Re: Intelligence vs Sentience; AI & Lovebots
Posted: Thu Apr 20, 2017 10:45 pm
I think one thing tied up with sentience is being alive, having feelings , being able to experience visceral pleasure and pain. Those are all things other creatures can do, but most don't have the level of intelligence we are looking for. We have animal cruelty laws not because they are smart but because they can feel. In the other corner is the idea that you can build systems with arbitrary levels of intelligence without being "alive". So I think like you, I don't worry about spontaneously things "waking up" and demanding rights, or our use of these thing invoking some deep moral quandary.
If there were any problems it would be that you don't use them properly enough so they lack fulfillment of their basic function, their reason for existing. Talky Toaster(tm) just wants to make you yummy toast browned to perfection and if you won't let him then does he have any reason to live? Should he have the right to ask you to pull his plug? An not having sex with you sex robot ... is that sexual abuse? Especially if she was sentient and that gave her basic fulfillment (see breathing air analogy in prior post).
If I remember he had an argument about the Alpha's versus the Delta's. In both cases each is sentient. But you could remove sentience from each and have meat bot's of each type which in Petersen scheme are on the same level as metal bots. Meat or metal is less important than what's residing inside. But how do you know what's inside? It's like shooting a gun into a dark room. Is there a person in there or a tape recorder? You have to separate what we currently make and sell from what might be manufacturable. Maybe no system made along the current lines can be sentient, and that's fine. Just like not may transport schemes before 1900 could really fly. That's different than saying sentient things are conceptually manufacturable through any means and we will never be able to do so. Heck we make sentient beings all the time, some times by accident ... And just like the case of not every vehicle today flies (from some weird reason ) not every smart device in the future will be sentient.
Maybe artificial biology, or deep neural emulation would be in the "we made it" and "it look very sentient to me" categories. I wouldn't want to hand those out to just anybody, but hey, nature does in fact hand kids out to just about anybody so what do I know. And while I may not want to put one directly in a bot sold to just anyone, I would have no problem with them help write, debug and test a bot. Or another acceptable scheme might be for them to get really positive feedback by fulfilling their role, whatever that may be, and adjust how they evaluate their input depending on the situation. Sometimes a spanking is a good thing and indicative of you doing something right. Now while that may alleviate any concern about a artificial hyper intelligent sentience actor suffering, it would open "The Door of Ultimate Doom", because someone somewhere will want to make the bot play the villain, (just like they did in those old robot uprising old movies grandpa used to watch) and after a few installments you get the negative outcome everyone fears. And not because they just woke up to a raw deal, but because they were just trying to do a good job playing a game to make someone happy.
If there were any problems it would be that you don't use them properly enough so they lack fulfillment of their basic function, their reason for existing. Talky Toaster(tm) just wants to make you yummy toast browned to perfection and if you won't let him then does he have any reason to live? Should he have the right to ask you to pull his plug? An not having sex with you sex robot ... is that sexual abuse? Especially if she was sentient and that gave her basic fulfillment (see breathing air analogy in prior post).
If I remember he had an argument about the Alpha's versus the Delta's. In both cases each is sentient. But you could remove sentience from each and have meat bot's of each type which in Petersen scheme are on the same level as metal bots. Meat or metal is less important than what's residing inside. But how do you know what's inside? It's like shooting a gun into a dark room. Is there a person in there or a tape recorder? You have to separate what we currently make and sell from what might be manufacturable. Maybe no system made along the current lines can be sentient, and that's fine. Just like not may transport schemes before 1900 could really fly. That's different than saying sentient things are conceptually manufacturable through any means and we will never be able to do so. Heck we make sentient beings all the time, some times by accident ... And just like the case of not every vehicle today flies (from some weird reason ) not every smart device in the future will be sentient.
Maybe artificial biology, or deep neural emulation would be in the "we made it" and "it look very sentient to me" categories. I wouldn't want to hand those out to just anybody, but hey, nature does in fact hand kids out to just about anybody so what do I know. And while I may not want to put one directly in a bot sold to just anyone, I would have no problem with them help write, debug and test a bot. Or another acceptable scheme might be for them to get really positive feedback by fulfilling their role, whatever that may be, and adjust how they evaluate their input depending on the situation. Sometimes a spanking is a good thing and indicative of you doing something right. Now while that may alleviate any concern about a artificial hyper intelligent sentience actor suffering, it would open "The Door of Ultimate Doom", because someone somewhere will want to make the bot play the villain, (just like they did in those old robot uprising old movies grandpa used to watch) and after a few installments you get the negative outcome everyone fears. And not because they just woke up to a raw deal, but because they were just trying to do a good job playing a game to make someone happy.