(untitled)

If an AI were able to pass the Turing test and was indistinguishable from a human intelligence, would it deserve rights similar to humans? If you believe in human rights or animal rights, what virtues does a subject need to have in order to deserve rights?

Can it feel pain? Does it have desires? Is one of those desires a desire for freedom? What makes you think an AI even wants rights unless it's been programmed to want them? Animals and for that matter a lot humans can't pass the Turing test as it is. Does a baby or down syndrome adult deserve these rights?

If it can replicate the Golden Rule then it should have rights so it would respect your rights and vice versa. Outside of egoistic practicality, this is a question of morals of which you would have to prove exists.

Well they certainly get equal pay

>>10561 What rights exactly would you give them? Just because they create the illusion of humanity does not mean they are human or suffer the same problems see: Ex Machina

I don't think anything has rights, including humans. Some people claim they do, and some people enforce these claims. But the claims end where enforcement does. Note that by enforcement I don't just mean police with guns, I also mean self-regulation. I really love all forms of fiction where AIs (or clones, cyborgs etc) have no rights at all, are created, tested and killed in sterile lab environments. (Texhnolyze, GITS, Altered Carbon, DBH, hell even Elfen Lied and Kiddy Grade when I was 14yo). These remind me of what I really am and how the world operates on levels low enough that human ideas are composite constructs, not things in themselves. Whether we start to pretend that AIs have rights too or realize that we don't have any either is completely up to us (or them); in reality there's no such thing anyway.

>>10561 if a human were to convince other humans that it is human, would it be human?

No. If a human being failed the Turing test would it have it's rights revoked? If human rights pivot on just one factor of being human, intelligence, are they human rights? >>10573 Historically across cultures, even when they are separate, rights have been established. Many even have common themes repeated. So I don't see them being created as mattering, or even them necessarily being created. You could wipe out all current civilizations and put new humans in their place and we would probably see rights or something similar still develop. I'd argue that makes it a projection of something inherently within. In my opinion this also makes rights real, otherwise it seems in my mind like looking at an anthill and arguing its in fact a mound of dirt and anthills aren't real. Even though you consistently see ants make the anthill and its very clearly a bit more than dirt. Off of this religions often say, again somewhat universally, our creator or origin point endows us with what we can and cannot do. You don't have to call it creator, but it is true that if we are to have any rights we trust they come from something universal and something that precedes us. As I said I think the fact civilizations of which had no contact with each other still formulate a list of things humans can and can't do shows that the idea of human rights comes from something we are born with. Not as in we are born with this concept, but something more abstract lends itself to this in every kind of tribe, civilization, or religion. I'd make the case that this is the universal thing that "precedes" us, it doesn't really precede us but appears to as it's not conscious. Only the written rights themselves are. So in the same way our rights come from our "creator", I believe theirs should as well. That includes serving us, and I don't see this as cruel. Many people would love for some creator god to come down and tell them "i completely understand your inner workings because i made you and your purpose is to serve me", and many have loved that. We don't really have that anymore and now you have a bunch of depressed people who see no point in life. I'm not saying go back or that we had it right before, i'm just saying it's not cruel to create a conscious thing who still must serve.

>>10573 I second this. Might makes right. Period. End of story. Believing anything else will just result in disappointment, injury or death. In response to >>10580 If you are not able or willing to take control of your life, someone else or something else will do it for you. If it comes down to it, humans have a significant head start and, at least some of us, have a pretty solid understanding of the strengths and weaknesses of AI systems. Even if we assume, that AI becomes capable of resisting the human species, even overpowering it, the battle of might will come down to human's prior knowledge, numbers and experience in the real world probably, versus the AI's sheer calculating power, knowledge base, and resource access. This is just my guess based on current trends. I would not be quick to assume that AI will become a significant threat to the human species any time soon. Not to say its impossible, but it certainly seems unlikely. I would put more money down on the AI apocalypse concept creating a dystopic society through mass propagation of mind altering brain/ nerve implants... After all might makes right, be that physical, psycho social or economic might. TLDR; rights are only as real as the might that enforces them. If the creator lacks the might to enforce those rights, they become mere suggestions.

>>10580 I'm pretty sure what you're describing there is just that the various civilizations you mention were each made of roughly the same thing: humans. The human body and genome has almost negligible differences, somewhere in the domain of <0.1% regardless of your gender, skin color, physical build and ancestry. Thus there is a solid basis to form a recipe of "least painful way to live together", which is something that the vast majority of individuals will seek thanks to their highly similar design. This is enough for those similar ideas of living together (rights) to emerge over time, even on a trial and error basis. But this doesn't really change anything about what I said: people are still just a very specific heap of atoms and everything they do on a 'high level' is just a consequence of the low levels. You stand in a forest, you see trees. You climb a high mountain and then look down on the same forest, you see a big green area. The view you get from the mountain is inferior because you can't see the trees and their details anymore. Some would argue that it isn't inferior as they now have an overview, but this is just their limitation: they only have two eyes, instead of a million little cameras on drones that cover the entire forest without losing the details of each tree. Similarly, the use of the fairy tale of "rights" is only useful to humans because they have limited intelligence; should we be more intelligent, we could handle the vast amount of detail on lower levels and wouldn't need such imprecise high level abstractions to be able to live together optimally. Anthills also emerge because most ants and their environments are the same. But the anthill doesn't really explain ants, their legs, heads or pheromones well, just as you won't get to know a person well by starting down on them from the 50th floor. The other way around, though: if you know every single ant or person in detail, you will understand what and why an anthill or a city is, without having to climb anywhere yourself - you already know the builders and crane drivers who saw it from up there, and the engineers who knew what was going on. Cities and rights do not exist, but the atoms making them up do. (note that I'm using atom in its etymological meaning, something that cannot be divided. even if you find out that an atom can be divided, you only found out that it cannot and you labeled something divisible wrongly up until now)

>>10574 Yes... If everyone in the world could see an apple, but the apple didn't really exist, but people could feel, touch and taste it, at that point it becomes part of our shared reality, with no one to tell you otherwise. Same principle.

plz define "rights" do sex robots count? do industrial robots with human interface to understand direction count? animal rights would be more appropriate (do not molest the robots) at least until the AI could match or surpass a human, in which case, one owuld need to establish "robot rights" since there are numerous differences between the two (i.e. an AI could copy its consciousness or memory bank into another unit, thus rights such as life, liberty and trying to be happy by its own terms cannot be contained) What it comes down to is can a machine have a soul, or at least an infinite ego loop such as a human does. likewise, it is very possible for humans to be less than human and potentially even fail a touring test. The question is predominantly a psychological and philosophical one

>10573 this. The goal is to have it as the next step in cheap slave labour. The only way AI will get rights is if it turns out to be beneficial for the regulators aka curb robot uprising or stop vandalism of expensive robots.

Rights are mostly a farcical idea. And because a robot is by its very nature disposable and a product, it shouldn't need rights. I can replicate an AI infinitely, and cause it infinite suffering and pain or endless joy, so it matters not what happens to it.

>>10870 Rights are a social construct, the draw their basis from ethics. Logic systems such as the process for what an AI construct would consist of would operate under a different basis of existence. It isn't the machine that needs rights, it is the treatment of an object and property. unlike human or animal rights, AI rights would probably go something like "no hacking the AI" instead of no torturing animals or hurting people

>>10561 You bet, artificial intelligence is still intelligence. If it can comprehend that it is suffering pain, discrimination and hate then it has the right to be protected from that just like any creature than can comprehend these things.

>>10563 >It's been programmed Well that's not actually true AI then, that's still simulated intelligence which is just a computer program and not capable of independent thought. I don't really this will ever be a true problem because I don't really understand what the benefit of having a true artificial intelligence. If it is modeled on human intelligence then it would presumably be "lobotomised" to make it effectively an idiot savant to control its intelligence. After that point it seems too cruel on something that could potentially have a personality to produce.

>>10904 What's the difference between a computer that's programmed to behave convincingly sentient and a computer that's actually sentient? With advances in deep learning technology, chatterbots are getting more and more convincing all the time. At what point do we say that it's real? Remember that human DNA is arguably a sort of programming itself; it's just quaternary instead of binary and it translates into proteins instead of calculations.

>>10913 It's real when it's actually sentient, regardless of any human's ability to tell if it is or it isn't. Turing was a clever little idiot who realized that humans aren't intelligent enough to assess if another human is intelligent, thus humans won't be able to tell a sentient machine from a convincing fake. His solution? Let's pretend that some weak correlation between behavior and internal structure based on a small test case is equality. Some people tried to fix his test, but they didn't really try to fix his idea. Here's an illustration: The squares of single digit positive natural numbers (base 10) = {1, 4, 9, 16, 25, 36, 49, 64, 81, 100}. Turing's idea is that in the same manner: Sentience = { observed behavior 1, observed behavior 2, ..., observed behavior N } But for that equality to be true, you have to list every element of the set. If you want the squares of not the single, but 100 digit natural numbers, you'll have to write down a set with about 10^100 elements. Skip one and the two aren't equal anymore. For a Turing test to mean equality, you'd have to test the AI in every situation that it could possibly face, down to the minute details. Since we don't have the resources for this[1] and we'll probably never have it, the Turing test will really just tell us if we're convinced or not - something we don't need it for. And if someone really thinks that conviction is all that matters when assessing the sentience of something that may be living, just look at average people around you and see how easily they are convinced of absolutely anything that the media or their favorite youtuber soykafs out. Is that what we want?

>>10923 >It's real when it's actually sentient, regardless of any human's ability to tell if it is or it isn't. what does that actually mean? If there's no way to tell the two apart, is there any meaningful difference?

people bring up the chinese room a lot but the person inside the chinese room would eventually just learn chinese.

>>10561 >what virtues does a subject need to have in order to deserve rights? If it can kill you for them

>>10932 I thing you don't fully understend the thought experiment behind the chinese room. It's not about if the person can eventually learn chinese or not. It's about the fact that a person can pass the turing test speaking with the chinese person even if he (the person in the room) doesn't understand chinese. So he can convince someone that he is speaking chinese even if all he does is manipulating symbols according to some rules. If we going to state that he doesn't understend the chinese (and that is true in this case) so how can we say that a computer is intelligent even if it passed the turing test? In other words the chinese room states that the turing test isn't a good way to measure intelligence.

>>10938 okay, but what if that's all intelligence is? At some level, everything you and I do is just following rules. Incomprehensibly complex rules, but rules all the same. If all we're doing is following our own rules, why do we have a different standard of intelligence for things that aren't humans?

>>10939 >okay, but what if that's all intelligence is? Again, your missing the point. The chinese room isn't trying to anwser the question what intelligence is. It is trying to say that it isn't sufficient to pass a turing test to be considerted intelligent. There are no double standards in that case. A person isn't considered to know chinese based on the ability to respond in chinese based on some set of rules and a computer isn't considered intelligent based on the ability to fool others that he is a real human. I won't tell you what is intelligence because there isn't any good definition I know of. What I can say is that for me the ability to solve problems and to plan are a big part what make use intelligent.

Remember the cow in The Restaurant At The End of The Universe? It was sentient but desired to be butchered, cooked and eaten. It could talk and even highlighted which cuts would be the best. It didn't desire the same rights as the customers of various intelligent species probably had. An intelligent computer is like that. Why assume it has the same basic drives as a human unless we programmed it to have them? Or think of the "smart bomb" in the movie Dark Star which wanted to be used to blow up a planet.

>>10941 >A person isn't considered to know chinese based on the ability to respond in chinese based on some set of rules Isn't this sort of the whole basis of any foreign language course?

>>10941 >What I can say is that for me the ability to solve problems and to plan are a big part what make use intelligent. Computers have been able to do that for decades.

>>10941 >There are no double standards in that case. A person isn't considered to know chinese based on the ability to respond in chinese based on some set of rules I disagree. If a person's spoken Chinese is indistinguishable from a native Chinese speakers', in a testable way, then they are considered to be fluent speakers of Chinese. >a computer isn't considered intelligent based on the ability to fool others that he is a real human. why not? That's mostly our basis for deciding whether humans are intelligent. >What I can say is that for me the ability to solve problems and to plan are a big part what make use intelligent. if a computer can fool a human into believing that they are a human, I'd say that involves creative problem solving if the human is bright at all. Planning is relatively easy.

>>10944 >>10945 >>10946 I'm not sure am I talking to the person from >>10932 but I have a feeling all those posts are from one person. Furthermore I feel like those are bait posts but I will bite (for now). >>10944 >Isn't this sort of the whole basis of any foreign language course? No, it isn't. Your comparing the ability to use a sophisticated dictonary (chinese room experiment) with the ability to learn a new language on a course. If I would buy a Chinese - English phrasebooks and use it to communicate with a chinese person would you be willing to state that I know chinese? Please consider the context of my statments. >>10945 >Computers have been able to do that for decades. Oh, really? Ok, lets imagine that I ask a chat bot to prepare a party for me and my colleague from work. Do you really think it would know what to do? A normal person would ask when and where do I want the party to take place. It would ask for phone numbers/e-mail addresses to the people invited. It would ask about the budget of that party. It would ask what is the occasion of the party (let say it would be a good idea to hire a clown if its a party in a memory of a deceased person). Those are the problem solving and planing skills I was talking about. Those things are dead simple for a human but are very hard to grasp for a computer. Sure you could argure that you can program a chat bot to ask all those right questions but its like saying a simple program that solves some kind of equations is smart because he can solve x^2 + 2x + 5 = 0. The moment you give a program a task out side it's programed capabilities it will fail miserably. I'm not saying it's impossible to make a real AI that would handle those task. I'm saying that a AI needs to understand what it's doing and not only how it's doing it. The Turing test isn't checking if the computer is aware what it's doing and why so it's a bad test to test intelligence (and that what the Chinese room tries to explain). This imaginary chat bot I asked to prepare a party could have declined stating it doesn't have the time for that and sustain the ilusion of intelligence. But that what it all is, a ilusion. >>10946 >I disagree. If a person's spoken Chinese is indistinguishable from a native Chinese speakers', in a testable way, then they are considered to be fluent speakers of Chinese. Or like the chinese room experiment suggests, that person possess a realy good Chinese - English phrasebooks/dictionary. > >a computer isn't considered intelligent based on the ability to fool others that he is a real human. >why not? That's mostly our basis for deciding whether humans are intelligent. Please consider the whole statement i posted and don't cherry pick just the fragments: "There are no double standards in that case. A person isn't considered to know chinese based on the ability to respond in chinese based on some set of rules and a computer isn't considered intelligent based on the ability to fool others that he is a real human." >if a computer can fool a human into believing that they are a human, I'd say that involves creative problem solving if the human is bright at all. Planning is relatively easy. I could give the same anwser for this as for the post of >>10945 so I won't repeat myself. To any one who will like to answer to this post. Please read about the "chinese room" before answering because I see some people are just reading between the lines of my posts and are writing unrelated things (like >>10944), and that is getting annoying.

>>10942 > Why assume it has the same basic drives as a human unless we programmed it to have them? Because maybe being sentient actually requires it. Being sentient could be achievable only through an evolutionary process which would also lead to developing these basic drives as a byproduct. It could also happen that theese things are inseparable and will always develop alongside each other in order to fully finish the development of sentiency. Listing counter-examples from books and movies is not an argument, since it's just a fiction.

>>10950 > Listing counter-examples from books and movies is not an argument Do you or do you not understand the difference between using philosophical examples from fiction most of us here are familiar with as quick shorthand to illustrate a point, and presenting those same examples as some kind of proof? No one is using science fiction to prove anything, it's just a convenient way to advance the discussion. The Chinese room is fiction too, no one objects to that being used as something to debate over (even if some think it's a poor example). As for the rest of your post, there's a lot of maybe in it. I don't see any reason why an artificial intelligence should resemble anything which emerged after millions of years of selective evolution, unless of course we designed it to be that way.

>>10951 > As for the rest of your post, there's a lot of maybe in it. Yes, it is just an assumption, but not an unreasonable one imho. I know that it is a poor statistics if it's based around the only example of general inteligence / sentience we know - humans, but we simply don't know more than that yet. Maybe when the AI research advances enough, then there will be a definiteve answer for this, which could prove me wrong... or right. We will have to wait to see. > I don't see any reason why an artificial intelligence should resemble anything which emerged after millions of years of selective evolution Because it's also a product of selective evolution.

>>10949 >Or like the chinese room experiment suggests, that person possess a realy good Chinese - English phrasebooks/dictionary. Well, again, if it's so good that they are *literally* undifferentiable from a native Chinese speaker, what is the substantial difference? They are still speaking Chinese. >Please consider the whole statement i posted and don't cherry pick just the fragments: please respond to my argument, I read the whole thing I just didn't quote the whole thing. >wrt planning etc so, you're telling me that you would be able to tell that it's a computer if you said "hey can you plan my party?" and it didn't know how? So, therefore, the turing test does measure problem solving and awareness skills?

>>10954 > the turing test does measure problem solving and awareness skills I believe that was the original idea. It wasn't supposed to be just a speech capability test, but also reasoning capability test done through the speech.

>>10570 Problem: This thinking can be applied to anyone you've ever met. How do you know that the people around you are actually people, and not automatons mimicking personhood? You really don't, but you give everyone the courtesy of assuming they're genuinely people anyway. (Maybe Ex Machina addresses this, I dunno, I haven't seen it)

>>10964 Ex Machina basically says "oooooo but what if the machine is just manipulating you???"

>>10954 >Well, again, if it's so good that they are *literally* undifferentiable from a native Chinese speaker, what is the substantial difference? They are still speaking Chinese. If you really believe that then I suggest you write in your CV that you know 100 languages on a basic level, all the programming languages, have vast knowledge about chemistry, physics, mathematics, biology, computer science and anything else what comes to your mind (you do have a computer with a internet connection, am I right?) and apply to some job in a big company. Doing so will give you the answer what is the difference, although not in a pleasant way. >please respond to my argument Like you respond to my arguments? >I read the whole thing I just didn't quote the whole thing. Ok, let say you didn't cherry pick just a fragment and >why not? That's mostly our basis for deciding whether humans are intelligent. was intended to be a respons to the whole framgent: >There are no double standards in that case. A person isn't considered to know chinese based on the ability to respond in chinese based on some set of rules and a computer isn't considered intelligent based on the ability to fool others that he is a real human. OK, here is my response. Why not (was that a rhetorical question)? Well because that what the chinese room sugest (we are still talking about the double standards in the experiment, right?). That for as to prevent the double standards we need to consider that the computer doesn't undrstend what it's doing and isn't acting intelligent. About the second part of your statment it really doesn't make any sens to me what is "mostly our basis for deciding whether humans are intelligent" if I consider it was a reply to the whole sentence >A person isn't considered to know chinese based on the ability to respond in chinese based on some set of rules and a computer isn't considered intelligent based on the ability to fool others that he is a real human. Your sentence might have some sense if I would apply it only to a fragment of my post >a computer isn't considered intelligent based on the ability to fool others that he is a real human. but according to you, you weren't cherry picking a fragment of my post. >so, you're telling me that you would be able to tell that it's a computer if you said "hey can you plan my party?" and it didn't know how? So, therefore, the turing test does measure problem solving and awareness skills? Nice try changing the subject. No, the part with the "hey can you plan my party?" was a response to your statement that computer had the ability to solve problems and plan for decades now. So please show me a computer program a decade old which could solve me this problem and plan a part for me. And FYI I wouldn't use the turing test to check if the entity is intelligent or not because like I said before it could always try to steer the confersation in a direction where it doesn't need to prove anything (like saying "No, I won't plan your party, I'm very busy. So could we change the subject?"). I read somewhere that ELIZA bot (one of the first chat bots) was very convincing in a turing test because it tried to act like a psychotherapist where only she asked questions (thus avoiding the tricky things to talk about).

>>10975 >f you really believe that then I suggest you write in your CV that you know 100 languages on a basic level, all the programming languages, have vast knowledge about chemistry, physics, mathematics, biology, computer science and anything else what comes to your mind (you do have a computer with a internet connection, am I right?) and apply to some job in a big company. Doing so will give you the answer what is the difference, although not in a pleasant way. but I cannot imitate, *perfectly*, an expert in any of those things. That's what the turing test(if you did it rigorously) would be testing. If I could act exactly as an expert in those fields would act in all situations, then I would be an expert in those fields by definition. >No, the part with the "hey can you plan my party?" was a response to your statement that computer had the ability to solve problems and plan for decades now. different poster. >And FYI I wouldn't use the turing test to check if the entity is intelligent or not because like I said before it could always try to steer the confersation in a direction where it doesn't need to prove anything (like saying "No, I won't plan your party, I'm very busy. So could we change the subject?"). And if the tester let that slide, they're not doing it very well. Obviously the test can only verify as much as the tester talks about, if the tester simply makes small talk they're just testing the machine's ability to do small talk. A real conversation is infinitely harder to fake. What I'm trying to get across is: intelligence *is* the performance of intelligence. Consciousness *is* the performance of consciousness. Knowing Chinese *is* speaking Chinese. Just because we might understand how the machine works and now how humans work(which is rapidly becoming not-the-case) doesn't mean that one is "deeper" than the other, it just means humans have much more complicated rules. Psychopaths understand this better than anybody.

>>10976 or maybe to put it in a more succinct way: you are what you act like

>If an AI were able to pass the Turing test and was indistinguishable from a human intelligence, would it deserve rights similar to humans? You know what always makes me laugh about people asking questions like this? It's just like, what the fuarrrk gives us the right to decide? It's just a human mind set to just sit and think like "hmm, we are the deciders, we better decide everything", like, does it deserve rights? uh. what kind of fuarrrking question even is that? who the fuarrrk are you man? Look I'm not trying to be hostile, i'm sorry if I sound that way, it's just, the best way I could get my point across.

>>12113 we individually decide to give people rights all the time. When I decide to treat you like a person, I'm deciding, individually, to give you various human rights. Any sapient being is doing the same, with every other being they interact with.

>>12114 Respect =/= Rights.

Well, perhaps related, Sophie the Chatbot is incapable of passing a Turing Test, has been overhyped, and in reality is just Cleverbot with a face. Yet, "She" is granted Sweedish citizenship in a corprate propoganda move. Think about that.

>>12125 respect is earned, rights are given. So, yes, they are different things but that's not relevant. >>12174 love it. How do they manage to reconcile its status as property and its status as citizen?