Pentagon is Attempting to Incorporate AI-controlled Robots into the US Military

The Defense Advanced Research Projects Agency (DARPA) of the Pentagon has been testing completely autonomous artificial intelligence to handle drones bearing military weapons and licenses to combat enemy attacks. The US military requires human monitoring and supervision on all drone operations that use lethal force, so this is a significant deviation from normal procedure.

Last August, a drill involving numerous dozen AI-controlled drones and tank-like robots was held in Seattle. The AI-controlled drone swarm was expected to follow specific orders from a human operator, such as locating and destroying terrorist suspects, but they were free to choose their own form of action and strike without the need for human interference.

Rather than using actual weapons, the drones were equipped with radio identifiers that simulated weapons while detecting strikes and failures using telemetry and signal data. The drill purposefully used a huge number of robots to prevent human supervision. Simple goals (such as search and destroy enemies) were provided to the drones, with AI and algorithms deciding the form of action.

The objective of this exercise was to investigate the use of artificial intelligence (AI) in battle situations that are too complex and unpredictable for humans to manage. The Pentagon wants to see if it's possible to limit humans to high-level choices while AI handles complicated, critical decisions about how to accomplish certain tasks in an effective way.

Experimentation by the Pentagon to automate this element of conventional combat could pay off handsomely by exploiting existing developments in electronic warfare. This would make combat much more effective and less dependent on human decision-making.