Recently, two Air Force fighter jets engaged in a dogfight over California. One was piloted by a human, and the other by artificial intelligence (AI). This moment marked a significant leap forward for the U.S. Air Force in the development of AI technology, which has roots dating back to the 1950s. However, this demonstration is just a glimpse of what the future holds for AI in warfare.
The United States is striving to stay ahead of China in the AI arms race, particularly in the development of AI-powered weapon systems. The rise of AI in military applications has raised public concerns about a future where machines could autonomously select and engage targets without human intervention. U.S. officials assure that such scenarios will not occur under American control, but questions remain about what other nations, particularly potential adversaries, might allow. This uncertainty drives the U.S. military to rapidly advance its AI capabilities.
In the past two years, China has significantly ramped up its investment in drone technology, rapidly developing faster, smarter, and more adaptable devices for its navy, army, and air force. This aggressive push has not gone unnoticed by military analysts.
Unmanned aerial vehicles (UAVs) and other drone technologies are increasingly being integrated with manned assets in the operations of the People’s Liberation Army (PLA). As a result, the PLA now stands as the only military capable of competing tactically on a similar level with U.S. forces.
In May, Paris-based Naval News reported that satellite imagery of a dockyard near Shanghai revealed a “mystery ship” that seems to be purpose-built for hosting fixed-wing UAVs. If this is accurate, it could represent the world’s first naval carrier dedicated exclusively to drones.
More is known about the Type 076, an amphibious assault carrier currently under construction at the Hudong-Zhonghua shipyard in Shanghai. This vessel appears to be another global first, designed as a dual platform for both drones and helicopters.
Admiral Christopher Grady, Vice Chairman of the Joint Chiefs of Staff, emphasized the urgency of this development: “Whether you want to call it a race or not, it certainly is. Both of us have recognized that this will be a very critical element of the future battlefield. China’s working on it as hard as we are.”
The military’s use of AI has evolved from a combination of machine learning and autonomy. Machine learning enables computers to analyze data and develop rules, while autonomy allows these systems to act on those rules without further human input. An early example of this is the Navy’s Aegis missile defense system, developed in the 1960s and 1970s. Aegis could autonomously detect and intercept incoming missiles based on pre-programmed rules, though it lacked the ability to learn from its experiences.
A major breakthrough in AI occurred in 2012, when the combination of big data and advanced computing power allowed computers to analyze information and create their own rule sets. This moment, often referred to as AI’s “big bang,” enabled systems to achieve AI-enabled autonomy.
This month, Air Force Secretary Frank Kendall experienced this advanced technology firsthand during a dogfighting exercise at Edwards Air Force Base. He flew aboard Vista, the first F-16 fighter jet controlled by AI. While Vista represents the forefront of AI development, hundreds of AI projects are currently underway across the Pentagon.
At the Massachusetts Institute of Technology (MIT), service members have been clearing thousands of hours of recorded pilot conversations to create a dataset that AI can use to distinguish between critical and non-critical communications. The goal is for AI to prioritize important messages, such as runway closures, ensuring they are addressed promptly.
Another significant project focuses on developing an AI-based alternative to GPS navigation. In future conflicts, high-value GPS satellites may be targeted or interfered with, potentially disrupting U.S. military operations. The Air Force has been testing an AI program that utilizes Earth’s magnetic fields for navigation, a method that has proven effective in preliminary tests.
The Vista AI system, still in its training phase, operates with several safety measures. Mechanical limits prevent the AI from executing dangerous maneuvers, and a safety pilot can override the AI at any time. The AI does not learn during the flight; instead, it is updated with new data after each mission, allowing it to improve its performance in subsequent simulations.
As the AI continues to learn and adapt, it has already demonstrated its ability to outperform human pilots in some dogfighting scenarios. However, safety remains a top priority, with strict controls over the data used to train the AI to ensure it reflects safe flying practices.
The ultimate goal for the Air Force is to develop an AI system capable of managing a fleet of 1,000 unmanned warplanes currently under development by General Atomics and Anduril. This ambitious project highlights the military’s commitment to integrating AI into its future operations, preparing for a new era of warfare where AI plays a central role.
