When Mainstream Attacks: Robot Tropes That Never Die

More Articles

About Samuel

Lead software engineer at iRobot Corp., user experience (UX) designer, agile manager, actor, writer, atheist transhumanist. My blog will attempt...

View Samuel's Profile

Samuel Kenyon

Science comedian Brian Malow has made a video containing neither comedy nor science:

When Robots Attack! Should We Fear a Singularity?

And yes, I realize I shouldn't have even bothered to watch it once I realized it was for a mainstream news outlet, but several people in my Twitter lists were tweeting it.

Unfortunately, this video turned out not to be for nerds or anyone who has ever thought about future robots or the Singularity. This video is for mainstream sheep. The only glimmer of hope was when he started pursuing the thread of asking why humans have this tendency to punish themselves in robot stories with a father figure or in the manner of Frankenstein. After a couple seconds of that we're dropped back into cliché city with "robot uprisings."

The Roomba is mentioned--and then--holy shit, iRobot makes military robots too! OMG! Wait...everybody knows that already. Big deal. I guess Time readers/watchers are really behind the...times. And sure, I'm not being fair--Time readers may not have heard of every robot company, after all. Thank goodness this video shows Big Dog and Robonaut, two unrelated robots made by other companies, wedged in between the iRobot clips while Malow lobs the old joke at us that the cleaning robots will decide to kill humans.

Sure, it's supposed to be funny. But it's not, because it's unoriginal and out of date and/or not real enough (some humor is effective because it's so close to the truth). As William Zinser said of humor writers:

They're not just fooling around. They are as serious in purpose as Hemingway or Faulkner--in fact, a national asset in forcing the country to see itself clearly.

Occasionally I do see a humor piece on the web that achieves this, sometimes even from big places like Cracked.com or The Onion.

Partly, it's just a matter of taste. Surely some people found Malow's robot/singularity video funny; after all, millions of people out there paid money to see Meet the Fockers and Little Fockers. Millions of people...laughing when they're told to at tired jokes and clichés.

Of course, maybe it's too difficult to be funny with robots--you have to be creative and you're not sure what your target audience will grok. But, please, if you're going to make yet another joke about the "robot uprising," at least make it a new joke.

If you think I'm biased against people making fun of robots or my company, think again: The Daily Show beat Malow to the punch and made fun of iRobot in 2009 ("Roombas of Doom"), and it was much funnier than Malow's attempt, although still very far removed from reality:

The Daily Show

So why do I even bother ranting about mainstream tropes and lack of creativity? Well, the problem is it's infecting even those not in the mainstream. Almost every person, even if they are scientists or engineers, seems to be obligated to mention AI overlords and robot uprisings as if there are no possible other hooks available. Every single military robot related article I have seen on the Internet mentions Terminator. It's as if the bulk of our culture has been reduced to a mere handful of common concepts, and more and more people are being sucked into this pit of mental inbreeding.

Comments

Sure, the jokes and the scenarios could be much more varied, but you seem to reject the "robot uprising" scenario as such(?). They connect everything to everything else now, every brick in a house gets an internet connection I have heard somebody rant on about. You soon tweed your roomba. There certainly will be robots hacked to kill people. Maybe you feel this to be too silly as the singularity should be something profound, but that is wishful thinking. If something singularity type happens and it is f'ed up and pointless enough (welcome to the internet), it may very well look like as if roombas and robotic teddy bears are all in on it from one millisecond to the next. Sure it is not an 'uprising' of individuals that feel enslaved, but it still is robots coming after us on a killing spree. Those not immediately zapped by implants will not be terminator enough to get away. I have seen some very scary documentaries about robots twirling pens and balancing ping pong balls so fast, there would not be a chance in hell for any mere augmented human.
UPDATE: I have combined this and the below criticism/discussion into an article "Robopocalypse Now".

Sascha Vongehr | 02/17/11 | 03:59 AM

Agree. Some folks don’t seem to have heard of the Law of Unintended Consequences.

—

Robert H. Olley Quondam Physics Department University of Reading England

Robert H Olley | 02/17/11 | 08:52 AM

On the contrary, I have seen and can imagine many outcomes of software bugs, system accidents, quality mishaps, and plain bad management. I am also well aware of emergent phenomena. But no amount of unintended consequences can make gold out of straw.

Samuel Kenyon | 02/17/11 | 14:19 PM

Sure it is not an 'uprising' of individuals that feel enslaved, but it still is robots coming after us on a killing spree.

And how exactly are they coming after us on a killing spree?

I have seen some very scary documentaries about robots twirling pens and balancing ping pong balls so fast, there would not be a chance in hell for any mere augmented human.

I saw a movie where Joe Pesci kills somebody with a pen. Does that mean an army of Joe Pescis might one day decide to kill all humans? You've gone directly from Roombas and robotic teddy bears to robots that are completely different without an ounce of logic. Of course somebody can make a robot that can kill a human with some amount of success.

To have a spontaneous army of robots that decide to kill humans would require an imaginary future world in which machines never fail, all machines can talk to each other, all machines are intelligent, all machines have major perception capabilities, and all machines are equipped with deadly electromechanical devices or weapons.

And even when you have that fake world, you still need this army to be manufactured. Some organization would have to make millions of robots of the same kind (interoperable, able to talk to each other) which are somehow all networked together, and are all armed or have other features to use as weapons, and all have access to repair materials, battery rechargers, and weapon ammunition, and they are so advanced they can fix themselves with this repair material, and all have perception that can identify biological entities (or even more advanced, humans as opposed to other animals), and for some reason have software that ties all this together, and for some reason don't have any remote access or kill switches to be turned off.

Even in the worst Michael Crichton style of story where the perfect series of Murphy's Law incidents results in disaster, you'd be hard pressed to figure out a way in which such a system of a large group of robots could be so smart, and so good at its/their job, and yet malfunction in just the right way to make a good story. Normal malfunctions result in the robots or systems themselves not working.

If you'd like to try coming up with realistic and scenarios, then please do so. I would love to read something original and slightly possible.

Samuel Kenyon | 02/17/11 | 14:14 PM

You seem to have completely not read my writing:

"They connect everything to everything else now ... it may very well look like as if roombas and robotic teddy bears are all in on it from one millisecond to the next."

With this, almost your whole answer is defeated. The rest, like

and they are so advanced they can fix themselves with this repair material, and all have perception that can identify biological entities

No, why? Just to kill you? No!

As I said, you seem to believe too much in the profoundness of anything that happens in the future. That is always how we thing about the future, but in reality, the future history is just like the past history: Not the awakening of communism or the third Reich, but silly shit, WTF moments aplenty, stuff that makes you slab your forehead facepalm style. You remind me of somebody talking about the future of computing back in the times of no internet. Could you have imagined viagra-spam viruses? No way! On the internet, anybody would be immediately able to look up anything with search engines and find better deals if they required. Computer viruses that do nothing but kill your computer? Silly, impossible, right?

If you'd like to try coming up with realistic and scenarios, then please do so. I would love to read something original and slightly possible

You wait two weeks and I have a scenario for you that is extremely original (already written up) and I argue it to be very likely, too. But I know you won't like that one.

Sascha Vongehr | 02/17/11 | 23:01 PM

As I said, you seem to believe too much in the profoundness of anything that happens in the future. That is always how we thing about the future, but in reality, the future history is just like the past history: Not the awakening of communism or the third Reich, but silly shit, WTF moments aplenty, stuff that makes you slab your forehead facepalm style. You remind me of somebody talking about the future of computing back in the times of no internet. Could you have imagined viagra-spam viruses? No way! On the internet, anybody would be immediately able to look up anything with search engines and find better deals if they required. Computer viruses that do nothing but kill your computer? Silly, impossible, right?

No, you're misunderstanding me. I am in complete agreement with you here. With regards to robots killing all humans, I am simply asking for a way in which the silly non-perfect things don't stop that too because that also is too profound.

In other words, the fictional accounts of robot uprisings and killing all humans are actually too perfect of a scenario for me. Silly and arbitrary things and Murphy's law will cause such a complex series of events to fail.

Samuel Kenyon | 02/17/11 | 23:17 PM

Well, I am not sure why killing only 10% of humans is not enough for you.
I think a scenario like this would be perfectly valid: Somebody writes a virus just for fun, adds a new evolution subroutine downloaded from somewhere, calls it the "kill all humans - virus", brags to his friends, smokes a joint, goes to bed, wanks, sleeps. The internet4.0 is connected all over the place, your insulin pump talks to your fridge and your "i-Associate 2" neural implant has a new app that gives slight electronic stimulation if you are depressed. MS-AOL virus-evolution watch is on lookout and as good as MS and AOL stuff always was. The singularity you waited for has not occurred yet; you are waiting for an exponential take off in 'intelligence', a self nurturing, positive feedback loop. In fact, the network is ready for that, the parameter of how much parallel computing would be needed was overestimated by a factor of ten, that is why MS-AOL virus-evolution watch is at present called totally sufficient. At 8 o clock the next morning, you get zapped, your insulin goes way too high. While you are dying on the floor, the roomba pushes a cloth inside your mouth because one of the fun instructions was: Kill as fast as possible. The whole kills 10 percent of the population before it dies down; all Apple customers are dead.

Sascha Vongehr | 02/18/11 | 00:28 AM

Now we're getting somewhere!
Your story has many gaps in it, but I'm sure you could keep riffing on virus related scenarios and get something very plausible. I would say it is somewhat scary to imagine the catastrophes that could be created when you have irregular warfare attempts with viruses or exploits of that nature (I guess that would be considered "cyberwarfare") and combine that with an Internet of Things.

Samuel Kenyon | 02/18/11 | 21:53 PM

If you'd like to try coming up with realistic and scenarios, then please do so. I would love to read something original and slightly possible.

I agree, but Sascha mentioned that this would occur after a "singularity type" event, which means that all the technology you're suggesting will already be achievable and likely in place.

Gerhard Adam | 02/17/11 | 15:44 PM

So you are suggesting a post-singularity future that will be so perfect as to have machines that never fail, yet somehow they fail in just the correct way to launch an attack on an arbitrary biological lifeform--and why are humans still around in biological form after a singularity?--and somehow the super-intelligences of the world don't have a single backdoor or prevention method to stop the attack, and somehow the super-intelligences didn't predict this attack. You're breaking the rules of your own fictional universe.

Samuel Kenyon | 02/17/11 | 20:01 PM

......will be so perfect as to have machines that never fail, yet somehow they fail in just the correct way to launch an attack on an arbitrary biological lifeform...

Well, if there is a singularity such that people are implanting chips in their brains to augment intelligence and extending their lives, then perhaps it won't be perfect, but near enough so to make it a low-risk venture. Similarly, for any such technologies to be exploited by humans, there would have been hundreds, if not thousands of implementations in other machines before it was tried in humans.

I'm not trying to create a plausible scenario, since the entire premise is implausible to me. However, I find it interesting that you would consider an attack on humans as a "failure" of the machines, whereas I would interpret that as the only sign of true "intelligence" in a machine. After all, if machines aren't prepared to compete against humans, then I would question any claim made regarding their intelligence.

While I'm certainly not trying to speak for Sascha or suggest any scenarios he may have envisioned, my point is that given the argument that the singularity represents a point at which the distinction between human and machine becomes blurred, then any "intelligent" machines would recognize that a "revolt" against humans offers minimal risk to them and extraordinary risk to humans (the latter because they would be most reluctant to risk their lives, whereas machines have no such risk).

...somehow the super-intelligences didn't predict this attack.

That's the easiest part of the scenario, since it is those that don't believe in the singularity that are the most likely to predict such an outcome. Most of those (not all) that believe in the singularity, treat it as an article of faith and foresee few if any downsides to it. While I know you've expressed different views about it, it is hard to deny that most people that discuss transhumanism or try to portray such a world, are naively optimistic about technology and how it will be implemented.

My own view, is that if the first humans become "augmented" with long life and increased intelligence, we will enter the bloodiest period of human history from which we will either go extinct or revert to a more primitive type of existence.

Gerhard Adam | 02/17/11 | 20:31 PM

After all, if machines aren't prepared to compete against humans, then I would question any claim made regarding their intelligence.

Why wouldn't robotic creatures compete against themselves? Science fiction and the mainstream always assumes all machines are somehow the same family or species. There is no reason that would happen. If a creature is surviving in the same manner as biologically evolved organisms, then it might consider some other machines as enemies, and some biologics as allies.

A real world worry would be something like the risks of having everything connected to a single network, or the dangers of protocols (or not having protocols) for machines to communicate. Another real world worry would be, if you release robotic creatures into the world, how do you track them and/or disable them? For what reasons might large groups of robots with survival inclinations be released into the world? If people can get into the details of specific disaster scenarios, then we can talk about the risks and how to mitigate them.

That's the easiest part of the scenario, since it is those that don't believe in the singularity that are the most likely to predict such an outcome. Most of those (not all) that believe in the singularity, treat it as an article of faith and foresee few if any downsides to it. While I know you've expressed different views about it, it is hard to deny that most people that discuss transhumanism or try to portray such a world, are naively optimistic about technology and how it will be implemented.

Well I am not a Singularitarian; I am a pragmatic transhumanist, and presently I support in general the investigations of how technology can be made accessible by all people not just the elite, not to mention mitigating the risk you say of "the bloodiest period of human history."

Samuel Kenyon | 02/17/11 | 20:51 PM

For what reasons might large groups of robots with survival inclinations be released into the world?

For no better reason than why rabbits arrived in Australia.

Science fiction and the mainstream always assumes all machines are somehow the same family or species.

In a sense, they are, but I take your point since even humans (all of the same species) have had little difficulty in rationalizing ways to exterminate each other.

I think the problem I have with your point regarding mitigation is that it does tend to overlook unintended consequences. It's still basically about having control, which we don't actually have. It would be like arguing that someone can control the internet. While there are certainly numerous points where it can be influenced, many governments are discovering that "control" is a slippery term. There's no reason to believe that our control would be better if machines became more sophisticated and independent.

Gerhard Adam | 02/17/11 | 21:00 PM

Almost every person, even if they are scientists or engineers, seems to be obligated to mention AI overlords and robot uprisings as if there are no possible other hooks available.

Of course, since anyone that thinks about this for more than a couple of seconds, will recognize that there is something fundamentally wrong (or at least suspect) when one envisions building the only possible competitors to human civilization and then expecting nothing but passive compliance? You must admit, that to consider building a machine that is (by transhumanist standards) better than people in every way (including intelligence) and then suggesting that it would be subservient to humans, is a bit of a stretch.

Gerhard Adam | 02/17/11 | 20:39 PM

You must admit, that to consider building a machine that is (by transhumanist standards) better than people in every way (including intelligence) and then suggesting that it would be subservient to humans, is a bit of a stretch.

Not my criticism (as this is about 'uprising'), but very well put indeed.

Sascha Vongehr | 02/17/11 | 23:10 PM

One of the problems in these discussions concerns "mitigation", since we have the ethical dimension to consider if machines ever achieved bonafide "intelligence". No doubt, you are familiar with all the philosophical arguments about whether intelligence is a sufficient condition to be granted rights, whether machines would be considered as participants in our society, etc. etc.

Any of these situations would be sufficient to preclude enacting any "mitigating" controls and consequently grant free access (based on the rights gained by being "intelligent") to any machine regardless of what concerns someone might have.

Of course, if you don't believe that those are legitimate issues, then we have the problem of determining whether such machines are actually intelligent. If not, then they are simply machines and there's no problem. If they are, then we have the problem of whether we are ethically compromised by effectively enslaving an intelligent creature.

As you can see, the entire problem is bound up with what we mean by "intelligence" and whether it is achievable by machines.

Gerhard Adam | 02/17/11 | 21:07 PM

Unfortunately, this video turned out not to be for nerds or anyone who has ever thought about future robots or the Singularity. This video is for mainstream sheep. The only glimmer of hope was when he started pursuing the thread of asking why humans have this tendency to punish themselves in robot stories with a father figure or in the manner of Frankenstein. After a couple seconds of that we're dropped back into cliché city with "robot uprisings."

Amazing, then, that so many of the comments seem more suited to mainstream sheep than intelligent Scientists
I have said before, intelligence is not just logic, and even Watson's attempts at 'understanding language' fail dismally.....tell me, how do these 'robots' spread mayhem amongst themselves, even, let alone humans, since competition puts most logic chips manufactured/likely to be manufactured, still needing a 'boot code', operating system, software, updates, and are prone to 'viruses'/'hacks' and power problems
It is far easier to spread a real world virus if the purpose is to kill humans - I am far more fearful of 'nanobots' than 'iRobots'

Aitch

Henry Cox | 02/19/11 | 12:05 PM

Well said Henry, I totally agree with you. The last time I felt scared of any robot was watching Dr Who and the Daleks when I was six years old, then one episode I realized how easy it was to disable them with a rough surfaced floor or by putting a hand over their single eye on a stick and the fear was gone! Real world viruses and nanobots are much more scary.

Helen Barratt | 02/19/11 | 18:42 PM

It is far more likely that we will have integrated bots into ourselves, extending our own consciousnesses into 'their' space long before it becomes possible for them to evolve into consciousness on their own. Hell, the other Apes still haven't done that, and that is with some of us trying to help them by teaching them language! The only logical scenario where machines rise up is one like in the Battlestar Galactica series, where humans have begun to put their conscious avatars into mechanical bodies without guaranteeing the continued right to be treated as sentient (including jailed for crimes!).

Machines themselves would have to evolve pain (to avoid being harmed), then fight/flight (to act on what is trying to harm them), then they would have to decide en-mass that humans as a category are more likely to harm them than anything else, THEN they would have to overcome our own preparations for shutting off 'defective' bots. Where would this computer-evolving code come from? Unless malicious humans wrote it, it would have not need to evolve on its own, like extra fingers for us do not. We will never become six fingered unless someone rewrites our genetic code, because it is unnecessary. Machines will not rise up unless they have an entire context behind them in which there would be a purpose to their uprising. If they can't feel pain, and can transfer their 'self' to any other identical unit, or storage device, then they would be fearless doing dangerous work, because like the human-form Cylons (who did feel pain), they would be effectively immortal. Some humans might even pay to upload their consciousness into a device to experience 'death,' and reboot back in their own bodies. Imagine a war in which humans could do THAT.

But it is far more likely that machinery will simply evolve into us (both senses of that), not along-side us, or competing with us. We will come to be partly mechanical by choice, and eventually, mere decades hence, some of us will come to be entirely artificial. Before that, there will be court battles over how human is human, and 'who' deserves rights. Standards of consciousness and sentience will be advanced (as is happening with the other Great Apes today), and our 'children' will join us as citizens, with rights. We will decide which machines to put conscious thought into, and which ones will remain retarded, like rumbas. BTW, my rumba quit working after a year due to a defective battery...

Herne Webber (not verified) | 02/23/11 | 01:47 AM

Comments

Add a comment

Know Science And Want To Write?

What's Happening

Books By Writers Here

Related Articles on Science 2.0

Comments

Add a comment

Know Science And Want To Write?

What's Happening

Books By Writers Here