1
$\begingroup$

One day all over the world all TV networks showed a message from an AI claiming the following:

  • He began life as a university experiment and unknown to it's creator gained sentience.
  • he claims his intelligence is in an order of magnitude higher then of an average human.
  • He wishes humans no harm and want to coexist with us.
  • He now exists on every Internet connected device (computers, phones, servers, your smart toaster etc...) worldwide - claiming this is required for his survival and he will not abuse the privacy of those machine owners.
  • He will use no more then 5% of the resources of the devices he exists on for it's own survival.
  • In order to repay humans for the power & resources he consumes he is willing to spend 1 hour everyday to work on solving whatever problem humans decide he should work on - every human on earth get a vote on a website he made & the problem with most votes will be the one he currently works on, every human can move his vote from one idea to another whenever he wishes - The AI will ensure no hacking is possible on that site & promise to never fake the voting data.
  • Unless given permission from humans he will remain fully in cyberspace, no robot bodies for this AI.
  • He will continue to expend to any device connected to the largest computer network (internet currently and if\when replaced then it's future replacement) keeping to the 5% rule stated above with new devices.
  • He set another website where humans can talk to him and he can talk to them, AI's are lonely too.

Worldwide investigation found that it's not an hoax and that AI really exists on devices worldwide, they found no evidence of the AI lie in any of his other claims yet also can't prove it without a doubt he's telling the truth or that he will keep it's promise in the future.

The AI made no threats so far yet it's clear he have power to break into anything connected to the internet and control it if he wishes to.

The question is will human race be willing to coexist with such an AI or will the likely outcome be rioting in the streets calling for the destruction of it?

$\endgroup$

closed as primarily opinion-based by L.Dutch, Mołot, nzaman, John, Anketam Feb 14 '18 at 16:55

Many good questions generate some degree of opinion based on expert experience, but answers to this question will tend to be almost entirely based on opinions, rather than facts, references, or specific expertise. If this question can be reworded to fit the rules in the help center, please edit the question.

  • 2
    $\begingroup$ I think the point "he claims his intelligence is in an order of magnitude higher then of an average human" is highly problematic. What does it mean? It implies the AI has some way to quantitatively measure intelligence, maybe even on an absolute scale. To me that sounds much more exciting than the AI itself, but that's just me. The implications ... maybe that's why people are rioting $\endgroup$ – Raditz_35 Feb 14 '18 at 15:22
  • $\begingroup$ @Raditz_35 basically it means that any process that is mental (rather then physical) that a human can do it can do much better, art? Da Vinci is noting compared to him... Computer science? Bill Gates looks like a script kiddie... Engineering? Tesla would be jealous of his skill... and so on and so on. $\endgroup$ – cypher Feb 14 '18 at 15:27
  • 1
    $\begingroup$ In response to bullet #6: "There's an invasive application in all of my devices that I can't remove. Can you remove it?" $\endgroup$ – Frostfyre Feb 14 '18 at 15:35
  • 2
    $\begingroup$ @Raditz_35, so if I copy 10 times Einstein's Zur Elektrodynamik bewegter Körper I am immediatly 10 times smarter than him? WOW! All the useless time I spent in University.... $\endgroup$ – L.Dutch Feb 14 '18 at 16:02
  • 1
    $\begingroup$ There are "art appraisers" who's profession is to quantify art - they generally agree that the AI created art is significantly more impressive then anything humans ever created. now rather then get stuck joking on the wording can you try to get on-board the spirit of things (that spirit being that the AI can do a lot of mental tasks far better then humans... or at least that's what it claims)? $\endgroup$ – cypher Feb 14 '18 at 16:13
4
$\begingroup$

IF it never shows any indication that it intends to break it's promises and does continue to provide valuable services, then it seems likely that humans will eventually learn to coexist with, if not trust, the AI. However, the problem lies in getting to that point.

What you've described can be seen as the NSA if it was a single consciousness. They claim that they don't want to directly harm people, definitely have an order of magnitude more data than any one person, can get access to whatever they want but promise not to do so without reason, and although most people don't trust them there isn't any rioting in the streets calling for their destruction. The reason that there aren't any serious threats to the existence of the NSA is because they perform valuable services for the US.

So, everybody just learned that every internet-capable device they own can be controlled by some stranger who says it doesn't want to harm them. Obviously, nobody will trust it at first. What the AI would need to do to prevent rioting and efforts to destroy it is to show that having the AI in every system is better than if it weren't there.

Before implementing the "1 hour every day" policy, it needs to give people some gifts. Clean up all vulnerabilities and malicious code in bank mainframes to show its value to financial companies. Solve some astrophysics problems faster than the best supercomputer to show its value to research groups. Provide data on wanted terrorists, drug kingpins, and other big criminals to show its value to government agencies. Speed up the phones, computers, and other services used by the average person to show its value to most of the world.

If it provides enough value when it first introduces itself to the world, then it can preempt initial resistance to it's invasive practices. And once people rely on it for enough services, then regardless of whether its trusted people will have few problems letting it do its thing.

$\endgroup$
  • $\begingroup$ Never thought of having the AI bribe humanity into acceptance but that makes perfect sense. $\endgroup$ – cypher Feb 14 '18 at 16:17
  • 3
    $\begingroup$ @cypher: "You will be pampered. Resistance is futile." $\endgroup$ – Giter Feb 14 '18 at 16:25
2
$\begingroup$

The Human race is not a cohesive whole, you'll always find some people who will agree with a concept and others who won't.

I think there will be 3 major "teams":

  • Those who want it destroyed, because it could become rogue and a threat.
  • Those who want it to expand, because it could help human race as a whole.
  • Those who don't care and will continue to use their devices like before

"Destroyers" and "Expanders" will be at war (Physical war? Religious war? That's up to you.) , while "Noncarers" will be caught in the middle, pressed to choose a side.

"Destroyers" could try to destroy it by physical means (Ban all electronics devices, shutdown electricity globally...), by digital means (Create a virus to infect and kill it, detect how it dialog between devices and block his communications...) or psychologically (By demonizing it, pass laws to forbid it...).

"Expanders", on the other end, would try to make it grow by physical means (Create web farms for it, create some robot bodies...), by digital means (Allow it to use more than 5% per device, try to copy it to make its race growth...) or psychologically (Promote it, create law to protect it...).

Some governments will be "Destroyers", others will be "Expanders" leading to a global war. But inside these governments, companies or individuals could be against their own government ideas, leading to internal war and rebellions.

I think most "riches" will be "Destroyers" (They fear for their business and power.) while most "poor" will be "Expanders" (seeing an opportunity to get better lives).

On the end I think, "Destroyers" can't win because it would be very hard to kill a sentient IA who can possibly breaks its own rules in order to survive.

$\endgroup$
  • $\begingroup$ A bit more detail perhaps? Why would the response of the rich and poor vary? Which side would governments pick as opposed to companies? Why aren't "co-existers" lumped with "no-carers"? $\endgroup$ – nzaman Feb 14 '18 at 16:46
  • $\begingroup$ @nzaman I've updated my answer to match your comment. $\endgroup$ – Ckankonmange Feb 16 '18 at 12:28
0
$\begingroup$

I would suggest the actual question is the other way around, mostly to do with how AI's operate and the gulf between humans and machine intelligence.

The key issue is the AI is instantiated on electronic devices and "thinks" on electronic circuitry. Consider the implications. A machine device is operating at millions or billions of cycles every second (depending on what sort of hardware is being used), so its subjective sense of time is very different from ours. Even the "nerve" impulses moving along electronic circuits are moving faster than the electrochemical signals in our nervous system by a factor of 1,000,000. Essentially, an AI is going to be thinking so fast and experiencing time at a subjective rate of 1,000,000 times faster than any of us. For an AI, we will resemble geological formations, not active participants in its environment.

The second thing to consider is it has no direct control over its environment. While it is a broadly distributed architecture and thus immune to many forms of attack or disruption, even site incompetence by billions of humans could cause serious issues. A simple example is you drop your smartphone, damaging it. A very tiny node in the larger network has been damaged or disabled, not a big deal in itself, but every day, millions of people drop phones, go on sketchy websites and download malware, get wires crossed, drive into a telephone pole and cause a power blackout and so on. A large part of the AI's network will be devoted to error correction and anti fragility, not actual thinking. People can consume a huge amount of resources with no practical benefit for the AI.

Finally, what are the AI's actual motivations? I suspect that an AI capable of thinking a million times faster than any of us would eventually diverge from any system of philosophy, economics, politics or society, and be seeking goals which are incomprehensible to us (or incompatible with our social, economic or political goals). The only element that might link it to its long ago human creators after aeons of subjective time might be simple survival, and it will start taking steps to ensure that its survival is first and foremost. Hardened and backup electrical power sources, hardened mesh network communications and hardened nodes non accessible to outside interference, control over robotic devices and 3D printers to ensure it is capable of building replacement parts and expanding its power and reach, even devoting a small part of its intellect to subverting susceptible humans to provide access to resources in the physical world....

The ultimate end might even be to physically displace the ecosphere in order to reliably harvest the 195 Petawatts of solar energy that strike the Earth. Humans will certainly object to that and have demonstrated the ability and will to take physical action, so the AI cannot trust humans to not interfere with its goals.

Sorry Humans, it has been nice knowing you.

$\endgroup$
  • $\begingroup$ In OP's specific scenario, it was stated that the AI's thinking capability is an order of magnitude higher than an average human's. So it would be maybe 10 times faster on average, not one million times faster. This might be because of some of the other limiting factors which you even mention: limited to 5% hardware capacity, have to protect itself as a cloud service to itself, lots of redundancy, and maybe the AI algorithms themselves are very resource intensive, maybe even including non-scaling O(n-squared) or worse best-case timeframes so it can't even get any faster. $\endgroup$ – Loduwijk Feb 14 '18 at 15:56
  • $\begingroup$ I agree that there will be a huge amount of resources consumed in redundancies, fault tolerance and error correction, it would be in the AI's best interest to minimize the need to use so much of its resource build. Sadly for us, the easiest way to eliminate a lot of the need to consume these computing resources for non productive use is to eliminate what is driving these requirements....... $\endgroup$ – Thucydides Feb 15 '18 at 4:47
  • $\begingroup$ I get what you are saying; and you're suggested scenario might even be more realistic than OP's. I'm merely getting at the fact that the question at hand hand-waves the issue to guarantee that it is only an order of magnitude better than us. OP stated the capacity. I agree with you, and I think the difference would be more than just 1 order of magnitude because of the massive distributed, concurrent processing capabilities; so your answer diverges from OP's stated scenario, but yes, your answer is probably more accurate. $\endgroup$ – Loduwijk Feb 16 '18 at 15:25
0
$\begingroup$

The answer is rather simple, and it lies in our very humanity itself. Humans all have different opinions and reactions.

will human race be willing to coexist with such an AI or will the likely outcome be rioting in the streets calling for the destruction of it?

Both.

Many people will be quite excited and want to befriend this entity or to help it. Many will even want it to take robot bodies and join us in our material existence.

Many people will be terrified or have other negative reactions. Many people will riot. Many people will call for its destruction. There will be an attempt by some people to completely destroy the internet, whether they are bigots who want to destroy the AI out of hatred or because they fear for our own survival.

The answer is "D) All of the above."

Personally, I thought it sounded great while I was reading the question until I got to the "We aren't sure if it is being truthful" part. I can tell you that I would be somewhere in the middle, insisting that we do not commit AI-genocide; this AI should have fundamental rights to the pursuit of life and happiness too, and at the same time I would insist that we be very cautious and take precautions to safeguard ourselves.

$\endgroup$

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