector,” Morgan said. “I hope we’ll see more two-way movement and more continuous movement over the time of someone’s career. And again, selfishly, for me, this helps us grow our digital data and cyber savvy workforce with real world experiences.”

Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.

The Naval Surface Warfare Center Dahlgren Division works to ensure the reliability and safety and effectiveness of Navy shipboard weapons. All but a handful of its nearly 5,000 federal employees are civilians aided by several thousand contractors. For the second installment in this week’s series looking at Dahlgren’s activities, the Federal Drive with Tom Temin spoke with the division’s Technical Director Dale Sisson.

Interview transcript:

Tom Temin: Let’s just give the overview of Dahlgren from your standpoint as the top civilian engineer, if you will. This is not academic engineering at all, is it?

Dale Sisson: It is not. It’s definitely hands-on research, development, test and evaluation. We maintain a workforce of nearly 5,000 government civilians and another 4,000 folks from the contractor community that execute about a $1.9 billion portfolio annually, to deliver warfighting capability to our fleet.

Tom Temin: And the facilities here must be extensive, because of all the range of projects from stuff you can’t see like software to testing, firing, and so forth. Give us a sense of the range of what’s here.

Dale Sisson: Well, the range of what’s here is a great way to put it because we are anchored around the Potomac River test range, that includes really the length of the Potomac River to the mouth of the Chesapeake Bay, identified in certain increments that allow us to do open air testing on that range. So we fired ordnance down the Potomac River test range now for going on 104 years. And so that’s really the anchor for us as an organization, and where that hands-on research, development, test, evaluation and mindset started. Over the years that’s expanded greatly to include a very broad portfolio. And today is very software centric actually, you mentioned the software piece. Even hands-on work can be done in the software world. So we take great pride in having our technical workforce really ply their trade. So they have an opportunity to come here and work in their areas of expertise to do really critical mission work for the Department of Navy. So that includes the Navy and the Marine Corps. So as one of the 10 Warfare Center’s divisions within NAVSEA [Naval Sea Systems Command], we also serve as the Marine Corps’ warfare centers in a lot of ways as well. Our facilities, we’re larger than 300-plus buildings. So a tremendous footprint in that regard. That’s everything from internal areas to test electromagnetic effects, to software, collaboration, and coding spaces, the ability to do environmental tests and evaluation of our ordnance and weapons systems, and integrated warfare systems laboratory where we have the opportunity to bring combat systems online for the fleet. So we’re really all about doing very complex things in the naval warfighting environment. And we talk a lot about being the leader in warfare systems development and integration. To solving those complex problems that allow us to put ordnance downrange and hit an intended targets. There’s a lot of complex math, science, engineering behind that. And that’s really who we are.

Tom Temin: Yeah, so the original calculations done by ENIAC stopped there. And we have to keep working ever since then, huh?

Dale Sisson: Absolutely. And so as an organization, we’ve for many years been involved in advancing not only the computational scientific analytical capability, but the tools that go along with that. So many of the nation’s earliest computing advancements are a result of our work here. And it really was by necessity, as ordnance and weapons systems became more complex, and understanding the fly out of those weapons, trajectories of those weapons, and how to get them to more and more challenging targets. So really a lot of complex math and science behind that and engineering to make those systems real. And we’ve had to advance our disciplines, like systems engineering, for example, is a discipline that I think you can really say that NSWC Dahlgren Division has played a large part in developing that capability for our nation over the last 60 or so years, probably longer than that. So as we grow specific weapon capabilities, grow our workforce, we have to grow the disciplines within which we operate.

Tom Temin: And you mentioned a lot of how software has become important to calculating how all of these weapons work. Is there also research and development yet to be done? Or do you do it here on the weapons themselves? So the mechanical and chemical aspects of putting that shell on target?

Dale Sisson: Absolutely. There are a number of elements that goes into building a weapons system. We are very much a leader in the sensing side of things. So radar engineering, for example, and the development that goes on there. So understanding how to identify and track a target is a big part of weapons systems development and employment. So we really bring that to bear. Understanding the software side, as you mentioned, it really permeates now throughout everything. So one of the things we’re seeing is that the disciplines of engineering specifically, may not be as clear cut as they have been for so many years. We see the lines really blurring between mechanical engineers, electrical engineers, and computer engineers in particular, and the understanding that’s required for software development systems and employment and the hardware that goes along with that. So our folks really are very much multidisciplinary, though there’ll be founded in a specific area, but the complexity of the weapons systems really drives that. We’re not the ones that are working, the energetics formulation and things like that. We have our sister warfare centers that work in that area. But we’re really focused on that projectile, focus on the sensing element and then the fire control that it takes to make those two things meet up to a target.

Tom Temin: And what about industry and academia? There’s a lot of industry located physically near here. What role do they play? How do you interact with them? And is there also a grantmaking aspect to the research and development here?

Dale Sisson: Well, that’s a multifaceted question for sure. We’re about a $1.9 billion organization annually, as I believe I mentioned before. Of that over a billion goes outside of our gates to industry and academia. So the industry partnerships are huge. Whether it’s the local industry community here at Dahlgren, Virginia, or our location in Dam Neck, Virginia, of our 4,930 or so civilians, about 450 of those sit in Dam Neck, so down in the Virginia Beach area. So we have two primary localities that we are very involved with, for sure. So really rely upon those industry partners tremendously, and really also are heavily focused on the academic partnering side of things. That helps us on the academic front, really helps us to push the boundaries on technology, as well as grow our workforce. And we grow our workforce in a couple of different ways. It’s the growth of our existing team members and who we have. Typically we’ll have 300 or more members of our workforce and roles and graduate education at any one time. So we heavily invest in academic education in that way. We also want to have the leading professors from our nation’s leading universities here partnering with us on their areas of technical expertise. And then finally, we want to attract the best and brightest candidates out of these programs to join our workforce. And we know we had the opportunity to give them exciting work to grow their careers as scientists and engineers.

Tom Temin: And I wanted to ask you something specific that I know you’re working on. And that is the whole idea of autonomy, and of uncrewed vessels, in this case the surface, but I know it goes below the surface and in the air for that matter and other components of the Navy. But maybe discuss the idea of true autonomy, versus simply remote crew controls, or which look alike, maybe inoperational but they’re really not the same thing.

Dale Sisson: So we use the terminology of Intelligent Automation. Really one of our five technical strategic thrusts here. And the growth and evolution from what we would call remote control to autonomous systems is one that’s certainly iterative, and will be continuing for some amount of time for sure. With Remote Control Systems, you have the human in the loop, right? And that’s typically very important as we operate ethically in a warfighting environment. So as a nation, we’ll continue to evaluate how weaponized we want fully autonomous systems to be. But there are a number of ways of looking at autonomous systems. For example, a missile system has the ability to identify and reach targets and make some selection on its own there. We’ve worked in that kind of environment for many years. Some of the things that folks will talk about and maybe the more had been viewed as science fiction kind of realm that had the ability to come to reality, or automated soldiers, or sailors and marines and an airman and how we execute that. So how we weaponized some of those autonomous capabilities brings into our calculus, the system safety element. And so system safety is a big part of our portfolio. So we have a leading safety experts in these weapons systems to make sure we understand their operation. And the more autonomous a system gets, and the more decisions that system can make, the more it opens up your risk, the requirement on us to understand how to assess that risk, assume risk in the right places. As the United States of America, we operate under a set of set of ethics and a code of operation and how we not only look out for the safety of our personnel, but how we look out for the safety of those who may be impacted in a warfighting type environment. So all of those elements come into play. It’s not just simply the technology piece. We know a lot about how to do that technology piece. And we’re employing that for surface warfighting solutions. But now it’s also about rolling in the safety piece, the policy piece and how we employ such weapons.

Tom Temin: And as the top senior executive here, tell us what your day-to-day life is like, because if you wanted to visit one building a day, it would take you a year of working. So how do you know that the right things are happening? And what are some of your metrics that you watch to make sure everything is on track?

Dale Sisson: Yeah, well, that’s a great question. And one we’re always trying to figure out. Really the correct answer is that we have a tremendous leadership team here that surrounds me. I have 28 other executives that I directly supervise on the technical front as well as organizationally. So we have seven technical departments, three outstanding business related departments where we do contracts, all of our financial operations, human resources, security, and all of the things that are required to make our technical business tick. And so having that tremendous team around me is really what it’s all about. It’s really my role to set the vision and strategy for the organization, and then to work with that team to make it happen. So as the senior executive, I really get the good fortune of serving as the CEO of this operation and partnering with our commanding officer. And that requires a strong focus, not just on the technical front, but on the business front. I mean, we’re the stewards of $1.9 billion of taxpayer resources annually, not to mention all of the infrastructure that’s part of this organization. So we take that very seriously. Spend a lot of time in the metrics not only just on a technical front, but also on the business side. On the technical front. It really requires a strong personal engagement with the technical workforce. So I invest a lot of my time in doing that, whether that’s out in full day reviews with the technical departments and understanding the progress in specific technical areas. I am the steward of 37, what we call technical capabilities for the warfare centers. And so it’s my job to make sure that we within each of those technical capabilities that we really do have the technical expertise that we’re charged with maintaining and growing. And growing may or may not be in numbers, but it’s growing those disciplines that we’re responsible for. So there are a number of metrics and dashboards that we pay attention to. But it’s really about the one-on-one interaction with members of the team to see the progress that we’re making. You got to lay eyes on the systems and the technologies that you’re part of and have boots on the ground. And that’s the method that I employ.

Tom Temin: And when there are firing tests, do you go down and watch?

Dale Sisson: Sure, and we can do that in a number of ways, whether it’s in remote sites where we have full metrics and visibility and communication [in] what’s happening or actually standing out on the range. And our test range from firing systems down river to modeling and simulation events that are equally as exciting.

The National Geospatial-Intelligence Agency is looking to deliver software more like the tech industry under a new strategy that sets key metrics for both internal development teams and contractors.

“The NGA Software Way” lays out how the agency envisions delivering software faster and more consistently, as NGA’s technology priorities increasingly revolve around software-enabled capabilities like automation and machine learning.

Officials believe automation, artificial intelligence and machine learning will be key at NGA to analyzing a rapidly increasing volume of satellite imagery and other geospatial intelligence data that could overwhelm human analysts. NGA also recently took over Project Maven, a major AI program that’s been at the forefront of the Pentagon’s recent software development projects.

NGA’s new software strategy describes three key metrics as “availability,” “lead time for changes,” and “deployment frequency.” Each individual software product will have its own “product-specific metrics” as well, tailored to track how well the software is working for its users.

“We put this out for really anyone delivering software at NGA,” NGA Chief Technology Officer Alex Loehr said. “That could be government employees, industry, even commercial products that NGA is buying. There are significant parts of the software way that relate to how we want to work with those companies. And so we hope that this will set common expectations of how we can deliver useful software faster and for our mission.”

The software strategy complements the NGA’s recently released technology focus areas. The big priorities include assured positioning, navigation, timing and targeting; accelerated tasking orchestration; data access and data integrity; and analytic workflow modernization.

Loehr said the software strategy is an “implementation guide” for NGA’s technology focus areas.

“If the tech focus areas are the ‘what,’ the Software Way is ‘how,’” he said.

NGA wrote the “Software Way” based off of several existing documents, including the U.S. Digital Service’s “Digital Services Playbook,” as well as the U.K. government’s “Service Standard,” according to Loehr. The agency also looked to research and data from industry, specifically from the DevOps Research and Assessment, or “DORA,” a company owned by Google’s parent company, Alphabet.

Loehr said NGA took best practices from those documents and used them as a foundation for the software strategy, while taking into account the more unique needs of an intelligence agency.

“Some of those other documents are much more about citizen facing services,” he said. “At NGA, we do have some of those, but not everything we do is open and public. And so some of the elements from those other documents didn’t fit exactly, but we were able to build off the core of those documents in order to learn from those who came before and did a lot of really hard work and grow in a way that matches what we need at NGA.”

NGA published an initial version of the document last year and received more than 300 pages of responses from 47 companies.

“We got some feedback around things that were unclear, that didn’t make sense, as well as lessons that we learned about how we need to work at NGA and work with our industry partners to make this document successful,” Loehr said. “Some of that didn’t make it into the words of the document itself, but did start driving some work we’re doing to make sure that as we implement the NGA Software Way, we’re able to do it successfully.”

CORE developments

To help meet the goals of the strategy, NGA has established a Common Operating Release Environment, called “CORE,” to provide development teams with enterprise software delivery tools like version control, testing, and tracking and collaboration tools.

“Historically, we’ve let different teams choose their tools and their different processes of how they build software,” Loehr said. “That led to some really important things, but it also led to a lot of fragmentation. And what we’re trying to do is build one set of tooling and one set of processes.”

Many pieces of CORE are already in place and being used by mission critical applications in some cases, according to Loehr, including version control, the “CI/CD” pipeline, an API developer portal, and issue tracking and documentation spaces.

Enterprise workflow orchestration and messaging tools, respectively, are still “more in the beta phase,” Loehr said.

“The core of the CORE around the version control, the pipeline, the developer portal, all that is live, real and being used today,” he said. “And we are looking at growing that usage pretty significantly.”

Several years ago, NGA began developing an in-house software developer corp. Now, NGA is also looking to build out a key competency in the form of product managers who can shepherd a software project through development successfully.

“The person that acts as the interface between those end users and the development team and understands the vision for the product, creates the roadmap and makes sure that what is being built is actually both useful and actually used,” Loehr said. “That’s been a discipline that we are bringing into NGA, and then that we’re helping grow. I think will be really important for our future on how we make sure that we are building not just any software, but the right software, and it’s actually delivering on our mission.”

‘Build low, push high’

NGA is also increasingly developing its software in unclassified environments, called the “low side” in intelligence jargon, before it’s pushed to the “high side,” or a classified environment. The concept is “build low, push high,” according to Loehr.

“A lot of our workforce, and our contractor workforce doesn’t want to be in a [Sensitive Compartmented Information Facility] every day,” he said. “And also a lot of our software itself isn’t necessarily classified. The data that’s in it might be classified, and often not in all cases, but often our software isn’t.”

The CORE tooling includes the ability to sync software versions across classified and unclassified domains, Loehr said, a key process for speeding up development.

“Those process pieces are almost just as important as technology pieces,” he said. “And enabling us to build low and move high, I think will help us move faster and really increase the diversity that we’re able to have in the people working on our products and how that work gets done.”

DENVER — The National Geospatial-Intelligence Agency is evaluating the progress of “Project Maven” as it prepares to take over the artificial intelligence initiative and integrate it with a broader range of efforts to apply machine learning to geospatial intelligence.

The Biden administration is proposing to shift Project Maven to NGA as part of its fiscal year 2023 budget request. The program has been run out of the office of secretary of defense since its inception in 2017.

NGA Director Vice Adm. Robert Sharp said the agency would be “calling on industry” with regards to the Project Maven transition in the coming months. The agency has repeatedly stressed in recent years that it will need to use artificial intelligence and machine learning to process and analyze the fast growing amount of satellite imagery and other GEOINT data available from both government and commercial sources.

“We want to move forward together, so we can deliver GEOINT at the pace that our warfighters and decision makers need,” Sharp said during a Monday keynote address at the GEOINT conference here. “We have to be able to keep up with rapidly emerging digital trends. We have to be able to accelerate our ability to provide detections at the speed of mission, to give our customers tactical, operational and strategic advantage.”

NGA has been a partner to Project Maven since it started, helping to provide imagery and other data necessary for companies to train their algorithms, according to Mark Munsell, NGA’s deputy director of data and digital innovation.

And NGA has also been working on computer vision and machine learning projects, according to Munsell. He said NGA can “bonus off all the things” Project Maven has learned over the past five years and integrate the software projects into its own infrastructure.

“It makes a lot of sense to bring these things together,” Munsell said during a Tuesday media roundtable with reporters on the sidelines of GEOINT. “Doesn’t mean that we’ll always operate the same way that Maven has operated. We’ll do a really good assessment of what they’ve done. We bring in our subject matter experts, who are steeped in GEOINT for 30 years, some of the folks. They’re the ultimate customer now to be able to assess utility, assess how we can continue to support our combat support partners in the military, and make modifications moving forward to make it better.”

The agency is also hoping to avoid duplicative efforts, while sharing promising software across the military and intelligence components it supports.

“If you’ve developed an algorithm that goes after certain objects in certain geographies and biomes that is successful, we will capitalize on that as a community and ensure that if someone else in the community needs that, we can provide that as a service to them,” Munsell said.

Military AI pathfinder

Project Maven was established in 2017 “to accelerate DOD’s integration of big data and machine learning,” then-Deputy Defense Secretary Bob Work wrote in a memo at the time. The project fielded its first algorithms for processing images and video captured by surveillance aircraft later that year.

The program served as an AI pathfinder for the Pentagon, preceding the establishment of the Joint AI Center.

It also garnered major headlines in 2018 when thousands of Google engineers protested the company’s involvement in the project. The company ended its involvement, later stating it would not develop AI applications in the areas of weaponry or surveillance.

But Project Maven continued to grow well past the controversy. The Pentagon requested $247 million for the Algorithmic Warfare Cross Functional Teams, aka Project Maven, in fiscal year 2022 after receiving $230 million for the program in FY 21.  It has also been among the programs to pilot the use of a “colorless” software appropriation.

It’s unclear how much the Pentagon is requesting for the program in FY 23 as it shifts to NGA, because the intelligence community’s budget details are classified.

But the Defense Department’s budget documents provide some detail on how Project Maven has advanced in recent years. The program aims to “augment and automate” the processing, exploitation and dissemination for full-motion video feeds from a range of unmanned aerial vehicles, including “WAMI ISR,” which stands for wide-area imagery intelligence, surveillance and reconnaissance.”

The program is also fielding AI to automate analysis of military and commercial satellite imagery, according to the documents.

The program has also expanded beyond imagery in recent years. It also uses AI to exploit “CEM,” which stands for “capture enemy material” in intelligence parlance, as well as “maritime” intelligence and “PAI,” or publicly available information.

“Maven’s AI, deep learning, and computer vision algorithms and insights are developed for use in theater to detect, classify, and track objects within images (e.g., persons, vehicles, and weapons) as well as provide other insights, such as with CEM, text-based, and other projects,” the documents state.

NGA is taking over the project’s “GEOINT AI services and capabilities,” according to Sharp, and it’s unclear what will happen to the capabilities the project has created for other categories of intelligence.

With the transition not expected to become effective until Oct. 1, NGA is currently doing the “administrative work” to determine how to transition contracts, as well as get personnel and leadership in place, according to Munsell.

And even as NGA evaluates how Project Maven fits into its broader portfolio, Munsell also emphasized that there will be “no pause” in the project’s ongoing activities.

“The assessment is not a pause,” he said. “It’s our charge, as GEOINT functional manager, to help prioritize, to help others understand the investments, and then with data and statistics, offer opportunities to improve.”