In the second of an occasional series examining the far-reaching effects of computers on today’s society, MEL CROUCHER discovers that there really are such things as:


COMPUTERS CAN KILL! We all know about the rogue mainframes that have tried to start World War III (don’t we?), or the computers that drove a pensioner to suicide by charging £2172 for two phone calls (don’t we?), or the fantasy games that drive kids to murder (yes we do; see Monitor in Issue 54). But this month’s Monitor points the finger at robots and computers that haven’t just caused human deaths, they have actually committed MURDER!

In 1920, a Czech writer named Karel Capek coined the word ‘robot’ to describe a machine that looked and behaved like a human being. Six years later, in Fritz Lang’s silent movie Metropolis, the world’s first sexy robot was shown leading men to their death. It was not until 1950 that the sci-fi author Isaac Asimov set down a series of rules which he called the Laws of Robotics, a sort of mechanical Ten Commandments, and ‘thou shalt not kill’ came out way on top. In 1964 a sixty minute sci-fi film called I Robot was shown on television as part of The Outer Limits cult series. It concerned the trial of a tin man, accused of murdering its creator. But we had to wait a little longer for the real thing.

Our life is made by the death of those different from us!

Leonardo Da Vinci 1502

I gave you life! I created you! And for thanks you desire to murder me!

Mary Shelley Frankenstein 1818

I’d sing, I’d dance, I’d play my part. If only I had a heart!

The Tin Man The Wizard of Oz 1939

In February 1982, a maintenance worker at the Kawasaki plant in Akieski, Japan, got himself written into the history books. Kenji Urada has the dubious honour of being the first human being to be murdered by a robot. Instead of opening the robot’s safety gate — which was supposed to cut off its power — Kenji jumped over the barrier fence and accidentally hit the juice button. The robot took a look at him, decided that he was an industrial component, grabbed ahold of the poor man and turned him into sausage meat with a gear-cutting machine. Nasty!

In a survey of American factories where robots are hard at work, no less than four percent have had major robotic accidents, including heads bashed in by ‘intelligent’ tool arms, and two unfortunate guys hung in the air by their feet and sent along the conveyor belt to be turned into cars. Most deaths and injuries are caused by the fast movement of robot arms, trapping and crushing humans, or knocking them senseless into heavy machinery.

International safety authorities now recommend stringent procedures and precautions in the battle against robot murder. In the industrial world, costs are always cut and very few factories have actually installed these precautions. But it seems that the robots are getting smarter than we think. In April 1986, a car assembly worker was made redundant by a second-generation industrial robot. He decided to take his revenge by smashing the machine up with a lump hammer. As he raised the hammer to crush the control box, the robot spun around, changed direction and caught the unfortunate human by his goolies! As he doubled up, the robot arm flipped him back over the safety fence and switched itself off!


The great American power blackout of 1965, last year’s Chernobyl disaster and the explosion of space shuttle Challenger in 1986 were all caused by human error, when scientists ignored or misunderstood the computer data that was supposed to protect life. In the near-meltdown at the Three Mile Island nuclear power plant, the software miscalculated the design of the reactor cooling system. It is estimated that over two hundred people will die of cancer as a result of that radiation leak. What is even more frightening is when software deliberately causes tragedy.

Life is a great surprise, I don’t see why murder should he a greater one

Vladimir Nabokov 1962

Be peaceful, be courteous, obey the law. But if someone fingers you, send him to the cemetery

Malcolm X 1965

In 1983, a terrible flood along the Colorado River caused millions of dollars worth of damage. The Governor of Nevada admitted that the flood was caused by a ‘Monumental mistake made by the federal computers’. What really happened was much more sinister. The computers decided to ignore the rate of snow melt-off that Spring, because they preferred to deal in averages based on their own experience, and so a massive amount of water was kept dammed up, causing the computerised flood. Six people were murdered by this ‘logical’ computer decision.

In May 1987, a Canadian Therac-25 computer was treating cancer patients by bombarding their tumours with radioactive therapy waves. The computer decided to increase the dosage by ONE HUNDRED TIMES. Two patients were murdered and several more are dying right now as a result of receiving 25 times the fatal dose of radiation. Multi-million dollar claims against the software manufacturers are still unresolved, because they insist that their programs worked perfectly, and the computer went insane. The claimants insist that it was the software which went crazy.

Two years ago, a Korean Airlines passenger jet liner, Flight 007, was shot down by the USSR over two hours after it wandered into Soviet air space, just as it was leaving for international waters. Everyone onboard was murdered. The onboard software had become confused and got the jet lost in the first place, but it was the Russian software which insisted that Flight 007 was a hostile military target, ignoring radio and visual contact to the contrary. The Russian claim that a passenger liner was on a spying mission was hogwash, put out to cover up the fact that their software was not only hopelessly slow in response, but also murderously stupid.

More recently we can look to the Gulf for fearsome facts. Many outside of Iran accepted the shooting down of the commercial airplane by the USS Vincennes as a mistake, the real truth is frightening. Just as it was supposed to do, the Aegis computer system onboard the American warship identified the Iranian passenger plane as it left the airport at Bandar Abbas. Trouble is though, the plane was recognised as a 62-foot-long F-14 Tomcat fighter, when it was in fact a 177-foot-Iong Iran Air Airbus. Within seven minutes of take-off two electronically-guided missiles were unleashed, and within seconds, out of sight of the American warship, it was blown to pieces, killing all of the 290 people aboard. The Airbus didn’t identify itself, and the Aegis control system is programmed to treat anything it cannot identify as hostile and blow it out of the air — unless it is manually over-ridden!

When NATO computers decide to shoot down civilian aircraft or flocks of migrating birds, it is mainly because our five biggest early warning computers are Honeywell 6080s, designed in the 1960s for batch processing. After the Korean Airlines disaster, we took a long hard look at our own system, and decided to replace it as soon as possible with IBM-3083 machines.


The American spacecraft Gemini V splashed down 100 miles from its landing point because its computer software ignored the motion of the Earth around the Sun. In 1979, five nuclear power stations had to be shut down when an earthquake prediction program threw a panic. Instead of analysing the values of a set of numbers, it decided to add together their arithmetic sum. Probably the most ludicrous military computer cock-up ever is the case of the US Air Force’s F-16 bomber. Whenever these sophisticated fighting machines crossed the equator, the on board software got a wee bit confused and instructed the planes to fly upside down!

If the human race wants to get to Hell in a basket, technology can get it there by jet!

Charles Allen 1967

Beware of engineers. They begin with sewing machines and end up with the atomic bomb!

Lewis Mumford 1951

On October 5th 1960, World War III was almost triggered by computers for the first, but by no means the last time. The Ballistic Missile Early Warning System based in Greenland informed the White House that the United States was under a massive attack by Soviet missiles. The West went to Red Alert and NATO got ready to launch an all out retaliation, as the Americans waited for their computers to confirm the attack. The computers calmly flashed the message that Soviet missiles were indeed attacking with a certainty of 99.9%! Back in those days, the world had several minutes to decide whether or not to commit suicide. These days, computers will make that decision in a matter of seconds. The Russian attack? Well folks, what the computers had spotted was the RISING MOON! Nobody had bothered to teach them about such a common occurrence! The expression ‘loony’ has never been more appropriate.

In 1979 on November 9, what happened in the sci-fi film WarGames happened for real. A very serious ‘computer game’ was fed into the North Atlantic defence computers. This test data is frequently used to check out the missile warning system by playing war games against the computer. Trouble was that some bimbo connected the software to the real alert system and for six minutes we went to war! There were ten launches from the Northern USA and Canada before the error was spotted. Frightening? Not as frightening as what happened the following year when the software took over and went bananas.

In 1980 on Tuesday June 3, at 1:26am, the Nebraska command post computer reported that two missiles had been launched from Russian submarines. Just 18 seconds later, it spotted several more. By 1:30 it was cheerfully plotting Soviet Intercontinental missiles heading towards America. It was then that the Pentagon computers confirmed the attack! We got ready for The Big One. B-52 atomic bombers started their engines, the covers came off our land-based missiles. The ever-ready airborne command post took off from Hawaii and took control of US warships at sea. The generals and admirals were a bit worried that the attack didn’t seem to follow any logical pattern, but they got ready to retaliate just the same. Three minutes and twelve seconds into the so-called attack the computer monitors started flashing garbage and the war was cancelled.

Three days later, June 6 at 3:38pm, exactly the same thing happened. Once again the bombers started their engines and the covers came off the red buttons. Once again the screens went insane after a few minutes. The cause of this warmongering? A simple failure of a 74175 chip in a Data General communications computer and a smartypants bit of free enterprise on the part of machine. All defence communications are constantly checked out by filler messages, with a zero for the number of missiles that are attacking. When the chip failed, the computer decided to ignore all of its error correcting programs and fill in the amount of missiles detected with random digits!

But, I hear you say, insane military computers have NOT started World War Ill. And I reply, ‘not yet’. They already have committed wholesale murder, though. During the American war in Vietnam, the Pentagon used special computers in the jungle battlefields which reported real-time live action, so that the USA could make tactical decisions. Computerised information on combat sorties, fuel supplies, weaponry and enemy movements was sent back by the military to the politicians, and the war raged on killing tens of thousands of civilians. The trouble was the software was lying. It was reporting on ‘illegal’ but real targets in Cambodia (a country where the US forces were not supposed to be) and automatically converting the data to fake locations in North Vietnam.

You may be wondering why I haven’t told you about similar incidents in Britain. This is because my information is freely available from the US Senate Committee on Armed Services. They have laws over there where the public are allowed to know about all the computer cock-ups made by the military. If I told you about the colour of wallpaper in the Ministry Of Defence office I used to work in, I could be thrown in jail under our loony Official Secrets Act.

So what happens if someone gets killed or injured by a computer foul up? Well, under the Consumer Protection Act which came into force last March, the victim’s family stands a much better chance of claiming compensation from the software company responsible. Ronald Robertson of the company law specialists Stephen Harwood says ‘victims won’t have to establish fault, they only have to prove a causal link between the injury suffered and a defect in the product’. But certain folk don’t bother with the law. In the USA, one woman walked into the nerve centre of a Trident missile targeting system and smashed it to bits with a hammer. We must not applaud acts of violence, but some people seem to think that we should kill computers before they kill us.

In the next Monitor, Mel Croucher will be exposing the facts about less lethal computer crimes, but much more amusing ones!