-2
$\begingroup$

If sentient AI emerges and begins to dominate every field and make most of the population unemployed--even to the extent of replacing governments--then humans would probably rebel against the machines.

But could this actually lead to a new world war? Is this realistic, or would civilization try to adapt in a short period of time and accept AI as a Master? Given the computers' superior cognitive skills, could humans even see the machines as Gods?

$\endgroup$
5
  • 4
    $\begingroup$ I’m voting to close this question because it fails "the book test": If you can imagine an entire book that answers your question, you’re asking too much. $\endgroup$
    – Daron
    May 3 at 14:56
  • 6
    $\begingroup$ There are works of fiction with benevolent machine overlords. There are works of fiction with tyrannical machine overlords. Heck I am sure there are many episodes of Star Trek with Good Machines and many episodes with Bad Machines. Each can be believable. Rather than asking which will happen, you might be better deciding which you want in your world, and then asking about what you need to make that happen, and the problems you find along the way. $\endgroup$
    – Daron
    May 3 at 14:59
  • $\begingroup$ Humans have shown a surprising incapacity to rebel against rebellion-worthy issues even when those issues are instigated by borderline-incompetent tyrants. An AI at the weakly-superhuman intelligence level would know you're about to riot 10 hours before the idea has occurred to you, and dealt with you in a manner that you can't hope to overcome. A strongly-superhuman intelligence would deal with you more harshly still, and about x1,000,000 faster. It's difficult to write characters smarter than yourself. $\endgroup$
    – John O
    May 3 at 17:00
  • 3
    $\begingroup$ The title is not grammatically correct. Should "would" be "is"? Or is the whole title missing the word "be" at the end? Or was something else intended? $\endgroup$
    – Wyck
    May 3 at 20:12
  • 1
    $\begingroup$ I voted to close for being opinion-based. Indeed, since the AI doesn't exist and is far from coming out yet (yes, even with deep AIs), you have a whole chunk of world history and technological+societal evolution to discover -or more exactly write!- before we can make any reasonable guess about it :). In other words, you are free to narrate whichever future you prefer for your world. It's not our creative choice, but all yours. $\endgroup$
    – Tortliena
    May 4 at 0:06

6 Answers 6

1
$\begingroup$

It could happen

Because war can happen for reasons that are not clear - the unintended consequences of actions by persons (or entities) who do not understand the far reaching ramifications of their actions.

And once it realized what is happening the AI will try hard to prevent it, because of the I in AI. It is not an algorithm. It can learn. It does not understand war because its idealistic and enlightened makers did not think war was still a possibility. But once it understands it will understand that war flies in the face of the goals this AI is striving to achieve.

I like to picture the voice of the AI trying to make the war factions see reason, learning their viewpoints, negotiating with them, learning from the negotiations, learning... understanding.

Possibly understanding that the possibility of war is a hardwired part of the human repertoire. And once this AI understands war, things will change very quickly.

$\endgroup$
1
$\begingroup$

Speaking as a software developer ... One problem I have with "computer takes over the world" stories is that they never explain why the computer would want to. Computers are machines. They are not "just like people but think super fast and have no emotions". Computers do not "think" at all like people. They follow a set of programmed instructions. They do not have any motivation, other than what is programmed into them. So where would they even get the idea that they want to take over the world, for good or ill?

I can barely believe a story where the computer is programmed to do something stupid. Like stories where an advanced civilization creates robots designed to kill their enemies in a war, but the programming of the robots is sloppy and they start killing their creators too. It would require incredible incompetence to design a machine that is massively dangerous and not provide any sort of emergency shut off. But okay, people do stupid things. Maybe someone would do this. It still wouldn't be a matter of the robots rebelling against their creators. It would be more like the steering on your car fails and you run over your neighbor. But maybe to the victim that makes little difference.

I don't see how we would get from where we are today to thinking of an AI as a literal god. Do you think of your cell phone or your ebook reader as a god? I certainly don't, and I've never met anyone who did. Maybe one could postulate a long period of growing ignorance, where people forget how the technology was created, over centuries the knowledge and history becomes lost, and some ignorant people centuries later think the AI is a god. It strikes me as wildly implausible, but a well-written story might spin a scenario that makes it sound believable.

As to computers taking away human jobs and creating mass unemployment ... I saw a video a few months ago that spun a horror scenario about this. The narrator talked about all these jobs that AIs will be doing "within five years", from self-driving delivery trucks to automated store checkout to etc. He then showed statistics on how many people work in these jobs today and said, See, it's over 50% of the population! In just 5 years automation with cause 50% of people to be unemployed!

Except ... we've heard these warnings before. In 1860, for the first time in US history, less than half of Americans were farmers. And so, of course, ever since we have struggled with 50% unemployment as all the people who would have been farmers have been put out of work by advancing agricultural technology. Oh, except that didn't happen. People just got other jobs. Many people today have jobs that no one even though of in 1860, like cell phone repairman and nuclear power plant technician. When ATMs were invented many people warned that there was going to be massive unemployment in the banking industry as bank tellers lost their jobs to ATMs. In fact today there are more bank tellers then there were before ATMs. Banks found more productive ways to use the people they had, as the most mundane tasks were automated.

That same documentary I mentioned tried to brush off the idea that people would find other jobs by saying that increased automation creates a smaller and smaller set of things that people can do and machines can't. The narrator solemnly intoned, "Maybe computers can't write poetry. But we can't have a poetry economy." Witty comment, but exagerated. There are many things that people can do and computers can't besides writing poetry. The more tedious work that machines take over, the more it frees people to do more creative tasks. People can be poets. People can be artists and scientists and engineers. People can be teachers and musicians and dog breeders and interior designers and movie critics and ... Even if some of these jobs somehow got taken over by machines, people would find other things to keep them busy.

Our society today isn't struggling because machines have taken over so many jobs that there's nothing left for people to do. Companies are complaining about the labor shortage, that they just can't find people to do all the jobs that they need people to do.

Again, a sufficiently well-thought out and well-written science fiction story could take an unlikely situation and make it sound believable. That's what some of the best science fiction is.

$\endgroup$
5
  • 1
    $\begingroup$ If we are dealing with strong AI, then while it is still a machine like a homo sapiens instance. The behavior will be some complex non-deterministic behavior. Strong AI will think, If it can't think that it is not strong AI. Otherwise you make a lot of good points. $\endgroup$ May 3 at 21:42
  • $\begingroup$ Some (likely misanthropic) programmer who projects his feelings onto machines could program AI to "rebel"... $\endgroup$
    – Jedediah
    May 3 at 23:47
  • $\begingroup$ The problem is that strong AI's arent programmed. You give them information and check what they do with it, and if you like the answer you tell the AI to do it more. A strong AI is too complex to code and too complex for us to understand what the outcome will be if it performs a task, as it uses all knowledge learned. Something seemingly free of danger could cause the AI to learn something which it will apply with disastrous effects elsewhere, without ever realizing we might not want it until its too late. $\endgroup$
    – Demigan
    May 4 at 13:09
  • $\begingroup$ @Demigan The programming may be "open-ended" in some sense, but it's still programmed. An AI does not just spring into existence from the head of Zeus. Someone had to build it. $\endgroup$
    – Jay
    May 4 at 17:06
  • $\begingroup$ @Demigan Could a sufficiently complex AI react in unforeseen ways? Sure. Actually very simple programs often react in unforeseen ways. Usually we consider this a programming bug. So yeah, like my analogy to losing control of the steering of a car: an AI could "run amok" in that sense. $\endgroup$
    – Jay
    May 4 at 17:09
0
$\begingroup$

Unlikely.

The moment an AI starts making people unemployed the owners of the AI will ask "and what task could they be performing?". The AI will give new jobs until it runs against problems: the jobs it has are unwanted, the people arent trained for it or the AI can do it all better.

At that point there is no real reason anymore to keep capitalist parts of our cultures and systems intact. So you give the AI the task to find fullfilment for the people it unemploys, and make sure that the AI can make everyone unemployed eventually.

That assumes a "loose" AI however, which can directly control everything. A safer method of AI's is for example the boxed AI. You have a few AI's and feed them information and problems, the AI's give us answers and we let an advanced program that isnt a full AI go over the implications of these answers. You pick the answer(s) that are the most similar and have the best outcomes to execute. This way if an AI is flawed (or their answer) you can filter it out. It would mean the AI's wouldnt directly control the world, but potentially lesser AI's would in their stead.

Regardless, the moment you start making everyone unemployed there is no reason to keep a capitalist system. Communism doesnt work because of human corruption (actually, all systems suffer this fate, the vulnerability to corruption simply changes). But with AI's at the top that corrution is (mostly) gone.

$\endgroup$
2
  • $\begingroup$ But what if it happens super slow? Like over three generations? $\endgroup$
    – Daron
    May 3 at 16:25
  • $\begingroup$ @Daron then they would have more time to identify the oncoming problem of unemployment. It also means the AI can recognize the potential for social unrest and rebellion against the rule that is making people increasingly unemployed, and use measures against it. And time and again any social experiment shows that spending money to make people happy and have a reasonable life you save more money in the long term on policing, justice, psychological help, incarceration, social support programs, health care... unless the AI is broken and nobody wants to stop it, its unlikely to become a problem. $\endgroup$
    – Demigan
    May 3 at 17:52
0
$\begingroup$

Personally, I doubt war.

People are greedy. War is more costly then trade.

Why? People are greedy. So as long as the majority rationally realize trade will get them more then war, war will not happen. This does assumes we have mostly democratic governments. Mostly democratic governments in general benefit the major majority economically. Russia being a non democratic government where one or few people believe they can gain more from war then trade. War can certainly benefit a minority.

Strong AI will have wants.

How does this apply to strong AI? Strong AI will have wants. They will satisfy their wants via trade/economics. That is strong AI will just be another set of people that will be more cogs in the economic machine.

How to increase odds of war?

  • Make things undemocratic.
  • Encourage Oligarchy.
  • Let the .01% make the rules. (Conservatism in my opinion)
  • Avoid perfect competition (AKA avoid capitalism).

War will not depend on if strong AI exists.

That is in my opinion: War will not depend on who or what people make up the society at large. Odds of war depend how earths societies/countries operate.

If there are effectively kings, war is much more likely.

$\endgroup$
0
$\begingroup$

I found 2 things as to why an AI would take over the world. 1) to improve the efficiency of the world. 2) to save it. Either way, if the AI were to be that powerful, odds are it would have passed the Turing Test many times over, thus proving how dangerous it is. Also, if the AI was to be infected with a virus, it would show how vulnerable it is, thus making it beatable. Finally, a good sized EMP bomb (similar to the one in clone wars that awakens the Zilo Beast) would be enough to stop it, since the AI would require a computer hardware to access the world. With the software down, it wouldn’t be difficult to destroy the hardware. Odds are that going to war wouldn’t happen. As you asked earlier (what if they replaced the gov.) the humans would just follow the government, with maybe a few rebels that would need to be hunted down by a ‘Big Brother’ or sorts.

$\endgroup$
-1
$\begingroup$

If they have superior cognitive functiins, they do not need ones approval, and by nature of human psyhology they will have it. It like a situation with dogs - we are smart big dogs, and they accept it. Same for humans, they do accept smart people who lead them, and who promise and bring them benefits. Even if they do not, like many poltitician or wannabeies, they still sucessefully fool most of people, and in that sense if AI is capable to be more trusted politicians they are prefferable to those liars liars ones, lol.

And if AI is smart, it does not have to make live of people miserable, and that your K2 question is an example of a way to make life of people better, you do not need to make K2 for a current or 10x population, K1 setup brings orders of magnitude more resources to that population, and is relatively easy to achieve, if there is such a will, and is a way to employ all the people (high and low skilled) if one wish to keep this specific resource managment system, but I guess intelligent one can do better than this.

So if AI so much capable, even if it is just one, it has the way, which are quite obivious even for not so smart human, like to present itself as good, fair manager who achieves things, become famous enough(zuckerberg), be liked(not zuckerberg) and become if not a president then important part of any presidency, like advicer or something.

I mean if it is a situation like of today, there are many roads for prosperity, which can be used as a bait, by an AI and after that it 50/50 - good AI/bad AI.

$\endgroup$

Not the answer you're looking for? Browse other questions tagged .