Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.
There are two overarching themes to the Defense Department’s desire to use artificial intelligence: Agility and speed.
The Pentagon wants agility and speed for warfighters as much as they do for back-office functions, which is why DoD believes its new AI strategy and Joint AI Center (JAIC) will deliver on those themes.
With the release of President Donald Trump’s AI executive order on Monday, DoD took the wrapping paper off its new center and corresponding strategy. Dana Deasy, the DoD chief information officer, said AI is a key component to the National Defense Strategy (NDS), released in March.
“As stated in the AI strategy, the United States together with its allied partners must adopt AI to maintain our strategic position, prevail on the future battlefields and safeguard free and open international borders,” Deasy said during a briefing with reporters on Feb. 12 at the Pentagon. “One of our central aspects of our AI strategy is the need to increase the speed and agility with how we will deliver AI capabilities across every DoD mission.”
The JAIC, for which DoD did a soft launch in June, will be the main artery delivering AI governance, tools, standards and more to all military services and DoD agencies.
Air Force Lt. Gen. Jack Shanahan was recently named director of the JAIC, coming over after running DoD’s Project Maven AI program for the last two years.
“There are a few critical mission themes regarding the JAIC that are reinforced in the unclassified summary,” Shanahan said. “First, to accelerate the delivery and adoption of AI capabilities across DoD. This underscores the importance of transitioning from research and development to operational fielded capabilities. The JAIC will operate across the full AI application lifecycle with emphasis on near term execution and AI adoption.”
The next critical mission theme is to establish a common foundation for scaling AI’s impact.
“One of the JAIC’s most important contributions over the long term will be establishing this common foundation, enabled by enterprise cloud with particular focus on shared data repositories, reusable tools, frameworks and tools and cloud and edge services,” Shanahan said. “Third, to synchronize DoD AI activities. There are a lot of AI and machine learning-related projects ongoing across the department; it’s important to ensure alignment with the National Defense Strategy. And finally, to attract and cultivate a world class AI team.”
Shanahan said the center launched two pilots and one is around humanitarian assistance.
“This is one we would put under the category of an open AI mission. You will have people who are spring-loaded to work with the DoD on this, whereas on some of the other projects they might have some concerns,” he said. “We are bringing in everybody from [non-governmental organizations to the Department of Homeland Security] to FEMA. DoD is not the lead in these areas, but this is a project where everybody has something they want to contribute to.”
Trump signs executive order fostering artificial intelligence R&D in government
He said the first project is helping to identify fire lines based on the experience with the recent California wild fires.
“This isn’t a project the JAIC came up with on its own. This was a pressing need for the state and federal governments and all the people who are involved in disaster relief and humanitarian assistance,” Shanahan said. “We are starting with that, full motion video, then getting into still imagery and then getting into wide area full motion.”
He said the pilot also will test out how command and control and firemen on the front lines can access and use that information collected by the AI solution.
Deasy added DoD and its partners will access the lessons and the tools developed under this effort for other initiatives.
For the second pilot, DoD is applying AI tools to preventive maintenance for the H-60 helicopter.
Shanahan said the Special Operations Command needed some help to further its business intelligence initiative and that’s how JAIC got involved.
“They already had data coming off their helicopters that turned out to be pretty good and ready to train an algorithm against. That was our starting point with SOCOM and then with the Army with their AI task force, and then going to the other services because the Air Force and Navy have the same helicopter fleet,” he said. “I don’t want to make it sound like this is a massive solution to a whole bunch of different platforms.”
Shanahan said the purpose was to get early practice working on the AI pipeline with a known question from SOCOM and their helicopter fleet, and then, using an algorithm, to do certain failure mode detection and predict when that might happen. From there it will scale out.
“There are a lot of different predictive maintenance efforts across all the services,” he said. “We are just trying to get our arms around who is doing what and then how can we expand from this initial project.”
JAIC supports all military efforts
Shanahan added he expects to see the initial results from both pilots sometime over the next six months or so. These pilots, at least for the initial launch of the center, are exactly the approach DoD wants to use when it comes to AI.
Deasy and Shanahan said the JAIC isn’t spearheading the effort but rather supporting other parts of DoD.
“This was a great example of all the services had this common problem. They all use this asset. Data was already coming off this asset, and we could quickly get the services to agree on the problem that we were trying to solve,” Deasy said. “This was a great example of how JAIC was used in the way we see it being used in the future.”
Shanahan added the most important factor in any discussion about using AI, or any technology for that matter, is understanding and agreeing to the problem one wants to solve.
“It’s not AI for AI’s sake. But how does this feed into a business intelligence and data analytics question? It goes well beyond AI,” he said. “I think that’s important. It’s an enabler, it’s not the thing just to say we are doing AI.”
The JAIC will explore other potential pilots in 2019 and into 2020.
“We are beginning to work, and this is just in the early stages, with Cyber Command on a cyberspace related national mission initiative. And at the same time, we are remaining closely integrated with Project Maven, which is a component initiative under the undersecretary of Defense for Intelligence,” Shanahan said. “We also are exploring other substantial initiatives in fiscal year 2020 that could, just as an example, accelerate operations intelligence fusion, command and control and decision advantage for operations in a peer and near-peer competitor environment.”
Shanahan offered a little bit more detail on the work with Cyber Command. He said they are exploring applying AI tools to event detection, network mapping and compromised accounts. They will propose future projects based on such criteria as mission impact, speed to delivery, resource ability and the ability to contribute to a common foundation.
“Likewise, we want to form teams that will work with the services and combatant commands on potential component initiatives,” he said. “At the same time, we want to identify some smart automation initiatives that could provide near term dividends in terms of increased efficiencies and effectiveness for back-office functions.”
Deasy said he expects the JAIC to really pick up steam in 2020 when full appropriations come through. Currently, DoD is funding the JAIC by moving money around. Shanahan said the JAIC’s budget is under $90 million for 2019 and is mostly staffed by detailees.
But Deasy said by having Army, Navy, Air Force, Marines and other DoD components provide the staffing, the JAIC’s impact and value will be felt more immediately.
“It’s a real simple concept: Bring people in, give them a place to work and learn over the next two years and actually send them back out to their respective services with an enormous amount of talent and capability and know-how of what JAIC can do,” Deasy said. “It’s going to be made up of civilians, academics, the commercial world and the service side. We think that service side provides that linkage.”
Deasy said these detailees also will help ensure the services don’t ignore or overlook the JAIC’s role in governance or buy tools that are redundant.
With the president’s EO and now the official launch of the JAIC, Deasy said DoD is more than ready to begin working with the private sector and across the government to harness the potential of AI in every mission set.