Paul Scharre, a former Pentagon official who played a role in establishing the country’s policy on autonomous weapons, has written a new report which argues that these weapons are more dangerous in real-world environments than we think because they’re susceptible to design failure, hacking, spoofing and manipulation by the enemy. Scharre now directs a program on the future of warfare at the Center for a New American Security, which is a policy research group based in Washington, D.C.
There has been a significant amount of advancement in military technology over the past few years which has opened up the field for powerful killer robots to do what humans have to do now. The advanced technology has contributed to weapons systems that can make killing decisions without any human intervention.
Scharre is warning about the real-world risks associated with these weapons in his new report. He points out that autonomous weapons systems don’t have the flexibility that humans do which allows them to adapt to circumstances and take appropriate decisions. Autonomous killing machines can thus make mistakes that humans might possibly avoid and in a conflict that can prove to be very costly.
The United States doesn’t have an advanced fully autonomous weapons system yet, or, at least, it’s not public, however the Defense Department has asked for $1 billion to produce Lockheed Martin’s Long Range Anti-Ship Missile which many describe as a semi-autonomous weapon because the missile can fly out of contact with the controller and automatically identify the target it’s supposed to attack.
Even though in that particular missile system a human operator has to initially select a target, Scharre believes that having a human just to push buttons won’t be enough and that humans should actively be engaged throughout the entire process.