Artist’s conception of a lethal autonomous drone, attacking enemy tanks after visual recognition and AI confirmation.

1942

LETHAL MILITARY ROBOTS

Robots used in warfare include many examples in the 1900s, such as the tank-like Goliath used by the Germans beginning in 1942, during World War II, on all fronts where the Wehrmacht fought. The unmanned Goliaths were remote-controlled via a connecting cable and carried high explosives so that the machines could blow up along with their intended targets.

Today, drones (unmanned flying vehicles) can be armed with missiles to serve as an effective weapons system, but they usually require remote human input and authorization before being “allowed” to destroy a target. Of historical note, an MQ-1 Predator drone launched the first-ever deadly airstrikes from a drone in 2001 in Afghanistan. Debates continue regarding the possible future use of lethal autonomous weapons that could actually select and attack military targets without human intervention. Automatic defensive systems do exist today, including machines that can autonomously identify and attack incoming missiles.

Possible advantages of military robots are many: they never tire or show fear; they can rapidly perform maneuvers that would injure human pilots; and they could potentially save lives of soldiers and reduce collateral damage and civilian casualties. In principle, machines might be instructed to follow various rules, such as not to fire when unsure if a target is a civilian or a combatant, or whether lethal force is permitted. Perhaps damage potential to civilians might be programmed and restricted to be proportional to the size of a military target. Facial-recognition software could be used to enhance accuracy, and military robots could work side by side with soldiers, enhancing their abilities much like software or robotics may be used today to assist with surgeries. But how much independence should such fighting machines be given? Who is at fault if a robot accidentally attacks a school?

In 2015, a large group of AI experts signed a letter warning of the dangers of military use of offensive autonomous weapons beyond human control, which could lead to a global AI arms race. The letter was presented at the International Joint Conference on Artificial Intelligence and signed by such intellectuals as Stephen Hawking, Elon Musk, Steve Wozniak, and Noam Chomsky.