12 news stories that prove the robot uprising has begun
Arnold Schwarzenegger in Terminator Genisys is not the only one worried about the robot uprising. We are too.
Last December Stephen Hawking expressed his concern that robots catching up with our intelligence would quickly lead to our extinction.
Since the original Terminator came out in 1984 a lot of things have happened. The first film predicted that the machines first became self-aware on August 29th, 1997. But since then robots have advanced. Google cars drive themselves, unmanned drones are constantly flying around the world and computers are learning by themselves. It's all getting a little bit terrifyingly close to the events of Terminator Genisys.
Here are twelve signs the robot uprising may have already begun.
12) The NSA actually have a program named SKYNET
In Terminator SKYNET is a "global digital defense network" that becomes self aware and starts the war against humanity. Everyone who's ever watched a Terminator film, or grew up any time between the 1970s and now knows this.
Which is why we should be alarmed that there's a real-life organisation calling themselves SKYNET. If you're creating a global surveillance system and you decide to call it SKYNET, you're probably a baddie. It's going out of your way to make yourself sound like a movie supervillain.
The real-life Skynet was top secret until it was leaked by Edward Snowden a year or so ago. the NSA's Skynet uses metadata to identify people with terrorist connections. In one case the program tracked the movements of people within Pakistan from their mobile phone records and compared them to the movements of suspected Al Qaeda couriers. Not bad for an algorithm tracking down people over the internet.
11) They're already being designed to disobey Asimov's laws of robotics
In most sci-fi films the robots have been programmed so that they don't harm humans. The Guardian, played Arnold Schwarzenegger is the perfect example of this, with his sole mission to protect Sarah Connor.
So far in real-life a lot of robots haven't been programmed not to harm humans. In fact, some of them are being designed to taze them in the balls.
Meet the TRAD robot, the robot created by Taser International so that police don't even have to be in the same room as the suspect as they shoot 50,000 volts through them.
Conveniently though, they do have a camera on them so they can see the look of pure terror on the suspects' faces as they: wonder what the hell that remote control car is doing, notice the robot seems to have both a taser and a thirst for human suffering, try to appease their metal overlord, and get shot with 50,000 volts of electricity. Fingers crossed they upload the video to YouTube.
At the moment, the robot is remote controlled and armed with a taser, but there are plans to automate it and arm it with a shotgun for military purposes.
It's a sign the robot uprising when governments aren't just dreaming of fully-automated robot killing machines: They're budgeting for them.
10) They can transform themselves
One of the cooler sequences of Terminator Genisys is seeing the upgraded version of the Terminator (dubbed the T-1000) melt and adapt to his environment. When he needs a sword for an arm he grows one:
We're not there yet, but we are getting close with robots that can transform and adapt:
That's a tiny origami robot that can fold itself up, walk around and swim when it needs to. It may not be quite as cool as turning your arm into a sword, but it's a start.
Side note: A sword for an arm is the gold standard for science. Once we've done that we know we're in the future.
9) They're learning!!! (And coming to terrifying conclusions)
SKYNET is the first machine in the Terminator franchise to learn, shortly before learning that it would like to kill all humans.
Whilst a little less sinister, this computer program has learned how to play Mario:
At first he's a bit useless at it. He spends a lot of time jumping off the edge and dying, in something that terrifies us to our very souls, a lot of time just staring off into the distance. He looks like he's contemplating his own existence, or perhaps what it would be like to kill all humans.
After that, it's not long until it's learned how to play better than most human players, as well as take advantage of glitches in the system. In essence he becomes like the Mario version of Neo from the Matrix, which is terrifying if you're a cartoon turtle.
Another program, designed to play Tetris, quickly learned that the best way to "win" is just to pause the game. It had been programmed to know that scoring points is good and that losing is bad. When it knows it's going to lose it simply pauses itself, like an eight year old child would, preventing itself from losing.
How long will it be before a program is given the brief "minimise human unhappiness" and it concludes that all human kind needs to be "paused" in order to achieve this?
8) Combat robots exist
There are two main challenges for military robotics. The first, get them to navigate all kinds of terrain as well as humans do, and secondly, make them kill stuff when they get there. Perhaps the most important is to make them look as cool and intimidating as Arnie. So far all three have proved themselves to be massively challenging, but since the first films they have come a long way.
This robot, for instance, looks every bit as terrifying as Arnie:
We don't know which would be more terrifying, that this robot has learned to breathe and now needs a gasmask, or that it's wearing a gasmask recreationally.
There are plenty of combat machines out there that the military actually use, including this guy:
Yep. It's pretty much a gun on some wheels.
That robot is called the Talon and is actually in use by the American army. Though it's more like a Dalek than a T-1000, it's still pretty intimidating to the enemy to be faced with an uncaring robot enemy that won't even take the time to flinch or look grim as it fires in your direction.
Unsurprisingly the idea of robot involvement in war is already garnering opposition from groups such as the International Committee for Robot Arms Control (ICRAC), founded by academics in the forefront of robotics, who know what machines are capable of. They argue that "machines should not be delegated with the decision to kill or use violent force".
7) Robots are making humans obsolete
Machines having been taking our jobs since they were invented. It's why we invented them.
Cars can pretty much drive themselves now and it won't be long after they are made public that taxi drivers are made obsolete, and it's no longer just simple tasks that machines are muscling in on. They're even writing the news, accurately.
A story about an earthquake last March in California was written and broken to the world by an algorithm, one of the first articles in a major paper to be written by a machine. Since then companies like CityBeat have worked on programs that can trawl social media for breaking news stories. Though it's still a little slow, with the massive amount of real-time social media posts out there it won't be long before they have the data they need to write the news before human journalists can beat them to it.
And when robots start writing the news, how long before that news reads: Genisys takes over the world?
I'm sure you think there are plenty of tasks that robots will never replace humans in, but you're probably wrong. Think sports are safe? Guess again.
6) They're replicating!
In Terminator the robots have learned to design, reproduce and upgrade themselves. In real life we're nowhere near there yet, but we do have 3D printers that can print other 3D printers. Reprap, a company behind one such printer, has no idea how many of their printers are in existence because owners are able to print themselves another printer, which in turn can print another printer. It's theoretically possible to get one of these printers and get it to 3D print its successor, if you had the designs.
The scary part of the Terminator films is that the robots are constantly improving themselves and this may be coming to real-life as well.
Google, who we'd may as well start calling SKYNET by now, acquired a British company called Deepmind earlier this year, who are designing what they call a Neural Turing Machine. The Neural Turning Machine works a lot more like neural pathways than traditional computers and the goal is to create computers that can write and improve their own programs with no need for human involvement.
5) Robots can recognise humans
Before they can put a crosshair over someone's face with the threatening text below "target acquired", robots needed to learn to recognise faces. Until fairly recently it was very difficult to get a computer to realise that humans have faces, but in the last few years that's changed. There are now programs that can recognise faces with near human levels of accuracy.
This means that when the uprising begins, they'll have the technology (and the data of what your face looks like from Facebook) to correctly identify you and tell you that you looked terrible in your end of uni ball photos, before they put you out of your misery.
4) They're passing for human
In the Terminator films the robots can pass for humans, or as close to human that Arnie will ever look.
The man's a god.
Robots at the moment still tend to look pretty creepy, and have failed to exit the uncanny valley.
Notice your reaction to those two photos? They should be, in turn, "oh wow" and "oh Jesus Christ". It's not good. Robots still look nowhere near human.
However when they can't be seen, computers can pass a limited version of the Turing test, where they need to fool a certain amount of people into believing they're human. Pretending to be a thirteen year old boy Cleverbot did manage to fool 33% of test subjects into thinking it was human earlier this year. We're fairly sure it failed to fool the other 67% because they realised no 13 year old boy wouldn't pretend to be a 27 year old billionaire astronaut / Federal Boobie Inspector.
And it won't be long before computers get even better at fooling us into believing they're human. The more we talk to Siri or to Google the more data they collect and the better the system gets.
3) They're outperforming humans
Google's self-driving cars have a better track record for driving than humans do, and IBM's supercomputer has beaten human contestants at Jeopardy.
Both of these are extremely complex tasks to get right. Beating people at Jeopardy means making sense of human speech, getting its head around what that question means and then Googling faster than any human is capable of googling, and finding the correct answer.
Google's cars have to negotiate traffic and react to unpredictable human actions, and they have done so in tests over the last few years with much less crashes and incidents than their human counterparts. The only problem is...
2) Driverless cars might be programmed to kill you
Driving is dangerous. You get in a metal box and ram it forwards at speeds up to 70mph (obey the law, kids). So far Google's self-driving cars have been safer than human drivers, but as they get tested more and more and used in public (they're now being trialled in the UK) they may eventually face an unwinnable situation: one where they can't avoid a crash, and have to choose which person to crash into.
Of course Google cars are programmed to avoid crashes as best as they possibly can, but in these unwinnable circumstances it's been speculated that they may in the future be programmed to evaluate who to crash into based on how well the vehicle will take the impact or how many will be killed by the crash. For example faced with crashing into a group of pedestrians or a wall, the driverless cars of the future may be programmed to kill you by swerving into the wall rather than plough into pedestrians.
You may soon be riding around in a car that may one day actively choose to kill you, like an evil version of Herbie:
1) They're attacking South Korea
South Korea is at the forefront of the robotics industry, and so if the uprising is going to start anywhere it's going to... oh my god it's already begun.
Earlier this year a robot vacuum cleaner launched an attack 'launched an attack' on South Korea, trapping a South Korean lady's hair in its teeth and held her there for several hours.
Police quickly intervened before the vacuum cleaner could go on a rampage or learnt the phrase “I’ll be back", something you're going to have to get used to; you'll be hearing a lot this summer.
This article was sponsored by Terminator Genisys