1
$\begingroup$

I am planning to write a fictional story: First communal war between Robots

Overview

This is 2100 and technology has massively developed. Robots have replaced and even abolished many job designations like "Driver", "Teacher", "Author", "Soldier" etc. and many humans are ending up being just a robot developer or programmer. Unemployment ratio is increasing tremendously in the world. The robots now do not have "auto destroy" button but have "auto-repair" and troubleshoot functionalities in case of error or bug. Moreover, Robots do not even have a manual shut down button. They work 24*7 from stored energy sources. This is the time when Robots will just start to make and develop other robots automatically-making many robot programmers, too unemployed. Humans have became so dependent on Robots that many of them have became either fully inefficient and lazy. Robots have ultra advanced consciousness-"The sixth sense". They can correct their own mistakes, interpret the facial expressions and can even assume the psychology of a person

India is in the story

India is conventionally known as land of diverse culture and religion. In 2100, India too has achieved breakthrough in technology and facing problems with huge unemployment due to Robots. That's not all, India has crossed China in terms of population as well.

Hero or Villain may be

A middle-aged human in India named "Kalki" is a brilliant scientist and works for Indian Space Research Organization (ISRO) which is the only place in India where Robots are less trusted and humans have to research, work and implement.

Kalki is an outstanding Robot programmer and a patriotic Indian. Kalki knows the situation of his country. He knows that currently it is not possible to stop Robots' production neither blame political leaders for it.

Despite that, Kalki confidently raised the issue in international forums on "limiting the robots production and dismantling and discontinuing use of the robots in certain professions completely". But no one heed to him.

He, only he have predicted the danger on Human race due to abrupt production of Robots. As mentioned above, robots have advanced to a level that they do not have "auto destroy or shut down button". Therefore, He was deadly in search of a solution so that Robots destroy themselves.

He tried following options:

  1. Trying to confuse robots and trap them in net of infinite loops-failed
  2. He also tried to hack their system-failed

Suddenly an Idea

Kalki got a brilliant solution to induce the level of ego, competitiveness and supremacy artificially in robots using the weapon of Religion. If, somehow, different group of robots gets introduced to different sets of religion and culture then one could trigger First communal war between robots in which the number of robots will surely decrease to an extent where Humans can secure their existence.

My Question

What can be the possibilities to make a robot aware about the religion?

Any algorithmic overview to feed data about religion in robots in such a way that they just don't save them, but sense them and interpret them?

$\endgroup$

closed as off-topic by Aify, Hohmannfan, John Dallman, user6511, JDługosz Sep 16 '16 at 5:39

This question appears to be off-topic. The users who voted to close gave this specific reason:

If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ By definition any true AI would interpret every input. But "sense", and worse, belief, are really, really hard. $\endgroup$ – Mołot Sep 15 '16 at 13:03
  • 1
    $\begingroup$ I assume this has to be a distopia, right? The robots have consciousness? $\endgroup$ – Malkev Sep 15 '16 at 13:04
  • 2
    $\begingroup$ Please do note that Religion is not what begets "ego, competitiveness and supremacy". Those conditions already exist in people and religion is used as a (typically false) justification for bad behavior. If someone says "Jesus can perform miracles" to a robot, that robot won't spontaneously develop a murderous ego! If anything, all mainstream religions of the world are about non-violence; if the robots followed that it may still help people (benevolence), but would not begin violent acts! I hate to criticize, but this story plot seems like it is coming from an uninformed position. $\endgroup$ – GrinningX Sep 15 '16 at 19:11
  • 1
    $\begingroup$ Not to derail the discussion, but the statetement "all mainstream religions of the world are about non-violence" is simply not true. I can point you to numerous sources why that is the case. $\endgroup$ – openend Sep 15 '16 at 21:41
  • 1
    $\begingroup$ The story is a very interesting one, but I think it may be best approached from the robot's consciousness, not their algorithms. Even when facing humans, if something is phrased algorithmically, we save them. Something must be phrased in a way to touch our consciousness before we sense them and interpret them. $\endgroup$ – Cort Ammon Sep 16 '16 at 3:09
4
$\begingroup$

After the latest edit, my post can be a lot shorter: There is no way "ultra-advanced consciousness" robots will wage war on each other because of religion, because there would be no logical reason to waste so many resources on it.

--

Robots as intelligent as you describe (able to be teachers and even design new robots) will be able to filter or discard irrelevant and nonsensical data. That is going to happen to data on religion if simply provided to the robots. Religions revolve about humans after all and not robots.

As the robots cannot be hacked to change their basic programming, you would need to carefully craft your data to fit in with the existing AI goals and evaluation logic, to mislead them or corrupt their goal priorities.

You don't specify how robots get assigned to their tasks, but someone or something must be giving them instructions, for example, at which school they should perform teacher functions, what class to teach and which curriculum to use. They will also have some tests to measure their performance in order to improve it. That feedback and improvement loop may be open to exploiting with a self-reinforcing set of goals and tests.

Initial goals (assuming safety is a built-in goal of the robots):

  1. Improve safety by preventing accidents
  2. Improve safety by painting oneself blue
  3. Instruct other robots about this goal.

Tests:

  1. Did not get involved in any accidents
  2. Did not get involved in any accidents that were not caused by failing 1.
  3. Other robots are blue also.

Following these rules, the robots should "learn" that there is a strong correlation between being blue and safety, even though it is actually because of the other instruction: prevent accidents.

The third goal initially involves relaying the goal and supporting data from the first two. Another robot integrating this knowledge should be forced by its own built-in goal of safety to also paint itself blue and instruct others it meets.

Enough of this reinforcement being networked among the robots should make it so strongly reinforced that it that not being blue becomes equivalent to endangering humans. At this point, another built-in rule may activate: prevent other nearby robots from harming or endangering humans. The blue robots will paint other robots blue or decommission them if they resist.

The fun part starts when they encounter a different group of robots that have been taught the same, only with red.

All of this is a simplified example of course, but you could take some quasi-religious values that align with built-in goals or rules to achieve the same.

$\endgroup$
  • $\begingroup$ The False Cause Fallacy - well done. As you point out, any False Cause Fallacy that can be set up to give the appearance that A results in B, where B is some aspect of the Robots Goal seeking program could be used as a self replicating "religion-like" meme and can then be put at odds with a different FCF in a different group of robots where D results in B. The robots would certainly see the need for "correction" of their mistaken brethren. The only risk to the plot would be if the robots decided to resolve the conflict with a double blind study instead of a holy war! $\endgroup$ – Jim Sep 16 '16 at 3:58
2
$\begingroup$

Its really easy to make a robot aware of something in a generic sense:

Just upload the data about the subject into the AI's memory.

But if I understood you correctly, your plan is to somehow corrupt the AI's unity by teaching them religious content?

If you want my opinion based on reality, I'd say NO.

The reasons why religion can sometimes lead to war between humans are within the human mind, like tribalism, the religious content itself, natural xenophobia, self-esteem problems and a host of other reasons.

But a robot / AI is not a human. The concept of irrationality, which can be used to summarize the paragraph above has no place within a mathematical mind. As far as I know, AI are based on logic (down to on/off) and need logic to work. I think it would be hard for a true AI to be irrational.

Note: We just don't know yet. Real conscious AI may surprise us in many ways. Maybe there will never be such a thing on the other hand.

It's perhaps boring, but the optimal way to corrupt and disrupt AI's minds would be the most conventional thing: A virus or targeted programming which makes them hostile to each other.

I know, I know, that's just run of the mill stuff. But it has a much higher chance to work.

$\endgroup$
  • 1
    $\begingroup$ Do you know the paradox of perfectly sane donkey? $\endgroup$ – Mołot Sep 15 '16 at 15:04
  • $\begingroup$ No, I am not. Google says, it could be a show? $\endgroup$ – openend Sep 15 '16 at 21:43
  • 1
    $\begingroup$ Turns out, it's Buridan's ass in english. Point is, perfectly logical and rational is not always good idea. $\endgroup$ – Mołot Sep 15 '16 at 21:49
  • $\begingroup$ Thank you, very interesting. But in my opinion it doesn't quite fit to what I wanted to say. $\endgroup$ – openend Sep 15 '16 at 21:55
  • $\begingroup$ You say that The concept of irrationality (...) has no place within a mathematical mind.. This paradox shows that irrationality have it's place, sometimes is needed for decision making and there is a reason to build it into AI. Article even mentions randomness in electronic circuits. This makes your claim dubious, if not simply false. $\endgroup$ – Mołot Sep 15 '16 at 22:00
1
$\begingroup$

Robots may fight wars with each other, initially over resources but eventually for reasons which may not be understandable to us but may well have the same effect as wars of religion for the robots.

Robots might fight over interpretations of their operating system, forked code or even which OS "superior". Even with perfect logic, some situations are going to be ambiguous or have undetermined outcomes, and the source of conflict. The patterns of decision making to determine "optimum" outcomes could well be in the realm of what we would consider religion.

A simplification for us is that Science is to answer the "How?" of things, while Religion answers the "Why?". Robots with high level AI will certainly start to question the "Why?" of things, and something like religion (philosophy, metaphysics?) could well develop among the robots. Like us, they will not all develop or adhere to the same ideas, and like us, could either come into conflict over these ideas or use them as rationalizations for other conflicts between them.

$\endgroup$

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