The Air Force is moving “as quickly as possible” to integrate artificial intelligence and machine learning tools into its efforts to improve readiness levels and prepare for future conflicts, according to Chief of Staff Gen. David Allvin.
As part of its plan to “Re-optimizing” for great power competitionThe service is revamping how it drives readiness for contested environments in the Indo-Pacific by conducting new large-scale exercises, improving supply chains and more. Through those initiatives, the Air Force is exploring how AI and ML can improve readiness by improving command and control of operations and how the organization accounts for spare parts, Allvin said Wednesday during a roundtable with reporters.
But to do that, the service must integrate data from various siloed systems, he noted.
“I think one of the things we probably regret the most is the way we’ve grown our datasets and our individual networks in a customized and organic way, which makes it hard to connect them,” Allvin said. “The data is all there, you have to put it together, develop the algorithm, and get the insights that will help us learn faster.”
The Air Force is exploring command and control needs and the role AI can play based on its recent Bamboo Eagle events, a new series of exercises designed to test the service’s agile combat employment (ACE) concept that aims to disperse combat aircraft deployed from smaller bases while working closely with fleet mobility aircraft.
The exercises demonstrate what the Air Force must do to effectively execute this deployment and where it needs to improve, he noted.
“Part of that involves better assimilation of situational data to be able to say, ‘At this point in time, what do I know that I didn’t know when I made the plan?’ so that we can update the plan and perhaps reorganize how subsequent operations might unfold,” Allvin said.
The Air Force is planning another large-scale exercise, known as Return of Forces to the Pacific (REFORPAC), scheduled for summer 2025. Building on Bamboo Eagle, the exercise will focus on the service’s mission readiness if it were to deploy to a complex, contested environment, particularly how it would maintain operations and conduct logistics in the event of an attack.
While the challenges of AI in command and control efforts still need to be addressed, the technology is a bit more mature for how the Air Force understands and prepares its spare parts inventory. The service now has a better understanding of its individual weapons systems and predicts where they are, when they might fail and what additional parts it needs to have on hand, Allvin said.
“These are things that we can play with in advance and see whether this particular algorithm or this particular methodology worked and how much it increased the operational impact,” he told reporters.
The service already has some data analytics tools and platforms that help it make more informed logistics decisions, such as the Basing and Logistics Analytics Data Environment (BLADE) platform.
Allvin said the system has improved the service’s understanding of its inventory and other logistical factors that affect readiness levels over the past year. Going forward, the Air Force wants to use these data analysis tools to better identify which specific parts are most likely to fail and when.
“In the past, we’ve kind of lumped it all together in weapon system sustainment, and it’s kind of a peanut butter spread across a lot of different weapon systems, so everybody’s a little bit healthier (and) a little bit sicker,” he said. “When we try to … cherry-pick that and be more specific about whether you have access to those parts in a given time frame, it’s going to make that particular platform, or that set of platforms, more mission-capable, more quickly. It’s just more specific.”
