Exploring AI’s Role in Modern Aerial Combat
I’m flying a Typhoon fighter over the Irish Sea, facing a significant challenge: a hostile jet is on my tail. Despite my best efforts with the joystick and throttle, the enemy remains in pursuit.
Simulated Combat with AI
The Black Triangle Challenge
The enemy is represented by a black triangle on the computer screen, part of a series of three simulated dogfights. Unfortunately, I lose 3-0 to this ominous AI-driven opponent.
AI Development and Learning
This AI dogfighter, developed by Turkish aerospace engineers and commissioned by BAE Systems, is designed to assist real pilots in combat. Emre Saldiran, one of the engineers, explains how the AI learns through trial and error, reinforced with extensive data inputs.
AI’s Potential in Combat
Enhancing Pilot Efficiency
One of the AI’s objectives is to alleviate the information overload that fighter pilots often experience. Mevlϋt Uzun, another engineer, highlights how the AI can offer real-time tactical advice, such as adjusting speed or avoiding high-risk engagements, based on a quick assessment of the situation.
Real-World Applications
Currently, the AI program runs on a normal laptop PC, and similar AI technology is being tested by the US Air Force. In 2023, the US demonstrated an AI dogfighter flying an F-16 jet, showcasing years of development aimed at creating an AI that can outperform human pilots.
International AI Competitions
The AlphaDogfight Trials
In 2020, eight US AI companies participated in the AlphaDogfight Trials Event, simulating online dogfights between AI programs and a seasoned USAF pilot. The winning AI repeatedly defeated the human pilot, leading to further development under the Defense Advanced Research Projects Agency (Darpa).
Rapid Evolution and Integration
The AI developed by Darpa has evolved rapidly, but safety pilots are still required for takeoff and landing. The challenge now is for the AI to gain trust and integrate seamlessly with other forces.
UK vs. US AI Development
Focused Approach in the UK
The UK AI dogfighter, developed at Cranfield University, focuses solely on combat moves, allowing for rapid development. Dr. Uzun notes that this streamlined approach enables faster progress compared to the more comprehensive US projects.
Real-World Testing
Despite these advancements, AI dogfighters are not yet a match for experienced human pilots. Ben Westoby-Brooks, a former RAF pilot, successfully defeated the AI in simulated combat, though the AI’s potential to assist in real combat scenarios remains promising.
Future Implications
From Public to Defense Sector
Michael Hull of BAE Systems observes that innovations like AI often transfer from the public domain into defense, enhancing development speed. The AI dogfighter project, for example, benefited from publicly available air combat tactics.
Conclusion
The future of aerial dogfights will likely involve AI co-pilots that assist human pilots, reduce information overload, and enhance combat efficiency. While AI has made significant strides, it still has a long way to go before it can fully replace human pilots. However, its rapid evolution and integration into military training and operations suggest a transformative impact on modern aerial comb
