Four Reserve Citizen Airmen are part of a groundbreaking partnership between the Air Force and the Massachusetts Institute of Technology to advance the use of artificial intelligence – computer systems that are able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making and translation between languages.
In 2019, the Department of the Air Force joined with MIT to create the USAF-MIT Artificial Intelligence Accelerator. The venture brings together enlisted Airmen, officers, students and MIT faculty members and scientists who conduct cutting-edge AI research.
“Developing AI is absolutely critical for our military advancement,” said Col. Tucker “Cinco” Hamilton, the active-duty AF-MIT AIA director. “It is key for us to use the immense amount of information that is being collected through our personnel systems, our war machines and our satellites. While AI can unlock this data and create efficiencies and capabilities that can transform our systems, we must also understand that algorithms are somewhat agnostic to their role.”
An algorithm that is supporting finance systems one day could be in a war machine the next, Hamilton said. This amplifies the need to develop AI ethically and transparently. Additionally, Hamilton said since data drives AI, the way government systems collect, format and protect data is essential in AI and machine learning development.
Reserve members of the team bring a unique perspective to the AI Accelerator since they have experience in both the military and civilian innovation worlds. Lt. Col. Eugene Keselman, the individual mobilization augmentee to the USAF-MIT AIA director, for example, is the executive director of the MIT Innovation Initiative as a civilian.
“I’ve been in the Air Force for a bit – 12 years active duty and eight years Reserve. I think I can provide a good operational, administrative perspective to our director,” he said. “My knowledge of the systems, culture, processes and personalities helps in managing the relationship between the Air Force and MIT.”
As the Air Force’s first and only uniformed unit dedicated to the advancement and delivery of artificial intelligence, Keselman explained that the AI Accelerator’s responsibilities stretch across three main lines of effort.
“The first is to execute flagship AI projects and related work on behalf of the department,” he said. “The second is to develop scalable AI education, and the third is to lead dialogue in AI ethics and safety. This is all pursuant to the cooperative agreement between the Department of the Air Force, MIT and MIT Lincoln Lab (a Department of Defense-funded research and development center chartered to apply advanced technology to problems of national security).”
Maj. Jonathan Hudgins, an IMA who serves as the strategic advisor to the AI Accelerator director on outside collaboration and strategy, said most of the team’s current projects are active-duty Air Force-focused, but that could change.
“One project the team is working on right now is a project designed to utilize AI-powered magnetic navigation techniques for GPS (Global Positioning System)-denied environments,” he said. “This is especially important for Air Force aircraft that may have to operate in contested environments. If we were able to get more Reservists on MPA days, we could focus on Reserve-specific projects in the future.”
Lt. Col. Shannon Mann, an IMA who is serving as the AI Accelerator public affairs officer and marketing manager, said the Reserve is well-positioned to take on this special mission.
“The civilian experience in artificial intelligence and machine learning that our Citizen Airmen wield as part of their everyday jobs is of huge benefit to the future of this mission,” she said.
“What the AI Accelerator is doing can set the standard for the entire Air Force and Air Force Reserve,” Keselman said. “This isn’t just about the specific AI technologies coming out of MIT. This is about creating a new integration model for academia and government, specifically around tough tech like AI.”
In terms of readiness, Keselman said the overall goal is not just to be ready, but to be “ready for what” in a changing world.
“A simple mantra for us here at the Accelerator is ‘data is all around us, information wins and knowledge rules all,’” he said. “Whether it’s related to pilot training, pilot scheduling or more performant and accurate systems, we want to demonstrate the value of using data-driven insights to answer questions.”
Keselman explained that the opportunity for the Air Force Reserve within the AI Accelerator is two-fold. First, it can help establish the model for use in the Air Force and the entire government on how to integrate with the world’s foremost research institutions and properly execute technology translation and transition. Second, it can show how to utilize the Reserve structure to truly optimize the process by bringing to bear expertise and networks that are not endemic to the Air Force otherwise.
“Research shows diversity of viewpoints leads to better outcomes,” he said. “And the Reserve embodies that ethos.” #ReserveReady #ReserveReform #ReserveResilient
(Fitzmorris is assigned to the 302nd Airlift Wing public affairs office. She was temporarily assigned to the Air Force-MIT AI Accelerator when she wrote this article.) ■