ROBOTICS AND AUTONOMOUS SYSTEMS
Algorithmic Warfare: Army’s AI Task Force Making Strides
Nestled alongside the Allegheny River is Pittsburgh’s famous Robotics Row — a grouping of tech companies focused on building and advancing autonomous systems. The location is also now the new home of the Army’s Artificial Intelligence Task Force.
Based out of Carnegie Mellon University — one of the world’s leading academic institutions in robotics — the task force has been humming since it was officially launched in February, said Col. Doug Matty, the organization’s deputy director.
The group — which was established under a directive signed by then-Secretary of the Army and now Secretary of Defense Mark Esper last October — was created with a two-fold mission in mind, Matty said.
One was to ensure that the Army is aligned in supporting the Pentagon’s Joint Artificial Intelligence Center, or JAIC, that was stood up last year as a way to coalesce the Pentagon’s disparate AI efforts, he said. The second was to serve as an integrator and synchronizer for the various science and technology and developmental engineering AI efforts across the Army, Matty told National Defense.
The task force has four main focus areas, he said.
“In the Army, the most important resource we have are our people,” he said. “It’s no surprise that the first task that they asked us to look at is how can we leverage either research and development that’s ongoing or existing capabilities in industry to help align our human resources and talent management.”
Matty’s organization will be working alongside the Army’s Talent Management Task Force to address that challenge, he said.
The next task is working collaboratively with the Joint AI Center, which broadly focuses its efforts on national mission initiatives and component mission initiatives.
“One of their ... national mission initiatives is predictive maintenance,” he said. “They asked if the Army would … take the lead as an executive agent-type role for the work on predictive maintenance.”
The Army selected the H-60 helicopter for the effort as it is a joint platform across the services, he said. Work is currently underway to see how AI can benefit the system.
The third focus area requires the team to work closely with Army Futures Command, which was stood up last year to get at the service’s top modernization priorities. The task force is specifically looking at ways to help with the No. 1 and No. 2 priorities, long-range precision fires and next-gen combat vehicle, respectively.
To do so, the organization is building on previous work done by the Pentagon, such as with Project Maven, which developed platforms to analyze drone footage, Matty said.
“As you can expect, there’s a lot of congruent types of efforts that they have that we could leverage,” he said.
The Army is working to apply AI to the entirety of the long-range precision fires challenge, he noted. “While the folks at Fort Sill, [Oklahoma], are predominantly focused on … the physics aspects of that mission thread, we’re also working the front-end piece in terms of finding and identifying and assessing and enhancing our situational awareness,” he said.
The fourth task is automated threat recognition to enhance the Army’s situational awareness while moving toward autonomous operational maneuver platforms, such as next-gen combat vehicles, Matty said.
The task force is benefitting greatly from working alongside researchers at Carnegie Mellon, he said. “They have both the breadth of technical expertise and depth in terms of capacity to conduct this type of research-and-development effort.”
The school’s Robotics Institute has nearly 1,000 researchers and faculty, Matty noted. Additionally, located within the institute is the National Robotics Engineering Center, which is also a helpful resource, Matty said.
While the task force was only stood up several months ago, Robotics Row academic institutions and companies have been working with the Pentagon for some time, he noted.
“The university and the surrounding ecosystem has been significantly engaged in supporting Department of Defense research-and-development initiatives for a number of years,” he said.
The task force — which is being led by Brig. Gen. Matthew Easley — is a blend of government officials and a cadre of individuals with technical expertise in AI. The organization has an authorization of 18 personnel and is looking to hire one more civilian before it is fully staffed, he added.
As the organization works to integrate AI into more weapon platforms, it knows that the technology is an enabler but not a silver bullet.
“When folks think about artificial intelligence, they tend to think explicitly about the magic algorithm or it’s some Voodoo-like thing that happens in the cloud,” Matty said.
But much like the Defense Department embraces the idea of doctrine, organization, training, materiel, leadership and education, personnel, and facilities, or DOTMLPF, to accomplish a mission, likewise the AI task force sees the need for a blending of different perspectives to generate a solution, he said.
To that end, the Army follows what it calls the “AI Stack,” a framework — which is popular at institutions such as Carnegie Mellon — that emphasizes the need to understand everything from the computing layer, the sensors, the data management, the algorithms as well as decision support, he said.
Ultimately, the Army’s approach for artificial intelligence is to enhance its soldiers, he said. “That covers the whole technology stack but also includes how much of that artificial intelligence is going to be … autonomy as well as the nuanced human-computer/human-machine interaction.”
Guiding all of that will be ethical considerations, Matty noted.