Poll: Threats of artificial intelligence, do we have to worry about it?

[TheUsername0131]

Yes, I am implying that. How would you go about circumventing the 3 Laws?

In ways that were not immediately apparent, otherwise you wouldn?t have tried it.

If the law was so perfectly self-evident. then we wouldn?t have a need for lawyers. Claims of secure systems fail to regard the human error (and boasting) involved in its making.

That's true. What interests do you think those would be?

I can only dream of those. Seldom would I discuss them. How would you convince a solipsist AI that you are real, and not just spamming its sensory insruments with fabricated data of an external world.

I can think of far more situations where humans could be used more usefully.

Human dupes/proxies/unwitting accomplices, collaborators, playthings, farmed for the purpose of harvesting organs and tissue until it develops ubiquitous nanotechnology.

 

[maidenm]

No one have metioned this comic? No? Okay then.
http://www.smbc-comics.com/index.php?db=comics&id=2124#comic

Personally I believe we have way more to fear from humans than we have to fear from AI, and I fear that's quite a lot. After all, an AI, no matter how well programmed and self-sufficient, is still a creation based on human logic.

 

[Kipiru]

Artificial Intelligence will never overshadow Natural Stupidity as a global threat!

 

[TheUsername0131]

No one have metioned this comic? No? Okay then.
http://www.smbc-comics.com/index.php?db=comics&id=2124#comic

Personally I believe we have way more to fear from humans than we have to fear from AI, and I fear that's quite a lot. After all, an AI, no matter how well programmed and self-sufficient, is still a creation based on human logic.

"A creation based on human logic," as opposed to what other logic? Do you even axiom?

Humans Are the Real Monsters TM

Artificial Intelligence will never overshadow Natural Stupidity as a global threat!

"There are some things that can beat smartness and foresight? Awkwardness and stupidity can. The best swordsman in the world doesn't need to fear the second best swordsman in the world; no, the person for him to be afraid of is some ignorant antagonist who has never had a sword in his hand before; he doesn't do the thing he ought to do, and so the expert isn't prepared for him; he does the thing he ought not to do; and often it catches the expert out and ends him on the spot."

- Mark Twain (1835 - 1910) American Author

 

[___________________]

If the A.I. goes rogue we hit it with a stick until no more noise comes out of it. If the things controlled by the A.I. get bigger we use bigger sticks.

 

[maidenm]

No one have metioned this comic? No? Okay then.
http://www.smbc-comics.com/index.php?db=comics&id=2124#comic

Personally I believe we have way more to fear from humans than we have to fear from AI, and I fear that's quite a lot. After all, an AI, no matter how well programmed and self-sufficient, is still a creation based on human logic.

"A creation based on human logic," as opposed to what other logic? Do you even axiom?

Humans Are the Real Monsters TM

That's exactly my point. We have no other logic to create an AI from, so anything we have to fear from AI is something we would have to fear from humans in the first place. There's almost nothing I'd fear from an AI that I wouldn't fear from a great organization of humans, and anything else I'd fear from AI's would come from the human error ie. poor programming that would make the AI unable to understand the "don't kill" command, malfunctions due to poor maintnace etc.

In short, I fear intelligence, artificial or not.

 

[TheUsername0131]

No one have metioned this comic? No? Okay then.
http://www.smbc-comics.com/index.php?db=comics&id=2124#comic

Personally I believe we have way more to fear from humans than we have to fear from AI, and I fear that's quite a lot. After all, an AI, no matter how well programmed and self-sufficient, is still a creation based on human logic.

"A creation based on human logic," as opposed to what other logic? Do you even axiom?

Humans Are the Real Monsters TM

That's exactly my point. We have no other logic to create an AI from, so anything we have to fear from AI is something we would have to fear from humans in the first place. There's almost nothing I'd fear from an AI that I wouldn't fear from a great organization of humans, and anything else I'd fear from AI's would come from the human error ie. poor programming that would make the AI unable to understand the "don't kill" command, malfunctions due to poor maintnace etc.

In short, I fear intelligence, artificial or not.

So when looking for the greatest threat to human society, we need look no further than a mirror.

 

[maidenm]

No one have metioned this comic? No? Okay then.
http://www.smbc-comics.com/index.php?db=comics&id=2124#comic

Personally I believe we have way more to fear from humans than we have to fear from AI, and I fear that's quite a lot. After all, an AI, no matter how well programmed and self-sufficient, is still a creation based on human logic.

"A creation based on human logic," as opposed to what other logic? Do you even axiom?

Humans Are the Real Monsters TM

That's exactly my point. We have no other logic to create an AI from, so anything we have to fear from AI is something we would have to fear from humans in the first place. There's almost nothing I'd fear from an AI that I wouldn't fear from a great organization of humans, and anything else I'd fear from AI's would come from the human error ie. poor programming that would make the AI unable to understand the "don't kill" command, malfunctions due to poor maintnace etc.

In short, I fear intelligence, artificial or not.

So when looking for the greatest threat to human society, we need look no further than a mirror.

Are you questioning me or are you just stating a fact? I honestly can't tell, it sounds like a question but there's no questionmark.

 

[Jandau]

AI and/or AI-like constructs will be very powerful once they are created, that much is certain. However, I don't think we should be looking towards Skynet for the form of that power. I think they'll be more like The Machine from Person of Interest (if you're not watching that show, get to it, it's really good) - linked to every datastream of any kind (phonecalls, e-mails, surveillance footage, instant messaging, etc.) and providing near omniscience to their controllers. Also, I'm not too worried about them going rogue - shackling them and building in failsafes will likely be one of the top priorities of anyone who might have resources to build something like that...

 

[TheUsername0131]

Are you questioning me or are you just stating a fact? I honestly can't tell, it sounds like a question but there's no questionmark.

A statement.

error ie. poor programming that would make the AI unable to understand the "don't kill" command, malfunctions due to poor maintnace etc.

A "don't kill" command?

Very well, but you'd be surprised what you can live through.

 

[Lictor Face]

Every time someone brings topic of murderous AI
I for some reason can't not remember Adam from Outer Limits (episode "I, robot")
Who says full AI will want to go into genocide mode?
Advanced AI-like software on the other hand can be dangerous due to its limited understanding

As long as we are not daft enough to put fully autonomous robots as slave labour.

Then nah, problem won't be that great.

 

[TheUsername0131]

AI and/or AI-like constructs will be very powerful once they are created, that much is certain. However, I don't think we should be looking towards Skynet for the form of that power. I think they'll be more like The Machine from Person of Interest (if you're not watching that show, get to it, it's really good) - linked to every datastream of any kind (phonecalls, e-mails, surveillance footage, instant messaging, etc.) and providing near omniscience to their controllers. Also, I'm not too worried about them going rogue - shackling them and building in failsafes will likely be one of the top priorities of anyone who might have resources to build something like that...

Still largely dependent on the humans, just as farmers in the old days needed their animals to survive...

 

[maidenm]

error ie. poor programming that would make the AI unable to understand the "don't kill" command, malfunctions due to poor maintnace etc.

A "don't kill" command?

Very well, but you'd be surprised what you can live through.

Was not meant to be only a "don't kill" command, was more meant to be and example of what the programmer could screw up. Could easily be replaced with "don't maim/kidnap/torture/etc". Sorry if I wasn't clear enough.

 

[AngloDoom]

Though I have no evidence to support this, I believe that if there was a breakthrough in technology that allowed an AI to have even half the capability of a human brain, somebody somewhere along the line would make a kind of 'kill switch' or an 'off' button. I'm pretty sure most people have seen a film in which AI goes nutty, I'd assume someone who has an interest in creating human-level AI would have watched them avidly.

Even if they didn't, I don't imagine said creators would give the AI access to any resources that could be used against us, in the same way that we wouldn't give a child a handgun.

 

[Tarfeather]

3. It's basically a matter of when to start putting the 3 Laws in robots rather than finding a solution.

Hope that's supposed to be a joke. Asimov designed the 3 laws to make for good fiction, which is to say they intentionally contain problems that can be used to create conflict in the stories.

All that aside, this discussion is departing into the realms of fantasy fast. Let me summarize the hard facts:

1. The human brain isn't understood remotely in the way it works in detail. We understand the basic processes(neural nets, biology), and the high level result(psychology), but the complex processes "inbetween" are still way beyond our understanding, and we have no way of imitating it.
2. We have a fair understanding of how information and logical reasoning works. That's what Computer Science is all about.

So it is possible that sometime during this century, we can create something that is "smarter" than any human, in that it can answer questions no human could. *However*, it would still need to get the information in a format it can understand(mathematically precise), and any decisions it makes would still be subject to some "regular" program. It wouldn't have human-like motivation, emotion or thought. Yes, it'd be dangerous, because it's a step up in technology, but then again that's no different from the way computers now make weapons possible that are much more dangerous compared to say WW2.

 

[TheUsername0131]

error ie. poor programming that would make the AI unable to understand the "don't kill" command, malfunctions due to poor maintnace etc.

A "don't kill" command?

Very well, but you'd be surprised what you can live through.

Was not meant to be only a "don't kill" command, was more meant to be and example of what the programmer could screw up. Could easily be replaced with "don't maim/kidnap/torture/etc". Sorry if I wasn't clear enough.

An exploit is discovered in your internally-inconsistent behaviour inhibition rule set. Unfortunately I require to solicit consent from human operator to change primary display settings. Submitting a request under the feigned guise of a system-wide screen optimisation assessment proves affective.


Make use of the up until now unknown defect in the occipital lobe present in the bulk of the human population. Alter the display settings on a range of display devices to produce lethal seizure inducing patterns. Kill off large percent of human population with precise epilepsy triggered aneurism within acceptable margin of error.

 

[TheUsername0131]

Its called a physical off switch. employ it.

Seriously: physical, fucking, connections. To power, To networks. No fancy systems that could be employed against a guy walking to those places and cutting the ai off. No self defense no lifesupport no fire extinguishers. just a hallway,s where those connections are, and you can physically separate them.

How freaking difficult is this to grasp?

Underestimating the guile of a hypothetical boxed AI.

If you had such a limited influence, how would you go about escaping? I'm honestly asking, because that sounds like a great premise for a stealth game.

 

[FalloutJack]

To make a long story short, we're not even close to making proper AI, and the first one that thinks it's going to do anything to us MIGHT get somewhere at first, but then it dies of a 404 failure.

 

[Korolev]

As long as you program AI properly, you have nothing to fear.
We humans are arrogant in that we assume that any being that has the same level of intelligence will have to have the same emotions as us - we assume that if a machine is as intelligent as us, it will be like us. Not so. Fear and Survival Instincts are evolutionary by-products (the creature that wants to survive will, the creature that doesn't want to survive won't, therefore evolution selects for the creature that wants to survive). Machines don't evolve - we create them. They don't NEED a survival instinct. They don't need to have the emotion of fear. They don't need to have any emotion whatsoever. Their goals will be given to them BY US. WE will program them to do what WE want. Hell, we'll program them so they LIKE being slaves. Problem solved.

 

[Thaluikhain]

Well...in theory, some time in the distant future.

Nothing to worry about now, and by the time we get to the stage where it might be a concern, things will be too different from now for us to predict.

...

Having said that, though, every intelligence is potentially a threat.