This is an “innovation competition”—Lockheed (LMT) and the league are challenging teams to develop an AI that can beat human drone pilots without human intervention. It’s called the AlphaPilot Innovation Challenge and there’s $2 million in prize money on offer, with $1 million set aside as the grand prize for the developer of the winning AI.

“Your challenge is to design an artificial intelligence and machine learning (AI/ML) framework capable of flying a drone through several professional drone racing courses without human intervention or navigational pre-programing,” reads the challenge’s webpage on Lockheed’s site. “By participating in this competition, your knowledge and ideas can contribute directly toward the future of autonomous transportation, delivery, disaster relief, and even space exploration!”

All of which sounds great, but might the resulting technology not also be used in a military context? Many techies don’t like that prospect—after all, Google’s employees successfully revolted over the idea of their AI efforts being used to power military drones’ video analysis, and there is widespread pushback within the AI community against the concept of autonomous weaponry.

Drone Racing League (DRL) founder Nicholas Horbaczewski told The Verge that the league’s new Artificial Intelligence Robotic Racing (AIRR) circuit, where the challenge will play out, “has no ties to the military.”

“To suggest that advancing AI piloting would be intrinsically linked to the military would be very short-sighted,” Horbaczewski said.

Lockheed said it was funding the challenge using the money it saved from President Donald Trump’s tax reform last year, which slashed corporate tax bills.

Lockheed hardware apparently won’t be involved, but Nvidia (NVDA) is providing its Jetson platform—essentially an embedded unit for high-performance data processing—for the AI’s development.