— Let’s pretend it’s the not-so-distant future. There you are, standing in the pile of rubble that used to be your home, minding your own business, when suddenly you’re confronted by a hulking humanoid robot with glowing red eyes.
While this robot may not be shouting “Kill! Kill! Kill!” — thanks to its body language and the whirring saws it has where its hands should be, you’re about 99.999 percent sure that’s precisely what it has in mind.
The question then becomes, should you:
A) Punch the robot with your mighty human fists
B) Call it hurtful names like “lug-nuts-for-brains” or “rusty arse” or "inferior cybernetic unit with outdated software"
C) Throw mud in its eyes
If you chose answer C, then give yourself a pat on the back. You’re well on your way to surviving a robot uprising.
With the new “Terminator Salvation” film in theaters, robots running amok seem to be on everyone’s minds (or at least
on my mind). And so I consulted one of humanity’s foremost experts on the subject — Daniel H. Wilson, an honest-to-god roboticist (who got his Ph.D. from the Robotics Institute at Carnegie Mellon University no less) and the author of the book “ How to Survive a Robot Uprising.”
If the apocalyptic future depicted in “Terminator Salvation” were to actually come true, Wilson’s book contains all sorts of helpful advice — advice you'll find in sections titled “How to Spot a Robot Mimicking a Human” and “How to Fool a Thermal Imaging Target Tracker” and “How to Treat a Laser Wound.”
While Wilson takes a cheeky approach to talking about robots overthrowing their meatbag masters, he does believe that humanoid robots will
one day be used to fight our wars and his book offers an approachable look at real-world robotics through the looking glass of popular culture.
When it comes to triumphing over a good robot gone bad, for example, he points out that “sensors are by far the most vulnerable, exposed parts of any robot.” And so if you want to blind an attacking automaton, “a handful of dirt, mud or water will suffice.”
He also points out, “It’s hard for a robot to wipe mud from its eyes when it has whirring buzz saws for hands.”
After checking out “Terminator Salvation,” I checked in with Wilson to get his roboticist perspective on the film. Here’s what he had to say about real-world robotics gone Hollywood and the likelihood of a robo-pocolypse ever happening. (A quick warning: This discussion contains spoilers aplenty.)
Did the first “Terminator” movie affect you as a kid and your interest in robotics?
It affected me a lot as a kid to see a humanoid robot and to see inside it. You really saw all the tendons and the joints and the skull and the whole skeleton of the thing. It really gave you the impression that this was a real walking, talking, murdering wind-up toy. And then I started thinking, well how would that really work? Can we really do that?
I think that “Terminator,” even though it has robots that kill everybody, has been central in helping a lot of people get interested in learning about robots.
As a roboticist, what did you think of “Terminator Salvation”?
I loved it. People were sweating it for technical accuracy and that blew me away because the movie was incredibly technically accurate from my perspective.
Obviously it's a movie, but what seemed realistic based on what you know about robotics today?
The behavior of the robots I found to be very accurate. The T-600 attacking John Connor, I thought that was quite realistic. It looked like a piece of machinery. You could see the pieces moving.
I also liked the part where another T-600 got caught in a snare and was hanging upside down by its foot. It was just so logical the way it behaved — it hung upside down, it was able to target the humans, it shot at them but it wasn’t able to be as accurate. When the humans were gone, it realized that it had a different problem — that it was trapped upside down. It saw that its foot was the problem and it shot its foot off. It was just basically executing a solution to a problem.
Also, I’ve got to mention the
robots in the water. Those are similar to modular robots — they’re made out of individual modules that are connected together and are able to execute different gait patterns. For example, snakes don’t all travel in the same way. There’s side-winding, there’s inch-worming, there’s your conventional wriggling. There are all these different ways that a modular robot can locomote, and they’re applicable in different situations. They can swim like sea snakes in water, wriggle across rough terrain on land, or even climb trees. (For an example of what he’s talking about, check out this modular snake robot from Carnegie Mellon.)
One non-obvious thing I really did like — did you notice that in the kidnapping scenes people were constantly getting snatched up by these big claws and they were never injured in the least? You know what, that is accurate. If a machine has force sensors in its manipulators, it can — to an incredibly accurate degree — regulate the amount of force it uses. (Check out this article for a peek at just such a robot.)
In the real world today, there’s a lot focus on how we can improve our bodies
by incorporating technology into them. So it seemed to me that the cyborg character Marcus Wright was one of the more “realistic” things about the movie.
Absolutely. I think the whole course of human history has been marching toward that. As a species we are tool makers and we use our tools to increase our own abilities. You can’t hammer a nail in with your fist, so you use a tool for that. For us to incorporate these tools into our bodies is nothing new. Prosthetics have been around for literally thousands of years.
To sort of see the logical conclusion of that in “Terminator Salvation” was exciting. It tapped into something really ancient and really human. Also, it’s interesting to have that character say, “What am I?” This guy doesn’t know whether he’s human anymore and, you know, honestly that’s something that’s going to be affecting us in our real lives not too long from now.
Knowing what you know about robots, what didn’t work for you in the movie?
I was a little thrown by the very end and by the presence of a human heart in the cyborg. Although it was an infiltration unit — from a tactical perspective, it makes no sense to put a vulnerable beating heart right in the front chest cavity of an otherwise indestructible robot.
Also, the giant humanoid robot didn’t make any sense to me. First of all, it snuck up on them somehow. That thing would be loud and that thing would be smelly. It’s a gigantic machine. Don’t you notice if a cement truck is backing up to your house?
And something I have to register a formal complaint about — a Terminator should never, ever close in on its human target, get a hold of its human target and then throw its human target as far away from itself as it can. In what universe does it make sense to be big and slow, to close in your prey and then to hurl your prey across the room where it will inevitably get up and run away again?
Can you tell me some of the reasons why you think a robot uprising won’t happen in the future?
Any technology can be used for good or evil. So human beings will use robots to kill each other. And human beings will use robots to save each other from life threatening disease, from dangerous situations and all kinds of stuff. But a robot uprising is different because you’re talking about the machine itself deciding to attack humanity. I just don’t see that happening.
What that entails is solving something called the “strong AI” problem — creating a machine with human-level or better intelligence. If you solve that problem, you’ve also by definition solved every other artificial intelligence problem there is. That is really, really hard. And in fiction it often just spontaneously happens. In the “Terminator” films, Skynet just “comes online” magically. There’s just no indication to me that this could suddenly happen without the intense scrutiny of some of the smartest human scientists on the planet.
The second thing that has to happen — which I don’t think will ever happen — is the AI is going to have to decide that it wants to wipe out humanity. If I’m a super human intellect, why am I going to destroy the most interesting phenomenon that exists in the solar system and maybe the universe? You can study the composition of all the rocks in the asteroid belt and you can map out all the mating rituals of every flatworm and reptile and bird on the planet, but honestly I don’t think there’ s anything more interesting than human beings. And if I was a superhuman intellect, I’d want to keep them around.
Can you give me any reason why a robot uprising could actually happen in the future?
Human beings are not that great at looking at long-term outcomes. We do things like introduce rabbits to Australia. Somebody decided that it would be fun to hunt rabbits and they introduced them to Australia and they multiplied to the extent that they basically covered the content and killed countless indigenous species.
We often see our plans reproduce and get out of control. So, there could be a situation through human stupidity in which we create some kind of robotic technology that gets out of hand.
If I suddenly found myself in the world envisioned by "Terminator Salvation," what three pieces of survival advice would you give me?
One, always, always go for the sensors. Kyle Reese obeyed the number one rule of human-robot warfare when he neatly wedged a piece of rebar into the brain stem of an attacking Terminator. The maneuver damaged the Terminator's inertial measurement unit and disoriented it long enough for the humans to escape.
Two, stay off the radio. Whether the signal is encrypted or not, it is exceedingly easy to track the source of a transmitted radio signal. If you have to send a communication, do it while you're on the move and keep your message brief. If you can, let John Connor deliver the long-winded fireside chats, while you high-tail it for the hills.
Three, know your enemy. Clearly, Terminators excel at killing other Terminators. Learning to hack into or reverse-engineer the robots is a key survival skill in the future. If John Connor had bothered to perform even a cursory autopsy of the cyborg Marcus Wright instead of hoisting him dramatically from chains, he would have found an inhibitor chip placed conveniently on the back of his head.