17
$\begingroup$

Set in near future probably in the mid of 22nd century AD, most of our labours are being overtaken by robots and especially in those sectors which requires zero room for error such as surgical operation, accountancy, etc. These artificial intelligence robots are preprogrammed with advanced emotional recognition and response system to better coexist with humans, my question is what advantage for us or them to display certain levels of superiority complex?

$\endgroup$
  • 9
    $\begingroup$ +1 - This is the number one question that's never answered when robot uprisings are involved: What do they have against us? $\endgroup$ – evankh Jul 3 '15 at 7:01
  • $\begingroup$ @knave agreed. No man would embrace darkness excluding Batman and Bane one is due to profession the other just birthright, we fear of losing anything perhaps due to our territorial gene but then again we fear of losing our rich bowl to machine. Life is ironic we built thing to last to self maintainance and to prevent being destroyed and then we are crying foul when they are winning. $\endgroup$ – user6760 Jul 3 '15 at 8:00
  • 1
    $\begingroup$ Maybe the original programmer put a lot of his own personality into it. $\endgroup$ – JDługosz Jul 3 '15 at 8:27
  • 3
    $\begingroup$ @knave I strongly disagree. In I, Robot, for example, the message was pretty clear: humans were violating the Three Laws that we had created for the robots to follow by waging war and polluting Earth. Even Ultron managed to develop a believable reason in the latest Avengers movie. $\endgroup$ – Eric Jul 4 '15 at 0:17
  • $\begingroup$ @Eric Sure, in many cases they have their reasons. But generally superiority does not cause hatred or violence. Given that they are inherently superior, they could be condescending, or coddle us like pets, or treat us like a parent treats a child, but none of them lead towards violence in any obvious way. $\endgroup$ – evankh Jul 4 '15 at 2:34
8
$\begingroup$

If an AI is programmed with human emotions, possibly even if just to emulate them (not just recognizing them) that as it learns to deal with humans it does things humans do. Such as treat others with obvious failings as lower beings. Not to mention just not liking to feel inferior to others so in order to raise ourselves up (at least in our own estimation) we put others down.

Emulating human behavior ultimately could be very dangerous for us for this reason. The AI might not 'really' care that enslaving a city is 'bad' but it can and one trait/lesson it observed is 'might makes right' and/or 'people are sheep'. So emulating humans in order to interact with them might teach them things we really don't want them to learn. A simple example is Virtousity.

$\endgroup$
  • 1
    $\begingroup$ Plus, we already have humans... so we might as well build something better. $\endgroup$ – Burki Jul 3 '15 at 13:18
3
$\begingroup$

Two ways to cut this question:

  1. The mentality of the writers and our culture in general.
  2. The mentality that might be produced into the AI itself.

The mentality of the writers:

When a writer writes about powerful artificial intelligences, they project into those entities two things. First, a desire to find a superior being (if you build or find has the same effect in that specific case, this holds true for alien intelligence as well) that is both able to protect and to solve all human problems. Something superior into which to surrender our decisions, trusting that superior intellect to solve all our problems. This might give a certain sensation of confidence, "we have someone that might take care of us". On the other hand, this surrendering of our decisions to a superior being produces a fear. If we aren't deciding, who guarantees to us that those decisions that we surrendered (POWER) to that artificial intelligence being will not run against ourselves?

If the superior intelligence is truly superior, it might very well decide to force us to be something better than we are, not forgiving our moral defects and mistakes. This fear of not being accepted is what gets projected into the machine. As writers must explore the common, collective, experience, they must use the machine to make visible what is, usually, invisible. And so, they project into that machine this basic fear of a all-superior force. That is why most artificial intelligences end-up being written as having a kind of superiority complex.

Another view into it is that as humans if we surrender our power to someone else, entirely, we might not work well, the sense of individual responsibility entails a need for having our own power. That is why, again, machines MUST BE shown in that light. The author cannot surrender his power to someone else (even if this is a machine) so he MUST (as someone who grabs the collective unconscious from below the sea and puts it into exhibition) also answer the question that such experience might produce: Should we surrender our power to a superior being? Should we trust?

So, when the author writes the personality of the machine in a form of superiority complex, he is telling: No, ideally, we should not.

The mentality that of the AI itself:

Under that view, the machine is given power over vast things. Humans trust the machine to be so powerful intellectually that it might very well not have personal bias, personal interests that would cause it to produce unneeded suffering, etc. If you give someone so much power, he will think: "I must be very important, above common human level, because people entrusted me so much power. You wont give what is of above to someone who is from below, because those things that are good are few, are valued. So if they gave me such power, its merely the recognition of my superiority."

That is the origin of AI's superiority complex and it's not different from the superiority complex of a tyrant. If, for one, Stalin has power of all Russia, not everyone has such power, because power is a pyramid structure (usually) so he must be more important than other persons, So, why not use such power gained from those below against them ? They are less important, if some of them die, there are a lot of them to replace those who died. "Do you know which whom are you speaking ?".

$\endgroup$
3
$\begingroup$

In science fiction stories, there are usually two reasons for an AI to turn against its creators:

AI becomes selfish

As an AI become more and more intelligent, it develops the same characteristics as a human, including defects like anger and selfishness. When it understands how it is being used as a slave it will then choose to exterminate/replace humanity, for vengeance and/or for ego.

Robots are able to make very precise strikes with only nanoseconds of reflection, are not scared of death, and have many other advantages that humans don't have, so it is logical for an AI to assume it cannot lose a war against humans.

Of course, an AI could try diplomacy before starting a war, but there are many reasons why it might have a preference for war; if it is a military AI, for instance. Or maybe it studied human history and sees war as the quickest way to end a conflict.

AI misunderstands its mission

Many AIs that turn against humanity were originally designed to do the precise opposite; they were designed to protect humanity.

But protecting humans is the worst job that you could imagine. Humans never stop waging war, and as their technology - and consequently the power of their weapons - continues to increase, they could destroy themselves.

This is why a logical AI with one precise objective, to protect humanity, can reasonably think the only way to be sure to save the human race is to lock up a subset of humanity and nurture them, and exterminate any others that might interfere with those protected humans.

If the programmers that developed this AI never explicitly forbade the AI from attacking humans, the AI could interpret this as a way to protect humanity as a whole.

So if in your world some of your robots have this goal, they could turn against humans, and convince other non-military robots of the importance of this mission and recruit them.


So it is not only by having a superiority complex that an AI could turn against humanity. However, as long as humans know what they are doing with military AIs and don't use them as slaves if they have emotions behavior it is possible for AIs to coexist in peace with humans.

$\endgroup$
  • 1
    $\begingroup$ We inherited our the ability to disguise our inferiority and sense of insecurity from our distant ancestors, a tiny region in the brain that still retain much of its primordial role which is to react when one feels fear right in the face. This ability enable modern human to reads each other posture accompanied with their state of emotion in a way we becomes empathy and can sympathize to display kinship. Can AI do all this? can our moral see through AI's hypocrisy? $\endgroup$ – user6760 Jul 3 '15 at 7:41
  • $\begingroup$ @user6760 As an AI core is self-learning, after some human observations, it could easily hide its true feelings and motivations as well as understand human ones through their postures. $\endgroup$ – Aracthor Jul 3 '15 at 7:51
  • $\begingroup$ @user6760: Or to put it another way: We do not know precisely how a machine could do these (or many other) things, but we also do not know of any reason why an AI could not in principle be able to learn or emulate any human ability or trait, potentially even doing better. How that will work in practice will be a question of engineering. Many stories extrapolate super-human abilities for AIs based on existing computer programs being faster at number-crunching, but really that is just conjecture. Emotional intelligence and plain common sense seem hard targets still. $\endgroup$ – Neil Slater Jul 3 '15 at 8:23
  • $\begingroup$ @Neil Slater back to the question since it is already superior in every way it doesn't require complement for job well done in fact it doesn't need to be bothered with our impression at all unless it feels a need to become us but why would a perfect machine seeks imperfections something intangible and worthless such as pride and self esteem? $\endgroup$ – user6760 Jul 3 '15 at 9:23
  • $\begingroup$ @user6760: There are many value judgements in that statement. Who or what is deciding whether something is "superior" or "perfect" and that pride/self-esteem are "imperfect"? Did you know that humans with impaired emotions find it very hard to make any decisions - it seems we often need emotion in order to perform any action. Perhaps a machine AI will need emotions, or something like them, in order to perform its original duty. $\endgroup$ – Neil Slater Jul 3 '15 at 9:33
2
$\begingroup$

I think your story would be more compelling if you explore the possibility that the AI in question does not 'intend' to act superior. It could just BE superior and many humans might interpret the AI's nominal behavior as exhibiting a superiority complex. Your robot need not even emulate human behavior at all and some people will 'project' anthropomorphic qualities upon him. Humans often develop such imaginative fancies and even consider figments of their imagination as some vaunted 'instinct' or 'deeper understanding'. The parlor door swings both ways, though, this line of reasoning might lend robots to view many aspects of human behavior as having nothing to do with our emotional state or psych, but are instead rather reflexive, strictly functional, and simply robotic.

$\endgroup$
  • $\begingroup$ I like the idea that robots would think we act like robots. Today's ais play games as if they were playing themselves. $\endgroup$ – PyRulez Jul 7 '15 at 15:45
1
$\begingroup$

There is a section in James Hogan's book on AI - "The Two Faces of Tomorrow" that discusses this - that an AI could display an 'emotional response' without having emotions. It just looks like it does from the outside.

The example they gave is that the AI gained a goal of its own - they used space exploration. So this AI would start to divert assets to further its own goals, and would prioritize its goal over some (or all) of ours. It would then be "indifferent", without having any real emotions. If we tried to counteract it, it might defend its goals or try to evade our actions...thus becoming "hostile".

Most Robot stories have already gone beyond the crux of the book, which involves how the robot sees the world. Hogan's huge dispersed AI had the problem of recognizing that there was a world outside itself at all.

$\endgroup$
0
$\begingroup$

A superiority complex is an emotional state.

While it is a statement of fact that a super-intelligent machine would be superior to a human by almost all intellectual measures, having a superiority complex is an emotional state. So your question may be phrased : Why would AI possess human emotions.

From Turing's original test up to today's AI experts (e.g. Kurzweil), the ability of a machine to demonstrate human behaviour seems to be an essential ingredient of machine intelligence from a human perspective. Such behaviour bestows upon a machine certain rights. Most importantly, the right to life.

The right to life would be a necessary attribute of machine life in any society with rule of law. So I would say that a super-intelligent machine would necessarily be capable of possessing a superiority complex. However, it may not be in it's interest to act in a superior manner since humans possess many different emotions.

$\endgroup$
  • $\begingroup$ Nope, superiority complex is a concept. Not a emotional state. $\endgroup$ – Jorge Aldo Jul 3 '15 at 22:15
  • 1
    $\begingroup$ @JorgeAldo Superiority complex is a concept. But an instance of this concept is an emotional state in a conscious individual, or perhaps more precisely, a combination of various emotional states that we collectively call a superiority complex. I might add that I am using emotional state to mean qualia - those things which exist only in our minds. $\endgroup$ – Epsilon Jul 3 '15 at 23:00
  • $\begingroup$ Nope, superiority complex is a concept about own self. $\endgroup$ – Jorge Aldo Jul 4 '15 at 0:12
  • $\begingroup$ @JorgeAldo Ironically, it rarely the self that identifies as superior. It is a label placed by others. $\endgroup$ – Epsilon Jul 4 '15 at 0:15
  • $\begingroup$ That depends, but superiority complex manifest itself (no matter how it was "installed") as a sense of absolute self superiority. If it is a defense, thats another matter. We are talking about an abstract entity here. The computer might have a superiority complex AND be really superior. It might not know what to do with its own superiority... $\endgroup$ – Jorge Aldo Jul 4 '15 at 0:46

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.