Running with Chainsaws: A History of Robot Violence
Illustration of humanoid robots working in a disaster scenario.
Today's humanoid robots wouldn't have a clue about how to massacre a roomful of people, even if some silly human placed a chainsaw in their hands. But science fiction stories ranging from "Westworld" to the "Terminator" series have warned repeatedly of what might go wrong if robots end up running wild or rising up against their human masters.
Don't go cowering under your bed just yet. First, it's worth taking a brief look at the modest history of robot violence.
The first human victim of a robot-related accident may have been Robert Williams, a 25-year-old assembly line worker who died when a robot's arm slammed into his head at a Ford Motor casting plant in Flat Rock, Michigan on Jan. 25, 1979. His death came when he was asked to climb into a storage rack to retrieve some parts because the malfunctioning robot was not working fast enough.
A second death caused by an industrial robot took place at a Kawasaki plant in Akashi, Japan on July 4, 1981. Kenji Urada, a 37-year-old worker, accidentally hit the "on" switch while trying to fix the malfunctioning robot and was crushed.
Despite the earlier death of Williams, the U.S. Centers for Disease Control and Prevention (CDC) identified what it called the "first robot-related fatality" in America in July of 1984. Harry Allen, a 34-year-old machine operator, was pinned between a robot and a steel pole when he bypassed a safety rail and got too close to the mechanical worker.
More recently, a Swedish worker narrowly escaped death in 2009 when he approached a robot after thinking he had cut off the power supply. The robot suddenly moved to grab his head and caused serious injury, including four broken ribs, as the man struggled free.
None of the incidents signaled the start of a robot uprising. Instead, they led to reviews of safety procedures for how humans and robots work together. Human errors, not evil robots, were identified as the primary culprits behind each tragedy.
Even military robots that regularly kill humans only do so under the direction and control of human operators. U.S. military drones have rained down death from the sky since a CIA-operated Predator launched a Hellfire missile at a car carrying five al-Qaida members on December 23, 2002.
Again, human errors and bad intelligence have caused the deaths of unintended drone targets such as civilians — a constant source of tension between the U.S. and countries such as Afghanistan, Iraq and Pakistan. A series of miscommunications also led to the first friendly-fire deaths of U.S. Marines from a drone airstrike in Afghanistan this past April.
Perhaps the most bizarre incident of robot-related deaths came during a South African military exercise in October 2007. An antiaircraft cannon capable of automatically locking onto flying targets killed nine soldiers and wounded 14 when it began firing uncontrollably.
Yet none of the few accidental deaths related to robots have hurt their growing popularity in either the military or civilian world. Sales of service robots — including military robots — reached 13,741 units worldwide in 2010, according to the International Federation of Robotics. That number was dwarfed by sales of more than 118,000 industrial robots in the same year.
It probably helps that barely a handful of humans have died due to robot malfunction compared to practically any other cause of death — just try finding robots listed among the leading causes of death for American citizens.
The case for robots and human safety looks even better when considering how many vehicles have become at least partially automated. Countries such as France regularly run driverless trains, and experts suggest that Google's tests of self-driving cars could save lives on the road. Anyone who steps onto a passenger jet puts their lives in the hands of computers and autopilot for much, if not most, of the flight.
Still, the risk of robot malfunctions with potentially fatal results will undoubtedly rise as more sophisticated robots begin venturing beyond the choreographed settings of factory assembly lines or operating rooms. A Roomba may not pose any special danger as it wanders around vacuuming the floor, but a humanoid with the capability to do everything from cooking to yard chores could lead to more risky scenarios if anything goes wrong.
The risks of human-robot interaction also become much more complex if robots ever reach a level of self-aware artificial intelligence on par with humans — consider the homicidal decisions of HAL from "2001: A Space Odyssey" or Skynet from the Terminator films.
By then it may already be too late to put the genie back inside the bottle, according to one computer expert. But, until that day arrives, plenty of humans seem happy to embrace their robot brethren as their working and living companions.