Violators of the First Law: How Robots Kill Their Masters
In 1942, the science fiction writer Isaac Asimov penned the short stories "Liar!" and "Runaround" where he formulated the Three Laws of Robotics, the first of which states: "A robot may not injure a human being or, through inaction, allow a human being to come to harm." Unfortunately, in the world of real robots, much like in the world of real people, these laws don't apply to everyone.
South Korea under Attack
This November, in the South Korean province of Gyeongsangnam-do, at a pepper sorting plant of the Donggoseong agricultural complex, a robot killed a human. The victim was a 40-year-old worker who climbed onto the conveyor to check the functionality of the sensors of an automated arm. For the past five years, this "arm" had worked without any complaints, lifting heavy boxes of peppers and placing them on pallets. However, recently, the robot started malfunctioning. Eventually, on that fateful day, it mistakenly identified the worker as a box, grabbed him, and forcefully pressed him against the conveyor belt, crushing the man's face and chest. The injured man was taken to the hospital, where he later succumbed.
This event sparked a strong reaction among South Koreans, who have long been living in a true future where robots are replacing people not only in strenuous production but also in household chores and everyday life. Prior to the mentioned incident in South Korea, there was another – less tragic but equally horrifying. In 2015, a robot vacuum attacked a woman lying on the floor (note the now dangerous habit) and sucked in her hair. One might think, so what? The catch was that the woman couldn't fend off the attacker on her own. Miraculously, she managed to reach the phone to call for rescue. Only when they arrived and freed her did she catch her breath, realizing that tomorrow there wouldn't be her obituary in the newspapers with the words "died heroically fighting a robot vacuum."
The First Victim of Robots
Robots started attacking humans not today or yesterday but actually a very long time ago. Robert Nicholas Williams, a man who entered history as the first person killed by a robot, experienced this a quarter of a century ago, in January 1979. Williams, then only 26, worked as an operator at the Michigan Casting Center of Ford Motor Company, where Litton Industries built a five-story system for delivering parts. In particular, there were robotic carts that took out molded parts and moved them to shelves. Once the system had a breakdown in inventory reading, Williams climbed onto the racks to manually retrieve the necessary parts. The situation looked so ordinary that his colleagues began to suspect trouble only after half an hour. It turned out that when Williams climbed to the third level, one of the carts, weighing a ton, suddenly accelerated without reason and hit the man in the back. Williams died instantly.
When people read about the first human killed by a robot, many wondered why it happened in the USA and not in Japan, where robots are, as known, some kind of national sport. However, two and a half years later, in the summer of 1981, machines also went after the Japanese. Kenji Urada, 37, worked as a repairman at the Kawasaki Heavy Industries plant in Akashi. When he noticed a minor malfunction in one of the robots, he decided not to stop production, cutting off the entire line. Instead, leaping over the protective barrier, he started troubleshooting right in the working area. Accounts of what happened next vary, but most likely Urada got caught up in his work and didn't notice a transport robot approaching rapidly from behind, carrying parts for the robot he was repairing. Pinned between the two machines, Urada died. Labor inspection concluded that due to the active use of robots, which sometimes operate entire workshops automatically, human operators tend to lose vigilance and even treat their work carelessly.
Don't Sleep on Tesla
It's crucial to keep a watchful eye on even the most impressive and reliable robots. Joshua Brown from Florida gained fame as an enthusiast of Tesla's innovative technologies. In 2016, a viral video from his dashcam showcased how autopilot saved his car from colliding with a truck. That was in April; by May of the same year, Brown let his guard down, allowing the automated vehicle to end his life. Neither the driver nor his Tesla noticed as a tractor-trailer veered across the road. The car went under the trailer, transformed into a convertible, flew off the road, and crashed into a pole. Later investigations revealed that during the last 37 minutes of his life, Brown had his hands on the wheel for only 25 seconds (periodically touching it to prevent the autopilot, which was in a test mode, from stopping the car). Despite continuous warnings from Tesla to return hands to the wheel, the driver ignored them, suspected of watching a Harry Potter movie on a laptop.
No Gender Preferences
If you get the impression that robots only kill men, that's not the case. In 2009, Ana Maria Vital became the first female victim of robotics. In the Californian city of Irvine, the 40-year-old warehouse worker at Golden State Foods (a supplier of semi-finished products to McDonald's) noticed that the "arm" had dropped one of the boxes. The woman decided to assist the clumsy robot and entered the cage where it operated. However, before Vital could pick up the box, the "arm" seized her and crushed her. Probably, as in the case of the South Korean worker, the robot mistook the woman for a box.
One of the strangest cases in history involved another woman, 57-year-old Wanda Holbrook. For 12 years, she worked at the Ventra Ionia Mains automotive plant in Michigan. Many robots work alongside people there, strictly forbidden to leave the workshops they are assigned to. Despite this rule, in 2015, one robot started moving strangely—first, it independently overcame protective doors between workshops, then entered the workshop where Holbrook worked and smashed her skull by unloading a part of an automobile trailer fastening onto her head. The investigation commission could not determine the reasons for the tragedy. But the fact that the robot executed a series of instructions (which no one programmed it for) and even independently figured out how to open the complex lock mechanism of the safety doors between workshops—agree, it's frightening.
Some may argue that the mentioned incidents are isolated, with victims being individuals. So, let's consider a case of mass murder: In 2007, at the Lothla test range in South Africa, soldiers were training near the automated Oerlikon GDF-005 anti-aircraft gun, which the Swiss developer had supplied to more than 30 countries since the early 1980s. Suddenly, the combat robot began chaotic firing in all directions—nine people were killed on the spot, and 14 were severely injured. Considering that the anti-aircraft system fired 250 shots with 35-millimeter shrapnel-fragmentation shells, the number of victims could have been much, much higher. Military personnel later admitted that the Oerlikon GDF-005 had been acting up before—the only difference was that there were no casualties, so they never raised the issue.
So, maybe it's time to raise it now? Imagine a world in the near future where robots independently cradle our babies. Where, while we sleep soundly, they silently clean our homes and survey their surroundings. Where they prepare our food with sharp Japanese knives. Where they see us with their camera-eyes, hear us with their microphone-ears, and constantly analyze all our actions with their processor-hearts. Can we say with certainty that, in doing so, they will also genuinely love us with their artificial intelligences?