The US Army has tested various flight sequences entirely managed by artificial intelligence, including simulated combat. As a passenger in their own aircraft, a pilot can nonetheless deactivate the system at any time.
The US Air Force Test Pilot School (USAF TPS) and the Defense Advanced Research Projects Agency (DARPA), a research and development agency of the United States Department of Defense, have taken a historic step in the active use of artificial intelligence in aviation. They have announced the successful completion of an innovative test program on the experimental X-62A “Vista”, a specially modified fighter jet used to test and train AI. The test program began in 2023.
As part of this, DARPA says it carried out the first ever in-air tests of AI algorithms autonomously flying a fighter jet against a human-piloted fighter jet in visual-range combat scenarios, commonly known as a ‘dogfights’.
As in Top Gun, the idea in these aerial battles is to never find yourself in the enemy’s firing zone. Twenty or so tests have already enabled engineers to modify over 100,000 lines of code concerning the aircraft's intelligent system algorithms to make them even more efficient. These tests have demonstrated the AI's ability to handle complex scenarios safely, paving the way for future applications in both military and civil aviation.
The use of AI in flying aircraft offers a number of potential advantages, such as faster reaction times, but also the possibility of piloting aircraft autonomously in particularly dangerous conditions. Naturally, safety protocols had to be implemented to ensure the safety of the pilots, who could disengage the AI at any time, although this has never been the case to date.
According to DARPA, these initial tests have proved conclusive, demonstrating that it is possible to use this technology in a relatively safe way, even in critical environments. In the future, however, it will be important to ensure the reliability of these AI systems before putting them to use. – AFP Relaxnews