5 Ways You Might Not Realize The US Air Force Is Using AI

The U.S. Air Force is incredibly technologically advanced. From weapons to tools and vehicles, it’s on the cutting edge. It’s unsurprising, then, that the branch is harnessing the power of AI to attempt to tackle some of its great historical problems, such as the safe and efficient storage of munitions.

This doesn’t mean that AI is necessarily always a positive thing to embrace. On the contrary, for instance, AI smart glasses can be a considerable danger where operational security and privacy are concerned, which is why the U.S. Air Force has banned this popular tech among servicepeople in uniform. This is where the whole core debate about the technology comes in. If AI is being used irresponsibly, there’s the potential for risk. If, on the other hand, it’s harnessed in innovative and practical ways, it can be an enormous force for good.

This is exactly what the air force is trying to do; incorporate AI in a variety of ways across its operations to save money, time, and lives, both inside and outside of conflict. From a whole new kind of pilot training to finding ways to manage scheduling the maintenance of a huge array of vehicles and components, AI absolutely has its place in the present and future of the U.S. Air Force. Let’s take a closer look at some of these creative applications for the technology.

The United States has one of the largest air forces in the world. This means, of course, that it has access to an enormous range of powerful assets, from helicopters to ammunition. By the same token, however, it can hugely complicate the issue of arranging, organizing, and storing them all so they can be accessed when needed. Many of us use convenient AI tools to manage our work days and boost our productivity, and that’s exactly what one U.S. Air Force initiative is implementing in this area.

An app known as Automated Master Storage Planning could revolutionize the way that these ever-present issues of efficient storage are approached. Virtualitics solutions leader Justin Shehane explained to Air & Space Forces Magazine“there are specific munitions that need to be kept a certain distance from one another, and there are some that do have to be stored together […] there are many, many governing guidelines and safety constraints that we absolutely have to adhere to,” Shehane explained to the outlet. What the app will do, then, is use what Shehane describes as a “base configuration plan” to take into account inputted details about storage requirements and the space available, and then return an in-depth analysis of the best way to position everything, available to download and refer to freely.

Virtualitics notes that its Automated Master Storage Planning Software is just one component of its Integrated Readiness Optimization suite. It has the capacity to provide 3D visualization of placement and storage spaces, “dynamically adapt storage plans to evolving mission requirements, safety constraints, and real-time operational data,” and determine where there are components or items in inventory that could be put to better use elsewhere.

A mistake made in warfare, a calculation that’s off by a second or even less, could be the difference between success and a fatal disaster. A safe and practical way of implementing AI into the planning and decision-making process, then, could help to remove the risk of fatigue or a momentary distraction from a human serviceperson. It’s partially for this reason, then, that AI is being trained to fulfil fighter jet functions like spotting enemies that human pilots can’t see, and distinguishing targets from allies in the process before a grave error is made.

In 2025, Shadow Operations Center-Nellis’ Las Vegas location was the setting for the second Decision Advantage Sprint for Human-Machine Teaming test. The Air Force notes that the exercise centered on “design[ing] AI-enabled microservices capable of assisting operators with the “match effectors” function, which determines the best available weapon system to destroy an identified target.” Several different teams took part, and data was collected from human battle managers’ work as well as those same efforts when AI systems’ algorithms are available.

It was found that the AI systems were much faster to find answers to problems, and provided about 30 times as many of them, but generating a long list of possibilities and having the awareness to choose the appropriate ones are very different. The decision-making itself, it seems, is best left to servicepeople who have a better understanding of their connotations, but AI models specialize in analysing data quickly, and that’s where the two seem to work so well in tandem.

Another huge downside of having a vast fleet of sophisticated machines at your disposal, unfortunately, is that there are going to be the constant logistical headaches of malfunctions and maintenance to deal with on a gigantic scale. To facilitate this for the U.S. Air Force, the Predictive Analytics and Decision Assistant, or PANDA, is a huge AI asset.

PANDA is a software that allows the technicians of the Air Force to focus their resources where they’re most needed. In May 2023, speaking to C3 AILieutenant Colonel Michael Lasher of the Rapid Sustainment office explained the concept: “And so what we’re trying to do is take advantage of all the data that we have, whether that be historical maintenance data or onboard sensor data, telemetry data — really anything that we have that’s useful for the purpose of formulating that evidence of need to perform the maintenance.”

The system makes use of C3’s AI Readiness application and Agentic AI Platform. It takes all of this information, far more than human technicians could feasibly work through manually, and translates it all into a plan that can inform engineers when work needs to be done or when issues arise. More than 16 different platforms are compatible with PANDA, which is accessed via Cloud One to help prevent unauthorized access. The data that the system uses is extensive enough for it to determine future potential needs for specific vehicle components, thereby looping in other specialists who monitor supply chains. It’s about helping ensure that parts are available when they’re needed as well as, by monitoring individual components, predicting when that will be.

Fighter jets are some of the branch’s big-ticket assets. As the tide of airborne warfare turns increasingly towards the destructive strengths of drones, though, it’s apparent that the strengths of these weapons will need to be harnessed too. For the U.S. Air Force, with its wealth of resources, it’s possible to combine assets to best effect. This seems to be the plan going forward with the force’s drone development program.

In March 2026, the U.S. Air Force tested an unmanned fighter as a ‘wingman’ for F-22 Raptor pilots. This endeavor was not the first of its kind. The previous year, XQ-58A Valkyrie drones accompanied pilots of a F-16C and a F-15E in a training flight above Florida’s Eglin Air Force Base.  Brig. Gen. Jason E. Bartolomei explained, in a statement shared by Air & Space Forces Magazine, that “by developing and integrating autonomous platforms with manned systems, we can quickly adapt, increase combat effectiveness, and reduce risk to our aircrews in contested environments.” Again, there’s an important paralel here in that these are semi-autonomous unmanned aircraft, meaning that AI is used to augment capabilities but a serviceperson retains ultimate control.

In August 2023, a press release from the Air Force Research Laboratory noted that XQ-58A Valkyries were flown in another trial at Eglin Test and Training Complex. ThIs marked, according to the press release, “the first-ever flight of AFRL-developed, machine-learning trained, artificial intelligence algorithms” on Valkyrie models. The Air Force Research Laboratory’s Autonomous Air Combat Operations team determined the algorithms that the Valkyries would follow, and successes and further development in this area will mean more sophsticated, safer unmanned weaponry.

Needless to say, being a student pilot at the United States Air Force is surely an intense, challenging experience, and with the adoption of new aircraft, equipment, and regulations, training in a broader sense never really ends. It constantly evolves, and unsurprisingly, AI technology has been at the forefront of some of these developments. These efforts begin with student pilots, and for them, the Air Force is working on developing a specialized chatbot it calls IP GPT.

The intent is for it to be trained exclusively on flight manuals and aviation-associated data, without taking in a broader range of documents, as a lot of the more broadly-focused chatbots are designed to do. As Lieutenant Colonel Seth Hoffman of the Flying Training Center of Excellence put it to Air & Space Forces Magazine“I want the whole gamut of what a pilot may have to interact with to be within this data pond, but then that’s it.”

It’s largely theoretical as of February 2026, but there’s no denying the boon this could be for both students and instructors if performance data, lesson plans and other information could be collated using the app. There are some other AI capabilities that are already in broader use in pilot training.

For instance, pilots need to be well-versed in the use of AI functionality relating to the aircraft they’ll be operating. Exercises to facilitate this are critical, and one example was held at the Lincoln Laboratory’s Beaver Works and the Computer Science and Artificial Intelligence Laboratory at MIT in August 2025. Participants used RACECAR, a platform developed by MIT for the testing of autonomous vehicles, to hone their skills and experience the unique nature of real-time decision making using AI input at the CSAIL Robot Apartment Living Lab.