Automation based on artificial intelligence combined with human supervision to manage hundreds of aircraft and their weapons against as many targets will be critical to warfighting in the future, Air Force Secretary Frank Kendall said last week.
“And we have to do that as efficiently as possible,” Kendall said December 2 at the Reagan National Defense Forum. “That is something that with emerging technologies, and we call it AI but it’s automation basically, still with human oversight, you can aspire to actually manage those fights.”
Kendall said the fights are multi-domain, air, space, and surface. Sensors in space provide a picture of the threat in peacetime, giving whoever starts a war an immediate advantage by being able to engage targets before an opponent can effectively respond, he said.
The U.S. will not start a war, which puts a premium on having a space-based battle management system to track its platforms and weapons and the adversary’s platforms, weapons, and other targets, Kendall said.
“That helps our operators decide with a huge degree of automation, how to respond when an attack occurs, both from the point of view of protecting our own assets and going after the other sides’,” he said.
The integration challenge does not mean everything can be effectively integrated, Kendall cautioned. Instead, he said there must be a “reasonable technical goal that’s achievable, that’s quantifiable, and then set out to built that.”
Brian Schimpf, co-founder and CEO of Anduril Industries, said a problem with the Department of Defense is that it only has a few programs that involve mobilizing the industrial base around using AI. He highlighted the DoD’s new Replicator Initiative that is aimed at acquiring thousands of low-cost, autonomous unmanned systems as part of a potential future war against China in the Indo-Pacific region.
Schimpf and Kendall were both participating on a panel focused on fielding autonomous defense systems. Schimpf said the command and control of thousands of aircraft will be a tough challenge.
“How are you going to have these multi-domain effects be able to be coordinated and effective?” Schimpf asked. “And the reality is…it’s nobody’s problem in the department because it touches so many different parts of the department. You end up having this sort of aspect of this really kind of scattered across all the different PEOs, which are focused on just buying the capabilities that they need to buy.”
Maven, a computer vision program managed by the National Geospatial-Intelligence Agency to sift through vast quantities of imagery to identify potential targets for combatant commanders, is the last AI program developed at scale, he said.
AI efforts since then have “been very fractured,” he said.
Schimpf said that Kendall’s creation of more than a year ago of a new Program Executive Office for Command, control, communications, and battle management (PEO-C3BM) is a step in the right direction to help sort out the command-and-control issues associated with Replicator and future warfighting needs.
Kendall said the new office, which is led by Brig. Gen. Luke Cropsey, is “making great progress.”
However, Kendall warned that congressional delays in approving a federal budget for fiscal year 2024 are threatening this progress. He said Cropsey’s budget is supposed to double in FY ’24 but “he will not make that progress if we don’t get that money. If we’re going to take advantage of AI, if we’re going to take advantage of the automation that we can get out of AI technologies, we’ve got to put the resources against it, invest in real products are going to give us real operational capability. That’s the path we’re on. I think we’ve got a good solid path to do that.”
The federal government is currently operating under a continuing resolution that maintains funding at FY ’23 levels. The fiscal year began October 1.