We live in a world in which robots are becoming more and more a part of our lives, and automation is taking over more and more control of our society from humans. The rise of robots on our planet and their capabilities have picked up incredible speed, which of course has led to discussion on just what the odds are that they may one day decide that they have had enough of being our servants and rise up to defy us in a robo-apocalypse. Although movies have covered this premise extensively it seems the reality of it has yet to appear. However, it does not mean we are exactly safe, as in a sense robots have already struck against their masters on occasion, showing us that perhaps the robot uprising has already somewhat begun.
If we are ever subjugated by robot overlords, the dawn of robot attacks on humans will be marked as the date of January 25, 1979, which was when the first ever recorded lethal robot attack on a human occurred. On this day, Ford Motor Company worker Robert Williams was working at one of the car giant’s casting factories in Flat Rock, Michigan, in the United States, which utilized massive, 1-ton robot arms used for the parts retrieval system, which lifted and moved various parts from one end of the factory to the other. At one point, one of the units inexplicably slowed down, and instead of trying to figure out what was going on, Williams decided to go and just move one of the parts himself. As soon as he entered the retrieval zone, the robotic arm suddenly sparked to life again and swung around to strike him in the head with brutal force, killing him instantly.
Japan, which is perhaps known as a mecca for robotics and has the largest number of robot workers in the world, had its first official human death by robot in July of 1981, at a plant of Kawasaki heavy industries. On this day, engineer Kenji Urada approached one of the industrial robots in order to repair it. The robot seemed to be deactivated, and would have automatically been completely depowered if Urada had passed through the fence surrounding it, but instead he for some reason chose to jump over this fence to approach the seemingly docile machine. Even then the robot remained inactive, but it would prove to be merely a coiled beast ready to strike. As Urada began his repairs, the machine jumped to life and ferociously shoved the engineer into a grinder used to cut gears, killing him in a rather gruesome fashion. It is thought that he may have brushed up against the robot’s manual controls to turn it on and accidentally trigger its action. The accident caused a good amount of concern in a nation that has a far larger robotic workforce than any other country, and safety regulations were strictly reviewed by the Ministry of Labor.
It may seem that with more and more use of robots in the workplace that these sorts of mishaps would be unavoidable, but safety regulations and increasingly safe technology have helped to somewhat cut down on the number of such horrific accidents. Nevertheless, despite this there have been other disturbing deaths dealt out by robots in more recent times. In 2000, one construction worker was working on installing catwalks at Greeley, Colorado’s Blueshift Industrial for the purpose of providing safer access to the factory’s Alvey machines; robots designed to distribute packaged goods through a system of robotic arms, conveyor belts, and an extruder car, in this case situated within a large, cooled warehouse for packing meat. Unfortunately, it didn’t prove to be safer for the worker himself.
In this case, the robot arm was used to follow a preprogrammed order and reach out and grab packaged meat from one of the massive storage carousels, lifting it up and lowering it down onto a conveyor belt, after which the meat would be whisked away to a waiting truck. The worker at some point was struck by the hydraulic arm of the system and knocked 10 feet down onto the conveyor belt, after which the extruder car came barreling along on its course and crushed his head. The car should have stopped when a human being stepped onto the conveyor belt, but at no point did the machine’s safety shut-down features kick into effect, pointing to a possible malfunction. Or maybe the robot was just sick of moving and packing meat?
In 2006 there was a rather chilling deadly robot attack at the U.S. based firm Alliance Tech Systems. One employee entered one of the company’s many robotic workstations, in which a powerful robot arm with the somewhat ominous-sounding name of “The Degator” turned on the man, picking him up in an actual chokehold and effortlessly swinging him around to crush him against its cold metal frame. The worker’s body was found later still pinned up against the frame in the robot’s grip, and it was reported that the victim had died apparently struggling against the immensely powerful arm while reaching desperately for the reset button. This already seems like an almost planned attack on the employee, and making it all the more sinister was that when authorities checked the robot’s memory to find out what happened it was found that the event had been mysteriously deleted, despite the fact that this data should have been evident. Even creepier still, a full analysis of the machine showed nothing out of place, no abnormal movement at all, and no malfunction of any kind.
Not quite as deadly, but just as frightening, was another robot attack the following year, this time at a factory in Bålsta, north of Stockholm, Sweden. In June of 2007, a worker went to do maintenance work on a bulky robot designed to lift and move heavy rocks, cutting the power before approaching the hulking machine. Of course if you have been paying attention so far you know where this is going. Despite the fact that the man insisted later that he had shut down the machine, it nevertheless suddenly reached out to grab him in its vice-like grip, tossing him around like a rag doll and breaking four of his ribs in the process. Luckily, the terrified worker was able to escape, after which the robot went into a standby mode. A public prosecutor investigating the case would later say of the strange incident, “I’ve never heard of a robot attacking somebody like this.”
Just a month before, in that very same year, there had also been an attack carried out by a robotic arm at a CD case factory in the United States. The machine’s purpose was to remove CD cases from an injection-molding machine. During this process, the arm reportedly spun around off of its intended trajectory and violently struck a nearby worker who was troubleshooting the machine, sending him flying through the air. The man was then hospitalized with a fractured skull and broken ribs, and would die of his injuries two weeks later.
Even more recently, the year of 2015 produced two strange and shocking robot attacks on human beings. In July of 2015, a 22-year-old worker at a Volkswagen plant in the Kassel district of Germany was going about his duties as usual when a robotic arm malfunctioned and swung around to grab him and smash him against a metal plate with bone crushing force. The attack caved the man’s chest in, and he later died at a hospital. Oddly, the robot had never been programmed to move at that angle or in that direction at all, and the worker would have thought himself quite safe where he was at the time, making the cause of the attack rather mysterious. Also in 2015 was a case at SKH Metals, in India, where a 24-year-old worker was working with a robot used to weld sheets of metal. The worker’s job was to adjust the sheets as the machine went about its work, which was all preprogrammed, but the welding arm unexpectedly lashed out to stab him in the stomach, by some accounts also electrocuting him, and ultimately killing him. Again, the robot had no business deviating from its set course.
In the face of an increasingly automated, robotic workforce there is perhaps no way to really totally avoid malfunctions and accidents with people and complex machines in such close proximity, but not all robot attacks involve industrial machines, nor are these always the most deadly. One of the scarier reports in recent years occurred in October of 2007, in South Africa, when a prototype autonomous robotic 35mm anti-aircraft cannon was being tested in the field. The robot, which was using live ammunition, reportedly suddenly turned on the South African National Defense force personnel standing about and began firing wildly upon them, unleashing its full payload of 250 rounds of high explosive rounds weighing half a kilogram each onto the horrified onlookers, continuing to try and fire even when its magazines were empty. When the smoke cleared, 9 soldiers were dead and another 11 injured, making it perhaps the deadliest robot attack in history.
This accident in particular underscores the dangers of increasingly militarized robots and the goal of defense contractors to develop completely autonomous robotic weapons of war. These advances in AI would allow robots on the battlefield to make decisions or select and fire upon targets without any human input or interference whatsoever, and considering the lethal accidents with preprogrammed robots we have already looked at it all opens up a whole can of worms with regards to safety, ethics, legality, and accountability. After all, who can stop a robotic drone or other weapon from destroying what it wants, and how does it choose? What if it disagreed with its orders and who would be held liable if it killed unintended targets?
These are the sorts of issues that have fueled seemingly endless discussion and debate on the imminent development of fully autonomous robots, with the United Nations itself talking over the possibility of potentially completely preemptively banning the use of such robotic weapons in wartime. This sort of thing has been done before, most notably with the U.N. ban on lasers used to blind the enemy, which were nixed before they could really be fully developed. Such a ban would shut down the many projects currently on the table that seek to create fully autonomous killing machines. One Steve Goose, arms division director at Human Rights Watch, has said of the dire need for this thus:
In policing, as well as war, human judgment is critically important to any decision to use a lethal weapon. Governments need to say no to fully autonomous weapons for any purpose and to pre-emptively ban them now, before it is too late.
Not everyone agrees, and some have suggested that using such robots would actually reduce the number of casualties, as they would be more precise, would be better at determining whether to engage an enemy or not, and would not act on rage or fear to kill indiscriminately. The heated debate continues, but whether autonomous robots are banned or not, there are already some scarily independent robots already operating out in the field. For examples, although not completely out of the human loop, the U.S. Navy’s Phalanx gun system is capable of automatically tracking and engaging perceived threats, and the U.S. military also has highly advanced drones that also possess some degree of autonomy when operating or targeting enemies. Israel also has shown such advances in its drone, the Harpy, which can seek out and attacks radar stations on its own.
Although the cases we have looked at here are most likely down to human error and malfunctions, the threat of renegade AI is perhaps not completely far-fetched, as computing power grows exponentially and more advanced algorithms and ever sophisticated learning computers are experimented with. This has caused a considerable amount of serious consideration in recent days into the dark possibility that our robots and machines will at some point reach a point where they can defy us, perhaps violently. A very vocal recent detractor of continued AI development is none other than Tesla Motors and Space X CEO Elon Musk, who has in recent years launched a veritable crusade against the unfettered growth of artificial intelligence, calling it one of the greatest threats that mankind has ever faced, and rather starkly and dramatically saying:
With artificial intelligence, we are summoning the demon. You know all those stories where there’s the guy with the pentagram and the holy water and he’s like, yeah, he’s sure he can control the demon? Doesn’t work out.
Time will tell if such a dire prediction will ever come true, or if robots will ever reach a level of development in which they can truly act independently of humans and actually turn against us, but it remains a dark specter lurking about. Does a bleak future lie over the horizon in which sentient machines hunt us down or seek to destroy us? Will our technology undergo a metamorphosis from a tool to the agent of our downfall? It is hard to say, but in the meantime we have these weird cases of killer robots that cannot help but make us think about such things.