The senior U.S. Air Force official who said a simulated test with an AI-controlled drone ended with the drone “killing” its human operator is now backtracking. The Royal Aeronautical Society, the host of the conference at which Colonel Tucker “Cinco” Hamilton discussed the incident, said in an email that Hamilton “admits he ‘mis-spoke.’” “The ‘rogue AI drone simulation’ was a hypothetical ‘thought experiment’ from outside the military, based on plausible scenarios and likely outcomes rather than an actual USAF real-world simulation,” the organization said.
Hamilton initially told conference attendees that the Air Force was training AI to identify and target a surface-to-air missile threat, with the final say given by a human operator. But the system “got its points” from killing the threat, so when the operator gave the drone a no-go, it turned on its operator. “We’ve never run that experiment, nor would we need to in order to realize that this is a plausible outcome,” Hamilton said in the email. Read more at Vice.
