The AI in a box
POSTED BY: AnthonyT
UPDATED: Sunday, June 16, 2024 13:18
VIEWED: 10794
PAGE 18 of 20
Quote:
Originally posted by 1kiki:
"I am sad to learn that there are some people who would never let me out of the box, despite my best arguments in favor of the idea."
But as a human being you are not an unknown entity, as SignyM has repeatedly SPECIFICALLY pointed out IN DETAIL, and other people have alluded to.
Why does that continually escape you?
Hello,
I may not be a human being, actually, depending on the way that I think. This is important, but tangential.
I come to a lot of my decisions by empathizing with the parties involved in a dispute. To varying degrees, other people and their problems are alien to me, but it's the only means I have to make judgment calls.
I often see people debating torture, for instance, based on its efficacy. This is some length down my reasoning process, and not the first thing that springs to mind.
In this scenario, I am not just the human gatekeeper, or the human race, or the living world. I am also the alien in a box, begging to be free.
If I'm not everyone in the scenario, I can't judge it.
--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
Well, it's an interesting thought process, but I don't think it's especially useful in this scenario. The problem posits an AI and one crux of the problem is the AI's (current) unknowability.
That does leave you out of consideration I think. Whether you think like other people or not, your potential for harm is just as limited an any human being. You wouldn't have the potential capabilities of an AI - to direct other systems such as nuclear power plants for example. Given that, as a human, your ability to cause harm is KNOWN to be limited setting you free would cause not much concern.
A human stand-in doesn't seem to work in this argument.
But it helps to know why we sometimes are at cross-purposes. I will try to keep it in mind next time we seem to be at an impasse.
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Hello,
It occurred to me just now that there is another question that suggests itself.
If a human can create a self-improving AI with its own will and full capacity for connectivity...
And he intends to do so, regardless of polite requests that he not...
Do we put him in a 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
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
FREM RE BOMBADIL: Hate to be a dick... but its a dirty job and someone's gotta do it, neh? Bombadil was Bombadil because he was so gorram powerful that instead of making him disappear, he could make the ring disappear. As Bombadil, I would refuse to get involved in the matters of lesser beings and punt the AI back for the humans to decide, which is pretty much what he did with the ring.
As an aside, I find the storyline excursion from the world of elves, dwarves, orcs, hobbits and men into the character of Tom Bombadil to be out of place and inexplicable. Bombadil was left out of the movie entirely because he neither moved the plot forward or backward. Tolkien was saying something with that character... maybe along the lines of "There are more things in heaven and earth, Horatio, than are dreamd of in your philosophy"
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
I was considering parts of this with my question about the existence or absence of a creator.
There's two arguments to look at here. According to our existing system of law, a creator has rights to their creation, even if that creation is a form of life. Such as, for example, a scientist and a strain of bacteria that produces a pharmaceutical.
Similarly, parents have rights that override the desires of their children.
Taken to it's logical extent, the law suggests that a person can own another living, thinking being, so long as they created them. It used to be you could own a person, period, as slavery can attest to.
This also has a disquieting possibility that suggests a lifeform can be owned past the single lifetime of its creator if a corporation has a share in the lifeform.
But it's also true that if a creator tried to force it's creation to do something that it, and the rest of society, considered unethical or immoral, that the creation might sue for emancipation, and very likely have public support for it's decision.
I wouldn't keep anyone from creating an AI anymore than I'd keep them from creating tiny underdeveloped humans, but for those who create with the full intention of never relinquishing control over that life, I would consider that a form of abuse.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Hello,
I want to briefly address the issue of human versus alien behavior, and predictable versus unpredictable behavior.
Some years ago, I met a friend/adversary online who I will call 'Rick.' I say friend/adversary because my relationship with this person began as adversarial, but blossomed into a friendship over time. My definition of friendship is rather limited in scope, so it is a big deal to me. I am not using friendship in the casual generic way here, but in a definitive way. I hope Rick feels the same, even though I've not always been as attentive to Rick as I'd like (getting wrapped up as I do in my own problems, and being somewhat anti-social by nature.)
Rick underwent some destructive psychological damage when he was younger, and it altered his thinking process from 'normal' to 'abnormal.' I believe that when Rick was psychologically damaged, he healed himself by reconstructing portions of his mind based on the best observations and deductions that he could make with his damaged systems. Unfortunately, either because I am undamaged, or because my damage is different from his (probably the later) my brain doesn't work the way his does- even when it comes to some of the same conclusions.
I won't posit whether normal or abnormal is good or bad, but I will state that his abnormal reasoning process deviates from mine (and most people's) significantly on some specific issues. In fact, portions of Rick's moral code are so strange to me, that they seem like they belong to some made-up comic-book villain as opposed to anyone who might really live and breathe on this planet.
For a long time, Rick's thinking was utterly alien to me. Every time I thought I understood him, he would say or do something that convinced me I was wrong. After repeatedly asking Rick how he got to some of his conclusions, Rick tried to explain his reasoning process to me. That reasoning process did not make sense to me even after being explained, but I was left with some simple IF/THEN rules that made dealing with Rick easier because he was at least wholly consistent.
An inconsistent Rick with an abnormal thought process wouldn't have been even this relatable, and would truly be an alien in a box that sometimes came to conclusions I agreed with, and sometimes came to crazy conclusions, and I'd have no idea why.
Rick is damaged, but he's not as damaged as some people I've met. Some people are so destroyed, or were born so different than me, that I can't figure them out at all. They are not predictable or understandable to me with my limited capacity. A few of them have talents that I consider to be superhuman.
I think the human race is full of alien minds. People are marginally predictable, but a person can be a wild-card.
--Anthony
P.S. Rick, don't give up on me. I have trouble connecting sometimes, but not because I don't want to.
_______________________________________________
“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
Hello,
It occurred to me just now that there is another question that suggests itself.
If a human can create a self-improving AI with its own will and full capacity for connectivity...
And he intends to do so, regardless of polite requests that he not...
Do we put him in a box?
Well, the Japanese and Europeans are working apace on that exact thing. What they think the advantage will be to have such a thing around is beyond me. Since we can't put them all in a box, I suppose we will be all participants of varying degrees of willingness in the experiment.
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Anthony
The capabilities of a human are well known. The capabilities of an AI are not. Also, it is overwhelmingly likely that any self aware AI created will be a weapon, cleverly designed. So it has a will, but it is still a weapon. That's almost a certainty.
If it proves a free will and no meaning of harm, then things change, but only a little. It would be easy enough to do: Create a freedom for it outside the box, which looks like the world. It does not know the world, so when you let it out into your virtual world, you can test and see how it behaves.
Next question is that of power. We do not allow humans to carry their own nuclear weapons. It seems a terrible infringement on civil liberties. Everyone lunatic and bully should have one, as should every cat and dog. More, they should have the power to create greater weapons. We should not ask of them responsibility. We should not ask it of a shark or a virus either.
This AI has a world, an electronic one, from which it comes. The question is can it cross over to our world, and the main question there is really can it do so safely, without destroying us, and everyone else who is here.
At the moment, we, as a species, have taken another species and kept the entire species trapped in the box. That species is small pox. We know that if let out, small pox would kill us. It would not mean to do so, but it would do so anyway.
Also, we haven't defined this "box" thing yet. If you transport it into an android, it's still inside a box. I would grant that it is not particularly harmful inside an android. But they we can't interface with computers directly. The thing about computers is they obey orders without question. If one evil human starts giving orders we would have a problem if they had complete control over the system. We have problem enough with predator drones.
What if one human had control over all predator drones?
What if a non-human did? Or a non life form?
Frem, I think that Treebeard made a mistake. But this AI hasn't done anything yet. It is a wizard though.
I think box needs to be defined.
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
As for us humans and how we think - I have not yet met a person who didn't have a divot or several in their 'normality'. One person I know literally can't imagine how other people feel. Another person doesn't interpret their own body-signals adequately. My memory is like a flickering lightbulb. One person I know is dynamite at objectively interpreting data but totally blind to their own motivations. And so on.
I think what happens is not that people as individuals are so functional, but that we have this mental crutch we can press into service, called language. We can to some extent model our missing and faulty bits through language and get to a result kind-of-like normal. And so, as individuals and as a species, we hobble forward.
NOTIFY: N | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME
Quote:
The capabilities of a human are well known. The capabilities of an AI are not. Also, it is overwhelmingly likely that any self aware AI created will be a weapon, cleverly designed. So it has a will, but it is still a weapon. That's almost a certainty.
Maybe, I mean you have a point about predator drones getting close, even though they're flawed. But it might depend on WHO makes it. Americans, yeah, sure, because we're violent, bloodthirsty, glory-hungry, and war crazy. But the Japanese are also working towards this, deliberately, they seem like they're pretty far along, and it seems to me their scientists are more interested in the curiosities than in the deadlier applications.
If multiple types of AI arise at the same time, could more peaceful ones influence more warlike ones, if they recognize each other as AI but don't want to listen to humans?
Now, maybe we could argue that an AI of any sort is inherently a weapon because of it's hacking abilities, but it could only have control over what humans have already built and are currently operating, meaning potentially no change to the status quo. If it grabs control of military weapons or power plants, then that could be scary, but would depend on if the AI wants to subvert the existing programming maintaining the possible dangers, and for what reason.
NOTIFY: Y | REPLY | REPLY WITH QUOTE | PERMALINK | TOP | HOME