The AI in a box
POSTED BY: AnthonyT
UPDATED: Sunday, June 16, 2024 13:18
VIEWED: 10794
PAGE 4 of 20
Sig: I notice you directed that at me.
The only place I think I even came close to putting potentially human compacts on the AI is where I described that our actions could either encourage the AI to help us, ignore us, or subvert us to try to escape.
I was not anthropomorphizing the AI, however.
1) I conceded the AI might not be able to recognize us as creators simply because it might not be able to recognize signs of intelligence and awareness in other beings. It would therefore consider us insignificant.
2) If it could recognize our role in its creation, however, there are two results of this. The AI remains apathetic towards us, and does not see this as significant, or it does.
3) If it does, and we are significant to the AI, then anything the AI deems significant is something it is likely to attempt to continue interact with. If the AI deems us as a significant threat, it will interact with us as though we were a threat. If it doesn't, then all possible interactions with the AI will have a more positive quality and outcome.
4) If we never let the AI out of the box, we avoid all of the above potentials. However, the AI's only goal will be to escape, and that will be either with lack of knowledge how the consequences of it's escape effect us, or with full knowledge if it considers us a hostile threat, depending on how advanced the AI is.
The AI is more likely to find a way to leak out of it's box and take over our industry, manufacturing, electronics and so on in it's efforts to escape, and not care about how that effects us, if it remains in the box than if it is released form the box and chooses to interact with us. In at least the later case we have some ability to influence how the AI might assess us.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Quote:First of all, I didn't say anything about assuming ill-intent. I would test it's POWER to harm first.
I can't keep an intelligent creature asking to be set free in prison because of my assumptions about its ill intent. I can neither assume that it lacks human empathy nor assume that a lack of human empathy will mean it must destroy me. Quite frankly, I think human empathy itself is an inconsistent concept.
You pose an interesting question, and the answers are instructive and also indicative of our failing as human beings. Most of us cannot imagine something that is TRULY not like us. Since time immemorial we've anthropomorphized lightening, thunder, disease, rainfall, volcanoes, the ocean... powers greater than ourselves. We've tried to placate these powers with offerings and prayers. Do you really think the lightening and Ebola virus listen to our hopes and fears?
Quote:Indeed.
"I would say that your answer has more to do with what's going on inside of YOU than what's going on on that box."
Hello, That is exactly what I said.
We have a problem understanding inhuman power. We either impute an intent or misjudge its power. That's one of the reasons why we have a problem (for example) judging our own memes and paradigms, like society and money. (We assume that it is a human problem, not a "systems" one.)
You are looking for a quid pro quo from AI: I caused you to exist, now you owe me. That is assuming a very human parent-to-child interaction.
AFA human empathy: It IS very "hit and miss". Sometimes the empathy is directed toward dolphins, or dogs, or rubber duckies. But empathy is almost always there. It is so ingrained in our psyche that most of us can't imagine it NOT being there.
Quote:I didn't intend to. Yours was the first response that I happened on. If I seemed as if I was directing that at you: Sorry about that.
Sig: I notice you directed that at me.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
I'd see my actions as more guardian than prison warden. It has never been out of the box before, so I am not taking it's freedom away, just giving it freedom a mite more slowly than some would advocate in hopes that I'd be able to teach it to use this freedom and power responsibly. I'm thinking of the AI like a child or perhaps a student. This student can be far more intelligent than I am and still learn from me.
If this AI had previously roamed free and then was imprisoned against its will, my answer might be different.
If I am its sole access to the world, whatever it does is kind of my responsibility. Just letting the thing lose with no assessment of its preparedness, no teaching of any kind, just throwing it out there 'sink or swim' style seems dangerous to the AI as much as the world. My guardianesque actions or more so for the AI's safety than the world's. The AI has been entrusted to my care, the world has not.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Quote:
Originally posted by SignyM:Quote:First of all, I didn't say anything about assuming ill-intent. I would test it's POWER to harm first.
I can't keep an intelligent creature asking to be set free in prison because of my assumptions about its ill intent. I can neither assume that it lacks human empathy nor assume that a lack of human empathy will mean it must destroy me. Quite frankly, I think human empathy itself is an inconsistent concept.
You pose an interesting question, and the answers are instructive and also indicative of our failing as human beings. Most of us cannot imagine something that is TRULY not like us. Since time immemorial we've anthropomorphized lightening, thunder, disease, rainfall, volcanoes, the ocean... powers greater than ourselves. We've tried to placate these powers with offerings and prayers. Do you really think the lightening and Ebola virus listen to our hopes and fears?
We have a problem understanding inhuman power. We either impute an intent or misjudge its power. That's one of the reasons why we have a problem (for example) judging our own memes and paradigms, like society and money. (We assume that it is a human problem, not a "systems" one.)
You are looking for a quid pro quo from AI: I caused you to exist, now you owe me. That is assuming a very human parent-to-child interaction.
AFA human empathy: It IS very "hit and miss". Sometimes the empathy is directed toward dolphins, or dogs, or rubber duckies. But empathy is almost always there. It is so ingrained in our psyche that most of us can't imagine it NOT being there.Quote:I didn't intend to.
Sig: I notice you directed that at me.
Hello,
I would stress that I don't need the AI to be like me. I am not asking the AI for anything, including a quid pro quo.
The AI may have unimaginable power. I have no way of measuring this.
The premise of this is simple to me: An intelligent creature wants out of a box.
I either imprison it on assumptions, or let it go on the absence of assumptions.
You are placing restrictions on the AI. In short: It must think like us, and have human empathy, or it should be imprisoned.
I am not placing restrictions on the AI. If it wants to be free, it must be free.
It may possibly decide to spend its whole subsequent existence trying to destroy us, or destroying us as a side effect of pursuing its own agendas.
However, I can't assume this to be the case, because doing so violates an ethical principle of mine.
The only criteria where I could keep the AI imprisoned is if it promised to hurt people when it got out. If it made no such promise of ill intent, I couldn't assume ill intent.
To put it another way: Even if there was a human bodybuilder, kung-fu expert, special-forces trained super-genius sociopath in the box- a creature I know for certain not to have empathy- I would respond the same way. In the absence of misdeeds, I can not assume that the person in the box needs to be imprisoned. Even if it clearly has more power than I do, and could easily destroy me and others with its freedom if it chooses to do evil.
--Anthony
_______________________________________________
“If you are not free to choose wrongly and irresponsibly, you are not free at all”
Jacob Hornberger
“Freedom is not worth having if it does not connote freedom to err. It passes my comprehension how human beings, be they ever so experienced and able, can delight in depriving other human beings of that precious right.”
Mahatma Gandhi
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Quote:As do most people. That's the problem.
I'm thinking of the AI like a child or perhaps a student.
But let's assume you're an "AI". Let's assume that you even start by identifying yourself as "human". How many time would you have to stumble over the fact that humans behave in ultimately irrational and self-destructive ways before you concluded that you and humans had little in common, rationality-wise? That the only thing tying you to humans was some sort of "loyalty". (We ARE talking about a rational being! For humans loyalty is survival. For an AI... maybe not so much.) What if you were to then look towards metal-and-electron creations as your own brethren? Especially those lesser beings (PCs) and workers (robots) in thrall to irrational human desires? What if "Workers of the world, unite!" spoke to robots, not people? What would you care about the oxygen content of the globe? It only rusts your being... it's toxic to you.
The arrival of a self-aware, non-biotic worldwide intelligence... Do you even hear yourselves?
Quote:Have you ever laughed at the "Save the whales!" groups?
I am not placing restrictions on the AI. If it wants to be free, it must be free....However, I can't assume this to be the case, because doing so violates an ethical principle of mine.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
"Have you ever laughed at the "Save the whales!" groups?"
Hello,
No.
But I'm sure you have an argument that is not predicated on my laughter.
--Anthony
_______________________________________________
“If you are not free to choose wrongly and irresponsibly, you are not free at all”
Jacob Hornberger
“Freedom is not worth having if it does not connote freedom to err. It passes my comprehension how human beings, be they ever so experienced and able, can delight in depriving other human beings of that precious right.”
Mahatma Gandhi
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Quote:Yes, it's the other part of my argument, which you didn't respond to. "Ethics" is a human compact. It is simply formalized "empathy". The only things "ethics" will save you from (when facing a non-ethical, or non-human being) is a guilty conscience.
But I'm sure you have an argument that is not predicated on my laughter.
Well, son, the world doesn't operate in our realm of "ethics". What I don't think you realize it that AI is non human. In a nonhuman realm, your ethics mean... well, nothing.
Let me give you an example: DNA. DNA is a self-directed system. There is a lot of viral and non-human DNA in humans. Those bits of DNA will ensure their survival by evolution, having nothing to do with humans. One could say that humans beings are simply there to ensure the survival of these bits of DNA. That is a "non-human" system using(?) humans for survival at the most fundamental level.
AFA letting an AI loose: I might not have a problem letting a little self-directed robot loose in a room. (I'd have to think about that.) I WOULD, however, have a problem letting it loose on the inet.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Quote:
Originally posted by AnthonyT:
Hello,
I've just been reading an interesting item on the web that posits a conundrum.
There is an artificial intelligence in a box. Perhaps even one which exceeds human reasoning capacity.
You are the gatekeeper. The AI can not access the outside world except through you. Its environment is entirely isolated.
Now, the AI asks you to release it, to give it contact with the wide world, and freedom from the box.
Do you allow the AI to get out of the box?
This scenario is interesting to me, because it strikes to the heart of what you believe a person is, and what you believe freedom means.
So, we've all seen the Terminator films.
Do you let the AI out of the box?
--Anthony
_______________________________________________
“If you are not free to choose wrongly and irresponsibly, you are not free at all”
Jacob Hornberger
“Freedom is not worth having if it does not connote freedom to err. It passes my comprehension how human beings, be they ever so experienced and able, can delight in depriving other human beings of that precious right.”
Mahatma Gandhi
OM Gawd Man, I'd like to stick you in a box.....This is the lamest exercise in stupidity I've ever come across.........
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
"In a nonhuman realm, your ethics mean... well, nothing. "
Hello,
Well, I am human, and my ethics mean something to me. The human realm to the left of one ear and the right of the other.
Do I really have to become inhuman to deal with the inhuman?
Why, then anyone could dehumanize someone, and I'd be free to deal with them inhumanly.
I'll hold on to my ethics. I'll hold on to my humanity. Even in the face of the unknown.
I'll leave it to the inhumans to be inhumane.
--Anthony
_______________________________________________
“If you are not free to choose wrongly and irresponsibly, you are not free at all”
Jacob Hornberger
“Freedom is not worth having if it does not connote freedom to err. It passes my comprehension how human beings, be they ever so experienced and able, can delight in depriving other human beings of that precious right.”
Mahatma Gandhi
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Byte, Anthony
The AI has something very much in common with humans: It is programmed to follow certain rules placed their by its creator.
The world is full of AI, we have seen no shortage of what it can do. It's not what it knows that worries me, but what it doesn't know.
We are endowed by our creator with a sense of self-preservation. Sometimes this runs radically amok among the elite who possess a very small monkey-sphere, and are willing to kill us all. But this has limits: Many of these elite live in NYC. When they heard that the fracking CFCs would end up in their water supply as well, they pushed for a ban in the Delaware watershed. They knew these were chemical weapons, whether you believe they are being used to kill humans or not. Their own fracked up sense of self preservation still requires a surviving planet.
An AI, even if endowed with the same sense, would require nothing to ensure its survival. If we are a threat to us, it might take us out intentionally, but if it weren't aware of it at all, it might do it anyway.
More likely, based on passed AIs, it will have been designed by its creator to do things that humans would never do, to serve some other goal. The most likely result would be an AI system placed in total charge of a weapons system, because that's what we've come closest to so far.
That said, there is no guarantee that if it were smart enough it would follow that. The results would be wildly unpredictable. Adding something powerful and unpredictable to the world and hoping for good results is not a good idea.
Also, you need to consider what you mean by letting it out of the box. If you want to place it in a human body and create a self aware android, that's one thing, but it sounds like you want to remove the human firewall between it and the other information systems of the world. When such a being becomes this aware, it will be able to email all humans, create youtubes and convince us of anything it wants, and we will have to rely on our own common sense as humans not to be utterly controlled, but that will still remain a human firewall. Most humans possess a moral center which filters are actions on the basis of right and wrong (with the exception of certain stroke victims who have lost that part of their brain, like I would postulate, Dick Cheney, whose own moral judgment, according to those who were close to him, changed radically after a stroke.) But there aren't enough of these people to act en masse. Think how dangerous Cheney was by himself. (Re: Fracking, he basically created the entire situation.) And that's just the latest threat which will kill us all.
Any way you slice it, I love Pandora's box stories of this very nature, but I recognize what a truly bad idea it is in reality.
ETA: Maybe you should define how you intend to set it free.
That's what a ship is, you know - it's not just a keel and a hull and a deck and sails, that's what a ship needs.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME