AIAI-Controlled Fighter Jets May Be Closer Than We Think — and Would Change the Face of Warfare

By Arun Dawson

Published 17 April 2025

Could we be on the verge of an era where fighter jets take flight without pilots – and are controlled by artificial intelligence (AI)? US R Adm Michael Donnelly recently said that an upcoming combat jet could be the navy’s last one with a pilot in the cockpit.

Could we be on the verge of an era where fighter jets take flight without pilots – and are controlled by artificial intelligence (AI)? US R Adm Michael Donnelly recently said that an upcoming combat jet could be the navy’s last one with a pilot in the cockpit. That marks a striking, if not entirely surprising, shift in thinking about the future of aerial warfare.

The US Navy is not alone. Other programs to develop next generation fighter jets are also touting uncrewed options as a distinct possibility.

However, we have been here before. Senior leaders in the US Navy said they believed the last crewed fighter jet had been procured in 2015. As far back as 1957, premature obituaries were being written for the fighter pilot era. So is there anything different now?

The ability of a fighter jet to maneuver, accelerate, and maintain high speeds, crucial for air combat, is called kinematic performance. Estimates are as high as 80% on how much pilots reduce kinematic performance. Though this figure may be disputed, there is no question that uncrewed aircraft enjoy several key advantages.

Without the need for life support systems such as ejection seats and oxygen supplies, these aircraft can perform in ways that are beyond the scope of piloted aircraft. But additional trends are pushing militaries to reconsider the role of the human pilot altogether.

Systems enabled by AI are already demonstrating superior performance in military exercises. In existing remotely piloted aircraft, a human operator remains in control. This model is known as “human-in-the-loop”. AI is now enabling the possibility of human-on-the-loop (where humans take a step back, supervising and intervening if necessary) and even “human-out-of-the-loop” systems (in which AI selects and engages targets autonomously).

The latter category, while controversial, may offer decisive advantages. In scenarios where milliseconds matter, a fully autonomous system could outperform any human operator, to the extent that senior defense leaders have expressed a willingness to trust AI with lethal decision-making under certain conditions. Others add that autonomous systems could adhere more rigorously to the laws of armed conflict compared with a human operator.