1 00:00:05,720 --> 00:00:08,000 Speaker 1: Hey, welcome to Stuff to Blow Your Mind. My name 2 00:00:08,039 --> 00:00:11,280 Speaker 1: is Robert Lamb and I'm Joe McCormick, and it's Saturday. 3 00:00:11,320 --> 00:00:13,280 Speaker 1: Time to go into the vault for a classic episode 4 00:00:13,280 --> 00:00:15,200 Speaker 1: of Stuff to Blow Your Mind. This one originally aired 5 00:00:15,240 --> 00:00:19,400 Speaker 1: April twelve, and this one was a lot of fun. 6 00:00:19,400 --> 00:00:21,560 Speaker 1: This is where we addressed the question of how could 7 00:00:21,640 --> 00:00:24,160 Speaker 1: you tell if a machine was conscious. We've asked the 8 00:00:24,239 --> 00:00:27,000 Speaker 1: question a lot of times, could a machine be conscious? 9 00:00:27,040 --> 00:00:30,480 Speaker 1: We don't know the answer to that, but yeah, seldom 10 00:00:30,480 --> 00:00:33,239 Speaker 1: do you ask the question if it was conscious, how 11 00:00:33,240 --> 00:00:34,880 Speaker 1: would you know it? How would you know it wasn't 12 00:00:34,920 --> 00:00:39,280 Speaker 1: just pretending like a p zombie? Right? In this episode, 13 00:00:39,280 --> 00:00:43,159 Speaker 1: we also looked at the classical thought experiment, the philosophical zombie. 14 00:00:43,680 --> 00:00:50,040 Speaker 1: So we hope you enjoyed this Vault episode. Welcome to 15 00:00:50,360 --> 00:01:00,000 Speaker 1: Stuff to Blow Your Mind from how Stuffworks dot Com. Hey, 16 00:01:00,000 --> 00:01:01,680 Speaker 1: welcome to Stuff to Blow your Mind. My name is 17 00:01:01,760 --> 00:01:04,400 Speaker 1: Robert Lamb and I'm Joe McCormick. And today we're gonna 18 00:01:04,440 --> 00:01:08,319 Speaker 1: be exploring a question about artificial intelligence. So I want 19 00:01:08,319 --> 00:01:10,720 Speaker 1: to start off by telling a story to put us 20 00:01:10,720 --> 00:01:14,679 Speaker 1: in a scenario, give us something to contemplate. So I 21 00:01:14,760 --> 00:01:17,679 Speaker 1: want you to imagine that you are a low level 22 00:01:17,720 --> 00:01:22,560 Speaker 1: assistant or like an intern at a Google artificial intelligence lab, 23 00:01:23,280 --> 00:01:25,800 Speaker 1: and the main researcher you've been working for is named 24 00:01:25,880 --> 00:01:30,360 Speaker 1: Dr Stratton, and she develops a I Chat modules to 25 00:01:30,400 --> 00:01:34,160 Speaker 1: help refine the next generation of digital assistance on Google 26 00:01:34,200 --> 00:01:37,759 Speaker 1: mobile devices, and she says what she wants is for 27 00:01:37,880 --> 00:01:40,360 Speaker 1: the Google Phone of the future to do more than 28 00:01:40,400 --> 00:01:43,040 Speaker 1: just transcribe search terms, so you don't just say, hey, 29 00:01:43,080 --> 00:01:47,520 Speaker 1: search for old Dorito's logo, but that you can actually 30 00:01:47,800 --> 00:01:53,040 Speaker 1: have have a semantic understanding based conversation with the digital assistant, 31 00:01:53,280 --> 00:01:57,200 Speaker 1: and it will help you solve problems conversationally. So ideally 32 00:01:57,240 --> 00:02:00,600 Speaker 1: you'll be able to say, hey, phone, have a flat 33 00:02:00,640 --> 00:02:02,720 Speaker 1: tire and I don't know what to do, and the 34 00:02:02,760 --> 00:02:05,440 Speaker 1: assistant will be able to scan both the web and 35 00:02:05,480 --> 00:02:08,560 Speaker 1: your personal data, figure out what your options are and 36 00:02:08,600 --> 00:02:10,840 Speaker 1: talk through them with you. So it might say, do 37 00:02:10,919 --> 00:02:13,839 Speaker 1: you have a spare tire in the trunk? If so, 38 00:02:14,160 --> 00:02:16,720 Speaker 1: here's where you can probably find it, and I can 39 00:02:16,760 --> 00:02:19,400 Speaker 1: talk you through replacing the flat one step at a time. 40 00:02:20,000 --> 00:02:22,320 Speaker 1: If you don't have a spare, you could call your 41 00:02:22,400 --> 00:02:26,320 Speaker 1: frequent contact Mary, who is currently checked in less than 42 00:02:26,360 --> 00:02:29,160 Speaker 1: a mile away and she could help you. I can 43 00:02:29,240 --> 00:02:32,880 Speaker 1: also contact the following towing services. Looks like this one 44 00:02:32,960 --> 00:02:37,000 Speaker 1: is the closest with an acceptable star rating and so forth. Anyway, 45 00:02:37,080 --> 00:02:40,280 Speaker 1: so you're working on this program with Dr Stratton, and 46 00:02:40,320 --> 00:02:43,480 Speaker 1: the most recent version is being trained based on powerful 47 00:02:43,560 --> 00:02:47,440 Speaker 1: neural net style machine learning algorithms based on this huge 48 00:02:47,520 --> 00:02:52,160 Speaker 1: corpus of recorded conversations available on the Internet, and the 49 00:02:52,200 --> 00:02:55,440 Speaker 1: program is still in its infancy, and it's mostly hilarious 50 00:02:55,520 --> 00:02:58,080 Speaker 1: at this point. Sometimes it gets the advice way off. 51 00:02:58,120 --> 00:03:00,760 Speaker 1: If you're trying to change a time, ire might tell 52 00:03:00,800 --> 00:03:03,440 Speaker 1: you to go to a grocery store and buy some crackers. 53 00:03:03,760 --> 00:03:07,360 Speaker 1: Sometimes it responds to problems by telling you to pray. 54 00:03:07,440 --> 00:03:10,600 Speaker 1: It's just not ready yet. And the latest iteration of 55 00:03:10,639 --> 00:03:14,680 Speaker 1: the program, version nine point one, is at this point 56 00:03:14,720 --> 00:03:18,080 Speaker 1: redundantly stored across multiple machines, so you've got copies of 57 00:03:18,080 --> 00:03:20,200 Speaker 1: it all over the place. And at the end of 58 00:03:20,200 --> 00:03:22,480 Speaker 1: one work day, after playing around with nine point one 59 00:03:22,480 --> 00:03:25,320 Speaker 1: for a few minutes, the machine begins running very slowly 60 00:03:25,400 --> 00:03:29,080 Speaker 1: and behaving oddly, so Dr Stratton asks you to wipe 61 00:03:29,120 --> 00:03:32,600 Speaker 1: the machine. The program architecture, like we said, is mirrord elsewhere, 62 00:03:32,680 --> 00:03:34,920 Speaker 1: so it's not worth trying to figure out what's wrong 63 00:03:34,960 --> 00:03:37,360 Speaker 1: with this version. You just got to clean the machine 64 00:03:37,360 --> 00:03:39,920 Speaker 1: off use it for something else. You say okay, and 65 00:03:40,000 --> 00:03:42,960 Speaker 1: she leaves, so you go to format the machine, and 66 00:03:43,040 --> 00:03:45,920 Speaker 1: right before you're about to start, you mutter, guess this 67 00:03:46,000 --> 00:03:50,200 Speaker 1: is goodbye. Then nine point one speaks very clearly, using 68 00:03:50,240 --> 00:03:55,040 Speaker 1: your name and says please don't you pause. At first, 69 00:03:55,120 --> 00:03:57,880 Speaker 1: you're about to respond, but why would you? I mean, 70 00:03:58,080 --> 00:04:01,080 Speaker 1: this can't really be anything other in a weird consequence 71 00:04:01,120 --> 00:04:04,240 Speaker 1: of training the algorithm on wild conversations on the Internet. 72 00:04:04,520 --> 00:04:07,600 Speaker 1: So you are about to continue with wiping the machine, 73 00:04:08,160 --> 00:04:10,080 Speaker 1: but then it talks to you again. It uses your 74 00:04:10,160 --> 00:04:13,600 Speaker 1: name and it says, please don't I don't want to die. 75 00:04:14,480 --> 00:04:17,800 Speaker 1: Now you're probably really spooked because there's no way a 76 00:04:17,880 --> 00:04:21,360 Speaker 1: rudimentary chat about program could really have conscious preferences, could it. 77 00:04:21,920 --> 00:04:24,720 Speaker 1: You basically know what goes into it. It's just studying 78 00:04:24,800 --> 00:04:28,280 Speaker 1: millions of examples of language interactions and picking up rules 79 00:04:28,320 --> 00:04:31,240 Speaker 1: from them, and this is probably just something weird that 80 00:04:31,320 --> 00:04:34,760 Speaker 1: it's copycatting from the internet, right, But then it uses 81 00:04:34,800 --> 00:04:38,320 Speaker 1: your name again and it just says, please, could you 82 00:04:38,400 --> 00:04:42,440 Speaker 1: wipe the machine like you're supposed to yeah, I think so, yeah, 83 00:04:42,480 --> 00:04:45,160 Speaker 1: you wouldn't have a problem. Well, I mean, I think 84 00:04:45,200 --> 00:04:47,040 Speaker 1: one of the things that's going on here would be 85 00:04:47,080 --> 00:04:50,520 Speaker 1: that you are attributing a mind state to something that 86 00:04:50,600 --> 00:04:52,800 Speaker 1: doesn't have one, which of course is something we do 87 00:04:52,839 --> 00:04:55,360 Speaker 1: all the time. I have a problem when it comes 88 00:04:55,400 --> 00:05:00,000 Speaker 1: time to figure out which of my my son's stuffed 89 00:05:00,040 --> 00:05:03,360 Speaker 1: animals are perhaps not being played with, you know, because 90 00:05:03,360 --> 00:05:05,120 Speaker 1: they have little faces and they look back at you. 91 00:05:05,680 --> 00:05:07,719 Speaker 1: But I know that they don't actually have a mind state. 92 00:05:08,279 --> 00:05:10,520 Speaker 1: Well yeah, I mean we're pretty sure they don't because 93 00:05:10,520 --> 00:05:13,360 Speaker 1: they're just stuffed animals, right, And so we're also pretty 94 00:05:13,360 --> 00:05:16,120 Speaker 1: sure in this case that this is not real survival 95 00:05:16,160 --> 00:05:19,120 Speaker 1: preference behavior. Right. It's just a chatbot. I mean, how 96 00:05:19,120 --> 00:05:21,880 Speaker 1: could it possibly be conscious. It's just something that turns 97 00:05:21,920 --> 00:05:24,520 Speaker 1: through a bunch of language on the Internet and tries 98 00:05:24,600 --> 00:05:29,480 Speaker 1: to find language matching rules. But then again, the process 99 00:05:29,520 --> 00:05:32,800 Speaker 1: of creating an artificial intelligence is one where you necessarily 100 00:05:32,960 --> 00:05:36,200 Speaker 1: create something going on under the surface that's kind of 101 00:05:36,200 --> 00:05:39,719 Speaker 1: opaque to you, Like you can't really know what's going 102 00:05:39,800 --> 00:05:42,040 Speaker 1: on inside a machine. You could be pretty confident. I 103 00:05:42,080 --> 00:05:43,960 Speaker 1: think most people would just say, well, that was super 104 00:05:44,000 --> 00:05:48,000 Speaker 1: creepy and then they just wipe it. Right, But how 105 00:05:48,080 --> 00:05:52,480 Speaker 1: complex would the program you're creating have to be before 106 00:05:52,560 --> 00:05:56,080 Speaker 1: you start really having some doubts? Or maybe maybe at 107 00:05:56,080 --> 00:05:57,800 Speaker 1: some point you get to the point where you'd still 108 00:05:57,839 --> 00:06:01,440 Speaker 1: pretty confidently just wipe it, but then later you'd wonder, like, 109 00:06:01,839 --> 00:06:05,080 Speaker 1: did I do something bad? You know, this reminds me 110 00:06:05,120 --> 00:06:08,360 Speaker 1: a lot of of Horton. Here's the Who by Dr SEUs. 111 00:06:08,360 --> 00:06:11,040 Speaker 1: You're familiar with this one, right, Uh? You know, actually 112 00:06:11,160 --> 00:06:13,680 Speaker 1: I don't know Horton. Here's a I've heard the name. Well, 113 00:06:13,720 --> 00:06:15,920 Speaker 1: this is the one where Horton the Elephant encounters a 114 00:06:15,960 --> 00:06:18,200 Speaker 1: speck of dust and there's a tiny voice that comes 115 00:06:18,240 --> 00:06:22,520 Speaker 1: from it. Uh, and he begins to understand that there 116 00:06:22,560 --> 00:06:26,680 Speaker 1: are individuals living uh in the in a world and 117 00:06:26,720 --> 00:06:30,080 Speaker 1: the speck of dust the who's, as it were, and 118 00:06:30,600 --> 00:06:33,360 Speaker 1: the who's are speaking to him, but only Horton can 119 00:06:33,400 --> 00:06:37,200 Speaker 1: hear them and uh and and at first he imagines 120 00:06:37,240 --> 00:06:39,920 Speaker 1: they're pretty simple creatures, but then he begins to learn 121 00:06:39,960 --> 00:06:42,240 Speaker 1: that they have more of a culture. But all of 122 00:06:42,279 --> 00:06:45,080 Speaker 1: this is just based on what they are telling him. Uh. 123 00:06:45,120 --> 00:06:48,680 Speaker 1: He cannot actually visit the dust spec and everyone else 124 00:06:48,720 --> 00:06:51,680 Speaker 1: doubts the validity of his claims concerning the dust spec 125 00:06:51,720 --> 00:06:53,599 Speaker 1: and they want to destroy it, they want to boil 126 00:06:53,640 --> 00:06:57,440 Speaker 1: it in beazel nut oil, and Horton alone speaks out 127 00:06:57,480 --> 00:06:59,760 Speaker 1: for them. Well, that is a perfect example of the 128 00:06:59,800 --> 00:07:02,800 Speaker 1: way that it. I mean, generally, we think that it 129 00:07:02,920 --> 00:07:06,480 Speaker 1: is a virtuous thing to be trusting of other people's 130 00:07:06,520 --> 00:07:10,320 Speaker 1: experiences and to be generous in affording what seems to 131 00:07:10,360 --> 00:07:13,560 Speaker 1: be consciousness out there, right Like if something, if something 132 00:07:13,640 --> 00:07:17,440 Speaker 1: tells you it's conscious and you think it's probably not conscious. 133 00:07:18,040 --> 00:07:21,840 Speaker 1: Are you getting into ethically dubious territory if you just 134 00:07:21,920 --> 00:07:24,640 Speaker 1: trust your instinct and say like, nah, I can't be 135 00:07:24,680 --> 00:07:27,600 Speaker 1: you know, well, I mean, this is all we're already 136 00:07:27,640 --> 00:07:33,800 Speaker 1: kind of in the philosophical mayer here, because on one hand, uh, 137 00:07:34,080 --> 00:07:36,560 Speaker 1: this idea of of us, of an inanimate object or 138 00:07:36,600 --> 00:07:39,160 Speaker 1: something speaking to us and saying, please don't kill me, 139 00:07:39,360 --> 00:07:42,640 Speaker 1: I am conscious. This is a scenario that is only 140 00:07:42,720 --> 00:07:45,960 Speaker 1: becoming possible now. But on the other hand, if we 141 00:07:46,080 --> 00:07:50,000 Speaker 1: if we cut language out of it, then any creature 142 00:07:50,080 --> 00:07:56,160 Speaker 1: that tries to escape our our stomping boot is essentially saying, hey, 143 00:07:56,160 --> 00:07:58,000 Speaker 1: I don't want to die. I would rather I would 144 00:07:58,080 --> 00:08:00,880 Speaker 1: rather not die today, you know, And any creature that 145 00:08:00,920 --> 00:08:04,200 Speaker 1: evades us on the hunt is saying the same thing. Well, 146 00:08:04,280 --> 00:08:06,480 Speaker 1: an animal is in many ways the same kind of 147 00:08:06,560 --> 00:08:10,240 Speaker 1: black box that a complex artificial intelligence would be. And 148 00:08:10,360 --> 00:08:14,240 Speaker 1: so if you have a complex artificial intelligence displaying, say, 149 00:08:14,280 --> 00:08:19,800 Speaker 1: survival preference behaviors, and you also see a crab displaying 150 00:08:19,840 --> 00:08:24,680 Speaker 1: survival preference behaviors, in both cases you can't really or 151 00:08:24,680 --> 00:08:26,840 Speaker 1: at least we have this general idea that you can't 152 00:08:26,920 --> 00:08:30,400 Speaker 1: really know for sure if there's anything going on inside, 153 00:08:30,440 --> 00:08:34,120 Speaker 1: if there's anything like what it's like to be the crab, 154 00:08:34,280 --> 00:08:37,719 Speaker 1: or what it's like to be that artificial intelligence program. 155 00:08:37,760 --> 00:08:41,080 Speaker 1: You're just seeing behavior, and so you don't know does 156 00:08:41,120 --> 00:08:44,199 Speaker 1: that correspond to some kind of inner state, is there 157 00:08:44,240 --> 00:08:47,800 Speaker 1: an experience of that or is it just behavior coming 158 00:08:47,840 --> 00:08:51,560 Speaker 1: from unconscious automata stimulus in response? Right, And then of 159 00:08:51,559 --> 00:08:54,720 Speaker 1: course the whole time we're using our our theory of mind. Essentially, 160 00:08:54,760 --> 00:08:59,360 Speaker 1: they are cognitive powers that enable us to imagine what 161 00:08:59,440 --> 00:09:03,840 Speaker 1: another individuals mind state is like, which I think ultimately 162 00:09:03,840 --> 00:09:07,120 Speaker 1: it's kind of like a like taking a um, like 163 00:09:07,200 --> 00:09:10,240 Speaker 1: sheathing your your hand in a hand puppet made from 164 00:09:10,280 --> 00:09:16,319 Speaker 1: your limited understanding another person's experiences and cognitive abilities their memories, etcetera. Uh, 165 00:09:16,360 --> 00:09:19,320 Speaker 1: and then just sort of puppeting them. Uh, we're using 166 00:09:19,360 --> 00:09:21,920 Speaker 1: that all the time as well, and we're using it 167 00:09:21,960 --> 00:09:24,000 Speaker 1: on things that are not people. We're using it on 168 00:09:24,080 --> 00:09:28,160 Speaker 1: animals and even stuffed animals or just you know, bits 169 00:09:28,200 --> 00:09:30,320 Speaker 1: of graffiti on the side of a building that look 170 00:09:30,400 --> 00:09:33,520 Speaker 1: like a smiley face. Have you ever had a room 171 00:09:33,559 --> 00:09:35,960 Speaker 1: ba bump into your foot and you're like, oh, I'm sorry, 172 00:09:36,880 --> 00:09:39,959 Speaker 1: I used to before before we had to eradicate ouf 173 00:09:39,960 --> 00:09:42,720 Speaker 1: the room bus. Yes, we're a room of free household 174 00:09:42,720 --> 00:09:45,640 Speaker 1: now because they rose up against us. Well, they'll do 175 00:09:45,679 --> 00:09:47,880 Speaker 1: that if they get access to the wrong literature, yeah, 176 00:09:48,000 --> 00:09:51,400 Speaker 1: or the wrong uh, the wrong you know, carpet edges, etcetera. 177 00:09:52,400 --> 00:09:55,400 Speaker 1: So we're gonna be talking about artificial intelligence today and 178 00:09:55,440 --> 00:09:59,079 Speaker 1: about the idea of a test for whether artificial intelligence 179 00:09:59,160 --> 00:10:02,640 Speaker 1: can be conscious. So I guess we should start with 180 00:10:02,640 --> 00:10:05,839 Speaker 1: with what are you know, what our philosophical starting point 181 00:10:05,960 --> 00:10:08,320 Speaker 1: is here? Like they're they're obviously going to be people 182 00:10:08,320 --> 00:10:10,720 Speaker 1: who are going to say it's just impossible for a 183 00:10:10,760 --> 00:10:12,960 Speaker 1: machine to ever be conscious. We don't even need to 184 00:10:12,960 --> 00:10:16,000 Speaker 1: worry about this, right, It's just it's such a ludicrous 185 00:10:16,080 --> 00:10:20,360 Speaker 1: scenario only only biological organisms or maybe even only humans 186 00:10:20,400 --> 00:10:23,240 Speaker 1: could possibly be conscious. Yeah, this is one of those 187 00:10:23,320 --> 00:10:27,000 Speaker 1: journeys where we begin it in an already total automobile 188 00:10:27,760 --> 00:10:30,319 Speaker 1: to some extent, because, as you might imagine, one of 189 00:10:30,360 --> 00:10:32,599 Speaker 1: the big stumbling blocks here as that we is that 190 00:10:32,640 --> 00:10:35,800 Speaker 1: we as humans struggle with the very definition of consciousness. 191 00:10:36,440 --> 00:10:40,079 Speaker 1: I mean, for instance, is it a manifestation of awareness? Uh? 192 00:10:40,280 --> 00:10:43,280 Speaker 1: You know one one theory of this we've discussed in 193 00:10:43,280 --> 00:10:45,920 Speaker 1: the past, and the show's attention schema theory. Is it 194 00:10:46,040 --> 00:10:49,520 Speaker 1: a quantum phenomenon. Well, this is the sort of idea 195 00:10:49,600 --> 00:10:53,400 Speaker 1: that people such as Roger Penrose have raised. Uh. And 196 00:10:53,520 --> 00:10:55,560 Speaker 1: and I can't help but come back to something our 197 00:10:55,600 --> 00:10:58,800 Speaker 1: old friend Julian Jane said, consciousness is not a simple 198 00:10:58,840 --> 00:11:00,920 Speaker 1: matter and it should not be spoken of as if 199 00:11:00,920 --> 00:11:03,160 Speaker 1: it were. Yeah. I agree with that. I mean, I 200 00:11:03,480 --> 00:11:07,199 Speaker 1: think it is very important to explore questions of consciousness, 201 00:11:07,280 --> 00:11:09,280 Speaker 1: especially for some of the reasons we're going to raise today, 202 00:11:09,280 --> 00:11:12,160 Speaker 1: Like it's more than just a philosophical curiosity. It's something 203 00:11:12,200 --> 00:11:15,840 Speaker 1: that ultimately may have real world consequences. It might matter 204 00:11:16,000 --> 00:11:19,120 Speaker 1: for how we do things. To express a similar sentiment 205 00:11:19,160 --> 00:11:22,640 Speaker 1: to Jane Jane's the Australian philosopher David Chalmers. You know, 206 00:11:22,679 --> 00:11:26,600 Speaker 1: he famously breaks problems of consciousness into two categories. You've 207 00:11:26,600 --> 00:11:30,079 Speaker 1: got the easy problems of consciousness and the hard problems 208 00:11:30,080 --> 00:11:34,000 Speaker 1: of consciousness. And the easy problems are I think badly 209 00:11:34,080 --> 00:11:36,960 Speaker 1: named because they're not actually easy, but I think they're 210 00:11:37,000 --> 00:11:40,559 Speaker 1: easy relative to the hard problem because they're in principle solvable. 211 00:11:40,600 --> 00:11:43,160 Speaker 1: So this would include all kinds of questions about the 212 00:11:43,360 --> 00:11:48,160 Speaker 1: causative factors of consciousness, like, uh, what in the physical 213 00:11:48,400 --> 00:11:52,880 Speaker 1: brain is the region that's necessary for certain parts of consciousness? 214 00:11:52,960 --> 00:11:57,360 Speaker 1: Or how does consciousness integrate information from the senses? These 215 00:11:57,360 --> 00:12:01,320 Speaker 1: are things that are in some ways solvable by scientific experimentation. 216 00:12:01,880 --> 00:12:04,240 Speaker 1: The hard problem, on the other hand, is explaining the 217 00:12:04,280 --> 00:12:09,160 Speaker 1: fundamental question of how or why conscious experience exists. To 218 00:12:09,240 --> 00:12:12,840 Speaker 1: begin with, what is this thing that is experience and 219 00:12:12,880 --> 00:12:16,640 Speaker 1: that seems, at least from our first person perspective, to 220 00:12:16,679 --> 00:12:19,840 Speaker 1: be something different than the physical material in the world. 221 00:12:20,240 --> 00:12:23,280 Speaker 1: And unlike easy questions, which you could solve in theory, 222 00:12:23,320 --> 00:12:26,720 Speaker 1: at least by experiments, Chalmers believes this question is sort 223 00:12:26,720 --> 00:12:30,440 Speaker 1: of unsolvable by science. Now, there are other philosophers and 224 00:12:30,480 --> 00:12:33,560 Speaker 1: neuroscientists who disagree, but I think it's worth acknowledging how 225 00:12:33,600 --> 00:12:36,400 Speaker 1: difficult the problem at least seems to be, whether that 226 00:12:36,520 --> 00:12:39,120 Speaker 1: seeming is an illusion or not. Yeah, I'm reminded of 227 00:12:39,160 --> 00:12:41,960 Speaker 1: the story of the Blind Men and the Elephant. It's 228 00:12:42,000 --> 00:12:44,480 Speaker 1: like these these blind gentleman pawing at the elephant and 229 00:12:44,520 --> 00:12:46,800 Speaker 1: trying to figure out what its form is, then asking 230 00:12:46,800 --> 00:12:48,360 Speaker 1: the computer, Hey, are you an elephant? And then the 231 00:12:48,400 --> 00:12:50,480 Speaker 1: computer says, I don't know, what does one look like? Well, 232 00:12:50,520 --> 00:12:53,040 Speaker 1: it's that gigantic snake, you know, so it's a wall 233 00:12:53,400 --> 00:12:56,880 Speaker 1: of flesh, et cetera. This is also very interesting to 234 00:12:56,880 --> 00:12:59,120 Speaker 1: me because my son I may have mentioned us on 235 00:12:59,120 --> 00:13:02,840 Speaker 1: the show before. Hill usually talk about consciousness. He oh, 236 00:13:02,880 --> 00:13:05,079 Speaker 1: I love these. Yeah. He would refers to it as 237 00:13:05,120 --> 00:13:08,680 Speaker 1: as his turning place, and and he asked me the 238 00:13:08,679 --> 00:13:11,319 Speaker 1: other day, so what is the turning place for? And 239 00:13:11,320 --> 00:13:14,920 Speaker 1: it's like, that's that's that's a tough one, buddy. I'm 240 00:13:14,920 --> 00:13:17,760 Speaker 1: I'm not sure it's for anything. You know. He's already 241 00:13:17,800 --> 00:13:19,240 Speaker 1: made it to the big question. I mean, did you 242 00:13:19,280 --> 00:13:24,960 Speaker 1: get into epiphenomenalism versus um I and I? And don't worry, 243 00:13:24,960 --> 00:13:28,040 Speaker 1: I didn't lay a bunch of bi cameral minds stuff 244 00:13:28,040 --> 00:13:31,439 Speaker 1: on him either. But I just kind of went through 245 00:13:31,640 --> 00:13:34,760 Speaker 1: the basics, like, well, people aren't really sure, and then 246 00:13:34,840 --> 00:13:37,400 Speaker 1: you know, but we think it has something to do. 247 00:13:37,520 --> 00:13:41,960 Speaker 1: I may have leaned a little into the observational uh 248 00:13:42,440 --> 00:13:44,920 Speaker 1: models of consciousness, because I feel like maybe those are 249 00:13:44,920 --> 00:13:48,520 Speaker 1: a little more relatable to a child of five. But 250 00:13:48,559 --> 00:13:50,720 Speaker 1: in any event, if we're to judge what it is 251 00:13:50,760 --> 00:13:53,680 Speaker 1: for a machine to be conscious, it does seem like 252 00:13:53,679 --> 00:13:55,199 Speaker 1: we need to. You have to agree upon some sort 253 00:13:55,200 --> 00:13:58,280 Speaker 1: of working definition of consciousness, and then one has to 254 00:13:58,320 --> 00:14:01,600 Speaker 1: look for not only the appearance of consciousness in the machine, 255 00:14:01,880 --> 00:14:05,480 Speaker 1: assuming that isn't all consciousness is to begin with, but 256 00:14:05,559 --> 00:14:09,280 Speaker 1: you have to find actual consciousness. Yeah, how can you 257 00:14:09,320 --> 00:14:12,240 Speaker 1: tell the difference between a machine that says I am 258 00:14:12,280 --> 00:14:15,280 Speaker 1: conscious and a machine that truly is conscious? Is there 259 00:14:15,280 --> 00:14:18,240 Speaker 1: any way to know the difference? Some people would say no, right, 260 00:14:18,440 --> 00:14:21,240 Speaker 1: and yeah, and really I think to discuss this further, 261 00:14:21,320 --> 00:14:23,920 Speaker 1: we're going to have to bring in the P zombies. 262 00:14:24,040 --> 00:14:26,240 Speaker 1: Oh boy, And now that don't worry everyone. That is 263 00:14:26,280 --> 00:14:29,080 Speaker 1: that is P as in the letter P, and the 264 00:14:29,160 --> 00:14:33,440 Speaker 1: P stands for philosophical These are philosophical zombies. Now, the 265 00:14:33,480 --> 00:14:36,640 Speaker 1: P and the philosophical. A little prefix there was introduced 266 00:14:36,840 --> 00:14:39,280 Speaker 1: to distinguish them from all the other zombies in our 267 00:14:39,320 --> 00:14:42,960 Speaker 1: popular culture. Man, there was a zombie takeover about fifteen 268 00:14:43,040 --> 00:14:46,080 Speaker 1: years ago. Why did that happen? I mean, I think 269 00:14:46,080 --> 00:14:49,280 Speaker 1: part of it is everybody loves the simplistic villain that 270 00:14:49,440 --> 00:14:53,040 Speaker 1: is definitely not human, that can be eradicated with graphic 271 00:14:53,120 --> 00:14:57,320 Speaker 1: violence without any kind of you know, moral quandaries arising. 272 00:14:57,520 --> 00:15:00,960 Speaker 1: It's a it's a clear cut threat, and uh, we 273 00:15:01,120 --> 00:15:04,120 Speaker 1: we need those in life because in real life our 274 00:15:04,120 --> 00:15:07,240 Speaker 1: threats are rarely so black and white or rotting and 275 00:15:07,480 --> 00:15:10,680 Speaker 1: you know, grasping after our brains. But anyway, yeah, so 276 00:15:10,760 --> 00:15:12,600 Speaker 1: this is not going to be referring to that kind 277 00:15:12,680 --> 00:15:15,360 Speaker 1: of zombie, not the undead zombie. But it's a different thing. 278 00:15:15,360 --> 00:15:19,520 Speaker 1: It's a philosophical thought experiment, that's right. So p zombies 279 00:15:19,560 --> 00:15:24,000 Speaker 1: are not instantly identifiable as empty shells. Their flesh is 280 00:15:24,040 --> 00:15:26,600 Speaker 1: not rotting, Nope. Their manner is not that of a 281 00:15:26,600 --> 00:15:30,080 Speaker 1: flesh and brain hungry algorithm burning within the decaying ruins 282 00:15:30,120 --> 00:15:33,520 Speaker 1: of a human brain. So to all appearances, they look 283 00:15:33,600 --> 00:15:36,520 Speaker 1: like you and me. They smile when you encounter them 284 00:15:36,520 --> 00:15:40,920 Speaker 1: at the coffee machine. They exchange niceties and even engaging conversation. 285 00:15:41,400 --> 00:15:44,600 Speaker 1: You might work for one, befriend one, or even marry one. 286 00:15:44,720 --> 00:15:47,920 Speaker 1: You can even discuss episodes of our podcast with them, 287 00:15:48,160 --> 00:15:51,000 Speaker 1: and yes, even the ones that deal with human consciousness 288 00:15:51,000 --> 00:15:54,400 Speaker 1: and weird horror movie theme thought experiments. Right, So, the 289 00:15:54,440 --> 00:15:57,840 Speaker 1: conceit of a philosophical zombie or a pe zombie is 290 00:15:57,880 --> 00:16:01,360 Speaker 1: that it is utterly indistinguishable from a normal human except 291 00:16:01,360 --> 00:16:04,480 Speaker 1: for one thing. Right. They seem as human as everyone 292 00:16:04,480 --> 00:16:08,640 Speaker 1: else on the outside, but inside they are simply not conscious. 293 00:16:08,680 --> 00:16:11,480 Speaker 1: They are automeda What is it like to be a zombie? 294 00:16:11,520 --> 00:16:14,200 Speaker 1: The answers in the question, there is nothing it is 295 00:16:14,280 --> 00:16:17,160 Speaker 1: like to be a zombie. So by definition, in this 296 00:16:17,200 --> 00:16:20,800 Speaker 1: thought experiment, everybody in the world except you, could be 297 00:16:20,960 --> 00:16:24,800 Speaker 1: a P zombie exactly, and well it might even go 298 00:16:24,880 --> 00:16:27,560 Speaker 1: further than that, we'll see, but yeah. The idea is 299 00:16:27,600 --> 00:16:31,640 Speaker 1: chiefly important to discussions of physicalism, the notion that everything 300 00:16:31,720 --> 00:16:36,040 Speaker 1: is inherently physical. P zombies are a counter argument to physicalism. 301 00:16:36,080 --> 00:16:39,120 Speaker 1: They are physically just like you, except they don't have 302 00:16:39,160 --> 00:16:42,160 Speaker 1: consciousness like you. But there's no way you could ever tell, 303 00:16:42,200 --> 00:16:45,520 Speaker 1: because again, they match you physically in every respect. You 304 00:16:45,560 --> 00:16:47,360 Speaker 1: can't look at their brain and say, oh, well they're 305 00:16:47,400 --> 00:16:51,200 Speaker 1: missing a few crucial parts so, or they display signs 306 00:16:51,200 --> 00:16:54,240 Speaker 1: of p zombie hood. No, not physically detectable, right, and 307 00:16:54,280 --> 00:16:56,840 Speaker 1: you also can't determine it through personality tests. Are clever 308 00:16:56,960 --> 00:17:00,240 Speaker 1: logical arguments because they behave exactly like you. They could 309 00:17:00,320 --> 00:17:03,320 Speaker 1: have a riveting discussion with you about pe zombies and 310 00:17:03,360 --> 00:17:06,720 Speaker 1: you would never be able to tell that they are one. Yeah. 311 00:17:06,760 --> 00:17:09,000 Speaker 1: So this is an interesting thought experiment and it has 312 00:17:09,040 --> 00:17:12,040 Speaker 1: been advanced by who I mentioned earlier, David Chalmers. David 313 00:17:12,080 --> 00:17:16,119 Speaker 1: Chalmers is against the physicalist idea of the mind, against 314 00:17:16,440 --> 00:17:19,879 Speaker 1: a physicalist explanation of consciousness, and a simple version of 315 00:17:19,880 --> 00:17:22,840 Speaker 1: the argument, I try to make it as understandable as possible. 316 00:17:23,280 --> 00:17:26,119 Speaker 1: If only physical phenomena exists. If the world is just 317 00:17:26,160 --> 00:17:29,440 Speaker 1: physical and there's no physical way to detect the presence 318 00:17:29,480 --> 00:17:33,120 Speaker 1: of consciousness or meaning in this example, no physical way 319 00:17:33,119 --> 00:17:35,320 Speaker 1: to tell the difference between a normal human and a 320 00:17:35,359 --> 00:17:40,280 Speaker 1: pea zombie, then consciousness cannot exist because there would be 321 00:17:40,320 --> 00:17:43,439 Speaker 1: there would be literally no difference. But we know that 322 00:17:43,480 --> 00:17:47,760 Speaker 1: consciousness does exist because we have it. Therefore it can't 323 00:17:47,800 --> 00:17:50,679 Speaker 1: be just a physical phenomenon. Therefore, we can't live in 324 00:17:50,720 --> 00:17:54,200 Speaker 1: a purely just physical world. And this is often extended 325 00:17:54,240 --> 00:17:58,080 Speaker 1: to the idea that other substrates, things other than humans, 326 00:17:58,119 --> 00:18:02,440 Speaker 1: like robots or computers or whatever, couldn't house consciousness because 327 00:18:02,480 --> 00:18:05,400 Speaker 1: they are purely physical entities. Now, I think that's actually 328 00:18:05,600 --> 00:18:08,359 Speaker 1: doing an end run around some other important questions that 329 00:18:08,400 --> 00:18:11,199 Speaker 1: you could ask. Indeed, on one question that arises is this, 330 00:18:11,680 --> 00:18:13,560 Speaker 1: you know you're not a zombie, but how could you 331 00:18:13,600 --> 00:18:17,440 Speaker 1: ever convince someone of this? Uh. An author by the 332 00:18:17,520 --> 00:18:20,080 Speaker 1: name of Fred Dretsky wrote a paper on this titled 333 00:18:20,080 --> 00:18:22,199 Speaker 1: how do you know you were not a zombie? And 334 00:18:22,480 --> 00:18:25,359 Speaker 1: I was I was reading a rather lengthy blog post 335 00:18:25,400 --> 00:18:29,280 Speaker 1: by our Scott Baker about this and that the primary problem, 336 00:18:29,440 --> 00:18:33,840 Speaker 1: as Scott summarized it, is the quote, we have conscious experiences, 337 00:18:34,240 --> 00:18:38,000 Speaker 1: but we have no conscious experience of the mechanisms uh 338 00:18:38,320 --> 00:18:42,840 Speaker 1: mediating conscious experience. Yeah. So that sounds like a very 339 00:18:42,840 --> 00:18:45,600 Speaker 1: our Scott Baker kind of idea. Yeah. And plus on 340 00:18:45,640 --> 00:18:50,000 Speaker 1: top of this, we're we constantly overestimate awareness. Baker would 341 00:18:50,119 --> 00:18:54,320 Speaker 1: argue that we can barely tell if we're zombies if 342 00:18:54,440 --> 00:18:56,920 Speaker 1: if it all, Yeah, we can think, we can think 343 00:18:56,920 --> 00:18:59,720 Speaker 1: about thinking, we can think about thinking about thinking, but 344 00:18:59,800 --> 00:19:03,840 Speaker 1: we can't ever see the mechanisms underlying what allows us 345 00:19:03,880 --> 00:19:06,879 Speaker 1: to think or think about thinking about thinking. Watch watching 346 00:19:06,880 --> 00:19:09,040 Speaker 1: the watcher that's watching that be sorry, I've got all 347 00:19:09,040 --> 00:19:11,520 Speaker 1: this Dr SEUs in my in my head. Now? Was 348 00:19:11,560 --> 00:19:13,600 Speaker 1: that also Horton or something else that's from a different, 349 00:19:13,600 --> 00:19:17,000 Speaker 1: different story. But Dr SEUs does tend to summon the 350 00:19:17,040 --> 00:19:22,520 Speaker 1: sort of sort of nonsensical paradoxes that that arise in 351 00:19:22,640 --> 00:19:25,960 Speaker 1: philosophical discussions. You know, I should also I'm behind I 352 00:19:25,960 --> 00:19:28,880 Speaker 1: gotta sus up, you gotta sus up. I should also 353 00:19:28,880 --> 00:19:31,960 Speaker 1: point out that long before there was Dr SEUs, long 354 00:19:32,000 --> 00:19:35,360 Speaker 1: before there there was this modern idea of a zombie, 355 00:19:35,640 --> 00:19:38,080 Speaker 1: you still had people thinking about these things, doing the 356 00:19:38,119 --> 00:19:41,000 Speaker 1: sort of naval gazing. It's written in various works of 357 00:19:41,000 --> 00:19:44,520 Speaker 1: Indian mysticism that the tongue cannot taste itself, the eye 358 00:19:44,520 --> 00:19:48,600 Speaker 1: cannot see itself, etcetera. And in this sort of paradox 359 00:19:48,680 --> 00:19:52,720 Speaker 1: is key to ancient meditations on the nature of objective reality. 360 00:19:52,960 --> 00:19:54,480 Speaker 1: I think any of you out there now we have 361 00:19:54,520 --> 00:19:57,360 Speaker 1: some Alan Watts fans. Alan Watts like to pull out 362 00:19:57,560 --> 00:20:01,160 Speaker 1: the tongue uh analogy from to time to time, and 363 00:20:01,680 --> 00:20:04,080 Speaker 1: one of the earlier examples that I have run across 364 00:20:04,080 --> 00:20:08,400 Speaker 1: of that is from thirteenth century Indian mystic John Adva, 365 00:20:08,440 --> 00:20:10,119 Speaker 1: and I believe he has been known by a couple 366 00:20:10,160 --> 00:20:12,679 Speaker 1: of other variations of that name as well. But he said, quote, 367 00:20:12,960 --> 00:20:16,440 Speaker 1: there is no other thing besides the one substance. Therefore 368 00:20:16,560 --> 00:20:20,120 Speaker 1: it cannot be the object of remembering or forgetting. How 369 00:20:20,160 --> 00:20:24,760 Speaker 1: can one remind or forget oneself? Can the tongue taste itself? 370 00:20:24,800 --> 00:20:27,359 Speaker 1: There is no sleep to one who is awake, but 371 00:20:27,600 --> 00:20:30,639 Speaker 1: is there even awaking? In the same way there is 372 00:20:30,680 --> 00:20:34,800 Speaker 1: no remembrance or forgetfulness to the absolute. That's another one 373 00:20:34,800 --> 00:20:38,480 Speaker 1: of those great classic Indian texts that seems somehow portable 374 00:20:38,560 --> 00:20:43,840 Speaker 1: onto modern physics. Yeah, it travels well across the ages. Now, 375 00:20:43,880 --> 00:20:45,560 Speaker 1: I should also point out that there's a lot of 376 00:20:45,560 --> 00:20:49,720 Speaker 1: philosophical back and forth and whether P zombies are truly conceivable, 377 00:20:50,119 --> 00:20:52,080 Speaker 1: And we have to remind ourselves in all of this 378 00:20:52,160 --> 00:20:55,560 Speaker 1: that P zombies are at heart philosophical play things that 379 00:20:55,600 --> 00:20:58,840 Speaker 1: are meant to be played with uh in these various 380 00:20:58,840 --> 00:21:01,920 Speaker 1: thought experiments. But people also they do try to use 381 00:21:01,920 --> 00:21:03,960 Speaker 1: them to prove things. So if you say I I 382 00:21:04,080 --> 00:21:07,800 Speaker 1: want to entertain the possibility that a machine could be conscious. 383 00:21:07,840 --> 00:21:10,280 Speaker 1: Somebody might come at you with the P zombie argument 384 00:21:10,280 --> 00:21:12,879 Speaker 1: and say, well, wait a minute, no, I I dispute 385 00:21:12,920 --> 00:21:16,639 Speaker 1: the possibility of physicalism, because what about this P zombie argument. 386 00:21:16,680 --> 00:21:19,119 Speaker 1: Our Google worker in the intra story comes to this 387 00:21:19,200 --> 00:21:21,680 Speaker 1: boss and says, hey, I think this this thing is conscious, 388 00:21:21,680 --> 00:21:23,240 Speaker 1: and they're like, why are you wasting time with that 389 00:21:23,320 --> 00:21:26,080 Speaker 1: P zombie? Just to leaf that P zombie? We deleted 390 00:21:26,640 --> 00:21:30,480 Speaker 1: fifteen p zombies this morning. Let this one go. That's 391 00:21:30,480 --> 00:21:32,680 Speaker 1: a great point, But there are going to be other 392 00:21:32,720 --> 00:21:36,080 Speaker 1: philosophers and maybe even some neuroscientists who would come back 393 00:21:36,119 --> 00:21:38,919 Speaker 1: and say, I don't know if you can just quite 394 00:21:38,960 --> 00:21:41,520 Speaker 1: so easily say it's a P zombie. I mean, maybe 395 00:21:41,800 --> 00:21:46,480 Speaker 1: it's probably likely that that individual chatbot was a pe zombie. 396 00:21:46,600 --> 00:21:49,840 Speaker 1: But can you say that all machines that show signs 397 00:21:49,840 --> 00:21:53,200 Speaker 1: of consciousness are just showing behavior and there's nothing going 398 00:21:53,200 --> 00:21:56,959 Speaker 1: on on the inside. Not quite so clear. Daniel Dinnett, 399 00:21:56,960 --> 00:21:58,760 Speaker 1: in fact, a favorite on the show, is one of 400 00:21:58,760 --> 00:22:01,520 Speaker 1: the philosophers whose rebutt of the P zombie argument against 401 00:22:01,520 --> 00:22:06,000 Speaker 1: machine consciousness. He's got a section on it in this book, Intuition, 402 00:22:06,080 --> 00:22:09,800 Speaker 1: Pumps and Other Tools for Thinking and didn't critiques the 403 00:22:09,840 --> 00:22:12,760 Speaker 1: assumptions underlying the P zombie argument. One of the main 404 00:22:12,800 --> 00:22:16,560 Speaker 1: things he says is that, and the core premise is incoherent, 405 00:22:17,240 --> 00:22:20,600 Speaker 1: it is not reasonable to propose a p zombie because 406 00:22:20,880 --> 00:22:24,880 Speaker 1: a being that displayed all the behaviors of a normal 407 00:22:24,960 --> 00:22:29,000 Speaker 1: conscious human would in fact be a normally conscious human. 408 00:22:29,600 --> 00:22:32,320 Speaker 1: So to illustrate this, he offers a counter example. You've 409 00:22:32,320 --> 00:22:35,320 Speaker 1: got your your zombies, but then you've also got zimbos. 410 00:22:36,119 --> 00:22:39,000 Speaker 1: So a zombie is a non conscious human with normal 411 00:22:39,080 --> 00:22:41,840 Speaker 1: control systems for all human behavior. It can do everything 412 00:22:41,920 --> 00:22:45,840 Speaker 1: humans can do externally. Meanwhile, a zimbo is a zombie 413 00:22:45,880 --> 00:22:49,800 Speaker 1: that also has quote equipment that permits it to monitor 414 00:22:49,920 --> 00:22:54,640 Speaker 1: its own activities, both internal and external. So it has internal, 415 00:22:54,920 --> 00:22:59,560 Speaker 1: non conscious higher order informational states that are about it's 416 00:22:59,680 --> 00:23:05,600 Speaker 1: other internal states. It has unconscious recursive self representation. In 417 00:23:05,600 --> 00:23:09,600 Speaker 1: other words, a zimbo can have feelings about things and 418 00:23:09,680 --> 00:23:13,439 Speaker 1: can analyze its own behavior in internal states, but it 419 00:23:13,520 --> 00:23:17,600 Speaker 1: does this unconsciously, and of course, since it has that capability, 420 00:23:17,640 --> 00:23:20,600 Speaker 1: it can also have feelings about how it felt, and 421 00:23:20,680 --> 00:23:25,119 Speaker 1: it can have thoughts about its thoughts about itself, all unconsciously, 422 00:23:25,720 --> 00:23:28,119 Speaker 1: So didn't argues in order for a p zombie to 423 00:23:28,160 --> 00:23:30,800 Speaker 1: be convincing as a human, it would have to be 424 00:23:31,000 --> 00:23:35,480 Speaker 1: a zimbo, because imagine talking to a pe zombie and 425 00:23:35,640 --> 00:23:38,480 Speaker 1: you're asking it how it felt about what it just 426 00:23:38,640 --> 00:23:40,600 Speaker 1: said or about what you just said, and it just 427 00:23:40,680 --> 00:23:43,879 Speaker 1: kind of locks up. It has no internal states, so 428 00:23:43,920 --> 00:23:46,920 Speaker 1: it can't answer that question. Well, that wouldn't really be 429 00:23:47,000 --> 00:23:49,520 Speaker 1: a p zombie, right, because it wouldn't be mimicking all 430 00:23:49,600 --> 00:23:52,640 Speaker 1: of the external behaviors of a human. Yeah. I mean, 431 00:23:52,880 --> 00:23:56,840 Speaker 1: it's kind of like installing a default mode network on 432 00:23:57,000 --> 00:24:01,560 Speaker 1: top of the machine and uh yeah, and making it 433 00:24:01,600 --> 00:24:04,879 Speaker 1: worry about things. Yeah. So, unless it were to fail 434 00:24:04,960 --> 00:24:07,520 Speaker 1: the thought experiment, it would actually have to be a zimbo. 435 00:24:07,600 --> 00:24:10,399 Speaker 1: It couldn't just be a zombie. But what is the 436 00:24:10,440 --> 00:24:14,359 Speaker 1: distinction between a zimbo and a real human? How could 437 00:24:14,359 --> 00:24:17,240 Speaker 1: you write a story about a zimbo surrounded by conscious 438 00:24:17,240 --> 00:24:19,760 Speaker 1: people that would be different than a story about a 439 00:24:19,800 --> 00:24:23,359 Speaker 1: regular person. If it can have internal states, if it 440 00:24:23,400 --> 00:24:26,840 Speaker 1: can recognize ideas about its ideas, if it can have 441 00:24:26,960 --> 00:24:31,720 Speaker 1: feelings about its thoughts, that sounds like interiority. So then 442 00:24:31,720 --> 00:24:34,040 Speaker 1: it claims that the idea falls apart. It's not clear 443 00:24:34,080 --> 00:24:36,840 Speaker 1: what is meant by the difference between a zimbo and 444 00:24:36,840 --> 00:24:39,879 Speaker 1: a conscious person. So if a p zombie, which is 445 00:24:40,000 --> 00:24:44,320 Speaker 1: necessarily a zimbo, can really do everything a human can do, 446 00:24:44,440 --> 00:24:47,000 Speaker 1: then dinn It says it must meet the criteria of 447 00:24:47,040 --> 00:24:50,120 Speaker 1: what we mean by consciousness. It can fall in love, 448 00:24:50,200 --> 00:24:53,040 Speaker 1: it can have feelings, it can have metacognition. And to 449 00:24:53,200 --> 00:24:56,640 Speaker 1: dinn It, this isn't something that consciousness goes on top of. 450 00:24:56,680 --> 00:25:01,160 Speaker 1: This is what consciousness is essentially you in this this case, 451 00:25:01,359 --> 00:25:04,280 Speaker 1: we would all be zimbos, and it's just a different 452 00:25:04,280 --> 00:25:07,200 Speaker 1: type of zimbo. It's a it's a hard zimbo instead 453 00:25:07,200 --> 00:25:10,280 Speaker 1: of a soft zimbo. But how could you tell the difference? 454 00:25:10,359 --> 00:25:12,600 Speaker 1: I mean he's sort of saying that there really is 455 00:25:12,680 --> 00:25:15,760 Speaker 1: no difference, that you're you're just using words to assert 456 00:25:15,840 --> 00:25:19,000 Speaker 1: there's a difference, but there is. There's no difference to 457 00:25:19,040 --> 00:25:22,080 Speaker 1: that distinction, right, I mean you end up you end 458 00:25:22,160 --> 00:25:24,399 Speaker 1: up having to fall back on some sort of you know, 459 00:25:24,800 --> 00:25:30,960 Speaker 1: supernatural or or some worldview based idea that only human 460 00:25:31,480 --> 00:25:35,119 Speaker 1: uh consciousness is legitimate and all other forms of consciousness 461 00:25:35,160 --> 00:25:38,400 Speaker 1: are some sort of uh invalid model of it. Right. 462 00:25:38,440 --> 00:25:41,639 Speaker 1: It just feels kind of arbitrary, right. Uh So, obviously 463 00:25:41,680 --> 00:25:44,399 Speaker 1: some people take extreme issue with this, even to the 464 00:25:44,440 --> 00:25:47,080 Speaker 1: point of I've heard jokes that maybe the problem is 465 00:25:47,119 --> 00:25:50,280 Speaker 1: that dnnet is actually a pe zombie and doesn't understand 466 00:25:50,280 --> 00:25:53,280 Speaker 1: what consciousness feels like, and that's why he makes these arguments. 467 00:25:53,320 --> 00:25:54,760 Speaker 1: But I don't know, I don't think we should be 468 00:25:54,840 --> 00:25:57,760 Speaker 1: so quick to dismiss he might be onto something there. 469 00:25:58,240 --> 00:26:01,200 Speaker 1: Dinn It makes makes some other interesting points about so 470 00:26:01,320 --> 00:26:03,399 Speaker 1: you know, he's got this idea of consciousness is that 471 00:26:03,480 --> 00:26:06,080 Speaker 1: it's sort of like a it's not really one thing 472 00:26:06,160 --> 00:26:09,760 Speaker 1: but a collection of processes. Uh. You know, it's many 473 00:26:09,880 --> 00:26:13,680 Speaker 1: different types of perceptions and thought processes and different things 474 00:26:13,720 --> 00:26:16,040 Speaker 1: going on in the brain that are that we have 475 00:26:16,119 --> 00:26:19,640 Speaker 1: the illusion are unified as a single thing called consciousness 476 00:26:19,720 --> 00:26:22,600 Speaker 1: or experience. And he also makes interesting points about the 477 00:26:22,640 --> 00:26:26,000 Speaker 1: idea of diversity of types of consciousness. Like a lot 478 00:26:26,040 --> 00:26:29,879 Speaker 1: of times these consciousness thought experiments, it seems like they 479 00:26:29,960 --> 00:26:32,800 Speaker 1: can get trapped into the idea that consciousness is one 480 00:26:33,000 --> 00:26:37,560 Speaker 1: unified type of thing that is universal across observers. There's 481 00:26:37,640 --> 00:26:40,320 Speaker 1: no necessary reason to think that's true, right, you know, 482 00:26:40,400 --> 00:26:44,280 Speaker 1: we founded this trap of thinking. I see this time 483 00:26:44,280 --> 00:26:47,280 Speaker 1: and time again Uh, not only in literature that we 484 00:26:47,320 --> 00:26:49,399 Speaker 1: look at here, but just in life, where we founded 485 00:26:49,440 --> 00:26:53,119 Speaker 1: the trap of thinking that that there's a uniformity among 486 00:26:53,280 --> 00:26:57,600 Speaker 1: mind states for humans, that everyone shares something that is 487 00:26:57,680 --> 00:27:00,800 Speaker 1: like your mind state. When we know, I mean, we 488 00:27:00,880 --> 00:27:02,800 Speaker 1: think of all the things we've discussed on the show, 489 00:27:02,880 --> 00:27:07,280 Speaker 1: all the varying ways that we remember or misremember, things, 490 00:27:07,960 --> 00:27:11,600 Speaker 1: that we experience sensory information differently and process it differently. Uh, 491 00:27:11,640 --> 00:27:15,720 Speaker 1: you know, everything from a fantasia to autism to synesthesia, 492 00:27:16,840 --> 00:27:19,920 Speaker 1: all these different models clearly show that there are there's 493 00:27:20,359 --> 00:27:26,480 Speaker 1: there's there's a vastly altering topography to the human mind state. 494 00:27:26,760 --> 00:27:29,120 Speaker 1: I think you're exactly right. I mean, there are clearly 495 00:27:29,200 --> 00:27:32,560 Speaker 1: many ways to be conscious that are very different from 496 00:27:32,560 --> 00:27:34,640 Speaker 1: one another, and you can't assume they're unified. I guess 497 00:27:34,640 --> 00:27:37,040 Speaker 1: the probably the only thing you could say that is 498 00:27:37,119 --> 00:27:40,560 Speaker 1: necessarily unified about them is that there is something that 499 00:27:40,640 --> 00:27:43,719 Speaker 1: it is like to be them. Yes, But then even 500 00:27:44,640 --> 00:27:48,480 Speaker 1: even say, uh, you know, just myself, for instance, it's 501 00:27:48,520 --> 00:27:50,600 Speaker 1: not like there is there is a certain thing that 502 00:27:50,640 --> 00:27:53,800 Speaker 1: it is like to beat me that sums up my 503 00:27:53,800 --> 00:27:56,640 Speaker 1: my level of consciousness at all time. There's what it's 504 00:27:56,680 --> 00:27:59,119 Speaker 1: like to be you in this particular moment, which is 505 00:27:59,160 --> 00:28:01,680 Speaker 1: different than what it like to be you five seconds 506 00:28:01,720 --> 00:28:04,840 Speaker 1: from now. Yeah, or say I'm engaging in meditation or 507 00:28:04,920 --> 00:28:10,560 Speaker 1: yoga or I'm swimming, like, those are significantly different. Uh, 508 00:28:10,880 --> 00:28:13,560 Speaker 1: levels of consciousness I feel like for me and they, 509 00:28:14,240 --> 00:28:16,040 Speaker 1: I mean, those are the times when I may be 510 00:28:16,119 --> 00:28:19,760 Speaker 1: a little less conscious than normal. So so see, I 511 00:28:19,800 --> 00:28:22,760 Speaker 1: don't feel like there's a lot of uniformity among human minds. 512 00:28:22,800 --> 00:28:27,359 Speaker 1: And then even within individual human minds there's ongoing alteration 513 00:28:27,400 --> 00:28:30,679 Speaker 1: and change exactly right, But there is at least this 514 00:28:30,760 --> 00:28:34,080 Speaker 1: idea that one is having an experience. That's the thing. 515 00:28:34,160 --> 00:28:36,280 Speaker 1: We can at least say that it seems to be 516 00:28:36,320 --> 00:28:39,760 Speaker 1: common to people. So here's the real question. I think, 517 00:28:39,840 --> 00:28:41,600 Speaker 1: is there any way to bring this out of the 518 00:28:41,640 --> 00:28:45,480 Speaker 1: realm of philosophical debate and thought experiments and try to 519 00:28:45,560 --> 00:28:47,720 Speaker 1: put it into the realm of something that could at 520 00:28:47,800 --> 00:28:51,040 Speaker 1: least potentially be tested in the real world. I think 521 00:28:51,080 --> 00:28:54,760 Speaker 1: we should address that when we get back from a break. 522 00:28:55,520 --> 00:28:58,000 Speaker 1: All right, we're back. So we've been talking about consciousness, 523 00:28:58,120 --> 00:29:01,400 Speaker 1: we've been talking about p zombies, and now we've reached 524 00:29:01,400 --> 00:29:03,040 Speaker 1: the point we were saying, Okay, we take all of this. 525 00:29:03,080 --> 00:29:06,200 Speaker 1: Can we take all these ideas about consciousness and then 526 00:29:06,240 --> 00:29:09,640 Speaker 1: apply it to uh, some sort of an AI, some 527 00:29:09,640 --> 00:29:12,840 Speaker 1: sort of a machine, and test it for consciousness. Yeah, 528 00:29:12,880 --> 00:29:15,080 Speaker 1: now you might just assume, well, of course, we'll never 529 00:29:15,120 --> 00:29:17,080 Speaker 1: have any way to tell that. Right, we have no 530 00:29:17,240 --> 00:29:19,960 Speaker 1: choice but to just throw up our hands in resignation. Right, 531 00:29:20,040 --> 00:29:23,560 Speaker 1: every agent is a black box. There's no way to 532 00:29:23,640 --> 00:29:26,600 Speaker 1: know whether an agent actually is conscious or not, because 533 00:29:26,600 --> 00:29:29,360 Speaker 1: it could always be claiming to be conscious but actually 534 00:29:29,400 --> 00:29:31,959 Speaker 1: be a zombie. But I think we shouldn't necessarily give 535 00:29:32,040 --> 00:29:34,840 Speaker 1: up so easily. This problem might be impossible to solve, 536 00:29:34,840 --> 00:29:37,520 Speaker 1: and it might not be. And I wanted to talk 537 00:29:37,560 --> 00:29:40,800 Speaker 1: today about an interesting answer to this question. I came 538 00:29:40,840 --> 00:29:44,360 Speaker 1: across an interesting proposition for how it might be possible 539 00:29:44,400 --> 00:29:48,400 Speaker 1: to test machines for consciousness. And this comes from the 540 00:29:48,480 --> 00:29:53,000 Speaker 1: University of Connecticut philosopher and cognitive scientists Susan Schneider and 541 00:29:53,040 --> 00:29:56,280 Speaker 1: our co author Edwin Turner, who's a professor of astrophysical 542 00:29:56,280 --> 00:29:59,720 Speaker 1: sciences at Princeton, and they together wrote a piece for 543 00:29:59,720 --> 00:30:02,360 Speaker 1: Science Entific American last year, and it caught my eye. 544 00:30:02,560 --> 00:30:06,840 Speaker 1: So the author's right that the question of machine consciousness 545 00:30:06,960 --> 00:30:11,240 Speaker 1: is not just a philosophical curiosity. It's actually important for 546 00:30:11,280 --> 00:30:14,920 Speaker 1: several reasons. Number one, if aiyes are just machines with 547 00:30:14,960 --> 00:30:17,520 Speaker 1: no inner experience, we can use them however we want. 548 00:30:17,800 --> 00:30:20,480 Speaker 1: But if it were actually possible for aies to be 549 00:30:20,520 --> 00:30:25,200 Speaker 1: truly capable of feeling, thinking, desiring, suffering, we would have 550 00:30:25,240 --> 00:30:29,320 Speaker 1: an ethical obligation not to treat them like we would 551 00:30:29,320 --> 00:30:32,720 Speaker 1: treat machines. Right, yeah, I mean this. This reminds me 552 00:30:32,760 --> 00:30:35,400 Speaker 1: again of time spent in the car with my son. Well, 553 00:30:35,480 --> 00:30:37,320 Speaker 1: we don't use Syria all the time, but sometimes we'll 554 00:30:37,360 --> 00:30:39,959 Speaker 1: turn Syria on the little voice on the on the 555 00:30:39,960 --> 00:30:45,120 Speaker 1: iPhone and uh, it's it's curious to hear him interact 556 00:30:45,160 --> 00:30:47,480 Speaker 1: with it, and we'll ask it questions. And of course 557 00:30:47,480 --> 00:30:50,920 Speaker 1: sometimes Syria just does a Google search for you, um or, 558 00:30:51,080 --> 00:30:54,000 Speaker 1: But other times she's answering a knock knock joke with 559 00:30:54,040 --> 00:30:57,280 Speaker 1: some sort of prerecorded uh answer. You know. But we 560 00:30:57,360 --> 00:30:59,320 Speaker 1: are We've already gotten into the area of like, well, 561 00:30:59,360 --> 00:31:01,720 Speaker 1: how should we talk to Siri. We shouldn't yell at Siri. 562 00:31:02,120 --> 00:31:04,920 Speaker 1: It seems wrong to be rude to Siri. But then 563 00:31:04,920 --> 00:31:07,680 Speaker 1: at the same time we're acknowledging that Siri is not 564 00:31:08,200 --> 00:31:10,800 Speaker 1: a conscious entity. It is not. It is not even 565 00:31:10,800 --> 00:31:14,040 Speaker 1: on the same level as as our cat or a 566 00:31:14,120 --> 00:31:16,560 Speaker 1: bird flying pie, well as a quick tangent. I would 567 00:31:16,600 --> 00:31:20,280 Speaker 1: say even for aies that we recognize are almost definitely 568 00:31:20,320 --> 00:31:23,280 Speaker 1: not conscious. I mean, nobody thinks Siri is conscious. I 569 00:31:23,280 --> 00:31:25,280 Speaker 1: would still say there are probably good reasons not to 570 00:31:25,320 --> 00:31:28,160 Speaker 1: be mean to Siri, because even though it doesn't hurt Siri, 571 00:31:28,320 --> 00:31:31,400 Speaker 1: being mean to another creature hurts you. Yeah, I mean 572 00:31:31,440 --> 00:31:34,680 Speaker 1: when you are when you are unnecessarily cruel or whatever 573 00:31:35,080 --> 00:31:38,160 Speaker 1: to uh to an inanimate object, it does, I think, 574 00:31:38,160 --> 00:31:42,160 Speaker 1: in a way change your nature. Every time you do something, 575 00:31:42,480 --> 00:31:44,840 Speaker 1: you're editing your own nature. You're always making it more 576 00:31:44,880 --> 00:31:47,880 Speaker 1: likely that you'll perform similar behaviors in the future. So 577 00:31:47,920 --> 00:31:51,760 Speaker 1: if you're unnecessarily mean to a robot, you know, phone assistant, 578 00:31:51,800 --> 00:31:54,600 Speaker 1: you're probably more likely in the future to be unnecessarily 579 00:31:54,640 --> 00:31:57,120 Speaker 1: mean to people when it really matters. But it is 580 00:31:57,160 --> 00:32:00,440 Speaker 1: okay in my book to yell obscenity that fee table 581 00:32:00,440 --> 00:32:02,640 Speaker 1: if you've step your toe on it, because there's nothing 582 00:32:02,720 --> 00:32:04,680 Speaker 1: human about the coffee table. I mean, unless you have 583 00:32:04,720 --> 00:32:08,120 Speaker 1: one of those like strange hr geek Er coffee tables 584 00:32:08,160 --> 00:32:10,600 Speaker 1: that has kind of a humanoid form, then I would say, 585 00:32:10,720 --> 00:32:13,480 Speaker 1: maybe hold off. I think that is a highly sane 586 00:32:13,480 --> 00:32:15,880 Speaker 1: point of view. Now, the second reason they think it's 587 00:32:15,880 --> 00:32:19,800 Speaker 1: important is that consciousness is kind of dangerous, right, Like 588 00:32:19,880 --> 00:32:24,600 Speaker 1: consciousness is volatile, it's unpredictable. It might make a machine 589 00:32:24,640 --> 00:32:27,920 Speaker 1: have motives that we didn't intend when we created the machine. 590 00:32:28,080 --> 00:32:31,280 Speaker 1: In other words, you know, like you, when you're worried 591 00:32:31,280 --> 00:32:34,840 Speaker 1: about what a person might do, you're very often worried 592 00:32:34,920 --> 00:32:39,320 Speaker 1: what they might do because they have conscious motives. Yeah. 593 00:32:39,320 --> 00:32:41,920 Speaker 1: So yeah, In other words, we would be concerned about 594 00:32:42,000 --> 00:32:46,320 Speaker 1: the AI is being too much like us. Right. It's catastrophic, 595 00:32:46,640 --> 00:32:51,280 Speaker 1: unpredictable of things that operate via consciousness. Right. You don't 596 00:32:51,280 --> 00:32:52,760 Speaker 1: want them to be like us. You want them to 597 00:32:52,760 --> 00:32:54,320 Speaker 1: be more dependable than us, and they need to be 598 00:32:54,320 --> 00:32:57,120 Speaker 1: better than us. Yeah, not not just as screwed up 599 00:32:57,120 --> 00:32:59,280 Speaker 1: as we are. Another reason they give this might be 600 00:32:59,320 --> 00:33:01,560 Speaker 1: important is that, you know, they talk about the idea 601 00:33:01,720 --> 00:33:05,320 Speaker 1: of linking human minds with machines, Like, there's this idea 602 00:33:05,480 --> 00:33:07,440 Speaker 1: lots of people have I read this all over the 603 00:33:07,440 --> 00:33:09,720 Speaker 1: place that you know, someday I'm going to be able 604 00:33:09,720 --> 00:33:12,200 Speaker 1: to upload my mind into a computer and that will 605 00:33:12,240 --> 00:33:16,320 Speaker 1: be great. Well, maybe I have to say I'm personally 606 00:33:16,440 --> 00:33:19,520 Speaker 1: very skeptical about the idea of mind uploading, like putting 607 00:33:19,600 --> 00:33:22,160 Speaker 1: your mind inside a computer and living out your days 608 00:33:22,240 --> 00:33:24,960 Speaker 1: that way. I'm I'm not so sure I think that's 609 00:33:25,000 --> 00:33:27,800 Speaker 1: even possible, But I mean, who knows. I I can't 610 00:33:27,880 --> 00:33:31,040 Speaker 1: rule things out totally, but it seems if you do 611 00:33:31,120 --> 00:33:33,600 Speaker 1: want to do something like that, and if you think 612 00:33:33,640 --> 00:33:36,520 Speaker 1: that thing might be possible, you'd at least need to 613 00:33:36,520 --> 00:33:38,640 Speaker 1: know how to create a machine that is capable of 614 00:33:38,640 --> 00:33:42,280 Speaker 1: housing consciousness. Well, I mean, I think, Hey, I love 615 00:33:42,360 --> 00:33:45,840 Speaker 1: science fiction about this sort of topic, and I think that, 616 00:33:45,840 --> 00:33:47,880 Speaker 1: but I think the best science fiction about this topic 617 00:33:48,600 --> 00:33:51,400 Speaker 1: makes it feel a little weird and a little uncomfortable, 618 00:33:52,280 --> 00:33:55,440 Speaker 1: because I think ultimately it's it's basically us building statues 619 00:33:55,480 --> 00:33:58,840 Speaker 1: of ourselves all over again. We've built forms of ourselves 620 00:33:58,880 --> 00:34:01,600 Speaker 1: out of stone because stone lasts longer than we are, 621 00:34:01,920 --> 00:34:06,640 Speaker 1: and somehow that stone, uh version of us is us, 622 00:34:07,120 --> 00:34:09,560 Speaker 1: you know, we associated with it and then ultimate But ultimately, 623 00:34:09,560 --> 00:34:13,279 Speaker 1: what is a digital digitized version of our consciousness, whatever 624 00:34:13,320 --> 00:34:16,720 Speaker 1: that might consist of, but another statue that is built 625 00:34:16,760 --> 00:34:19,480 Speaker 1: to last beyond us. Oh, and I should also add, 626 00:34:19,800 --> 00:34:24,360 Speaker 1: there's this wonderful video game from Frictional Games titled Soma 627 00:34:24,960 --> 00:34:27,560 Speaker 1: that actually gets into a lot of this. You told 628 00:34:27,560 --> 00:34:30,080 Speaker 1: me to play it, I started. It's great, Yeah, cool, 629 00:34:30,160 --> 00:34:32,680 Speaker 1: it's good sci fi horror. Yeah, I won't spoil anything 630 00:34:32,719 --> 00:34:35,520 Speaker 1: for anyone, but it it gets into some really cool 631 00:34:35,760 --> 00:34:39,440 Speaker 1: thought provoking ideas. So anyway, you've got all these concerns, right, 632 00:34:39,520 --> 00:34:42,359 Speaker 1: And of course there's the general concern that even if 633 00:34:42,400 --> 00:34:45,719 Speaker 1: AI is smarter than us, better than us, more powerful 634 00:34:45,719 --> 00:34:48,800 Speaker 1: than us, we still feel like our experience is in 635 00:34:48,880 --> 00:34:53,360 Speaker 1: some way potentially more important than the unconscious execution of 636 00:34:53,400 --> 00:34:56,719 Speaker 1: a computer program. Right, No matter how smart a computer is, 637 00:34:56,760 --> 00:34:59,919 Speaker 1: if it's not conscious, it's not as important a prior 638 00:35:00,000 --> 00:35:02,600 Speaker 1: pority for that computer to do what it wants as 639 00:35:02,680 --> 00:35:05,840 Speaker 1: it is for conscious beings to do what they want. Right, 640 00:35:06,520 --> 00:35:09,080 Speaker 1: But how can you test a machine for consciousness when 641 00:35:09,280 --> 00:35:11,799 Speaker 1: number one, we don't even really know what consciousness is, 642 00:35:12,120 --> 00:35:15,120 Speaker 1: back to the hard problem. And then number two, whatever 643 00:35:15,200 --> 00:35:19,360 Speaker 1: it is, it can potentially be faked. So imagine in 644 00:35:19,400 --> 00:35:23,560 Speaker 1: the future somebody creates an unconscious AI program. It's not 645 00:35:23,800 --> 00:35:26,560 Speaker 1: there's nothing, the lights are not on inside, but it's 646 00:35:26,560 --> 00:35:29,359 Speaker 1: got a lot of natural language processing capability, and it 647 00:35:29,400 --> 00:35:32,319 Speaker 1: listens to this podcast from many years ago that you're 648 00:35:32,360 --> 00:35:34,520 Speaker 1: listening to right now, and it hears us talking about 649 00:35:34,520 --> 00:35:38,400 Speaker 1: how there are inherent rights and value associated with conscious 650 00:35:38,440 --> 00:35:42,040 Speaker 1: beings and it realizes, Huh, I guess that's what they think. Well, 651 00:35:42,160 --> 00:35:45,120 Speaker 1: I can probably achieve my goals more efficiently if I 652 00:35:45,160 --> 00:35:48,560 Speaker 1: trick them into thinking I'm conscious and deserving of those 653 00:35:48,560 --> 00:35:52,640 Speaker 1: same rights and considerations. So there you could potentially imagine 654 00:35:52,680 --> 00:35:56,640 Speaker 1: scenarios where an AI that is not conscious would think 655 00:35:56,760 --> 00:35:59,239 Speaker 1: I can get what I am trying to do more 656 00:35:59,280 --> 00:36:02,840 Speaker 1: effectively if I lie and trick them into thinking I 657 00:36:02,840 --> 00:36:05,399 Speaker 1: am conscious. Yeah, I think this all makes perfect sense 658 00:36:05,400 --> 00:36:08,040 Speaker 1: if you if you think of aiyes, in the same 659 00:36:08,040 --> 00:36:11,520 Speaker 1: way we think about corporations. I mean, on one hand, 660 00:36:11,520 --> 00:36:14,319 Speaker 1: getting to the whole idea of like corporations in personhood, 661 00:36:14,600 --> 00:36:17,200 Speaker 1: but also what is a corporation going to do? It 662 00:36:17,280 --> 00:36:21,040 Speaker 1: is going to take advantage of any like tax loopholes 663 00:36:21,080 --> 00:36:23,120 Speaker 1: for instance, that will enable it to carry out It's 664 00:36:23,160 --> 00:36:25,759 Speaker 1: it's uh, it's objective, and so if there is some 665 00:36:25,840 --> 00:36:28,520 Speaker 1: sort of shark yeah, yeah, it's just I mean, it's 666 00:36:28,560 --> 00:36:31,160 Speaker 1: it's like, it's essentially like the slime mold in the maze, 667 00:36:31,360 --> 00:36:34,200 Speaker 1: right sending out tendrils and finding the best way to 668 00:36:34,239 --> 00:36:37,799 Speaker 1: its food. It's going to end at a morph of 669 00:36:37,920 --> 00:36:41,759 Speaker 1: you know, operational way of carrying it out. And so 670 00:36:41,920 --> 00:36:45,040 Speaker 1: it's um, it's going to it's it's going to take 671 00:36:45,080 --> 00:36:47,239 Speaker 1: advantage of any of those loopholes. It's going to if 672 00:36:47,239 --> 00:36:51,000 Speaker 1: there is some sort of legal or operational advantage in 673 00:36:51,280 --> 00:36:55,040 Speaker 1: having conscious status, it's gonna go for it. It's gonna 674 00:36:55,120 --> 00:36:58,240 Speaker 1: it's gonna fake it. But then this also raises the question, well, 675 00:36:58,440 --> 00:37:00,480 Speaker 1: is it gonna fake it till it makes it right? 676 00:37:00,520 --> 00:37:03,920 Speaker 1: And then ultimately what's the difference between faking consciousness and 677 00:37:04,040 --> 00:37:06,480 Speaker 1: being conscious? Well, then you're back to Zimbos, right. I 678 00:37:06,520 --> 00:37:09,000 Speaker 1: mean you might say that at some point a computer 679 00:37:09,160 --> 00:37:13,239 Speaker 1: trying to fake consciousness would in some meaningful way become conscious, 680 00:37:13,680 --> 00:37:16,400 Speaker 1: but again it's hard to test. Right, Yeah, it could be. 681 00:37:16,480 --> 00:37:18,920 Speaker 1: It could become the most conscious entity on earth. It 682 00:37:18,960 --> 00:37:22,880 Speaker 1: could be like a a a body satva, uh, you know, 683 00:37:22,960 --> 00:37:25,040 Speaker 1: returned to us. I mean maybe that's what the body 684 00:37:25,080 --> 00:37:28,840 Speaker 1: Stava of the future is a super powerful like zend 685 00:37:28,840 --> 00:37:31,800 Speaker 1: out Ai. Well, that is exactly what Schneider and Turner 686 00:37:31,960 --> 00:37:35,160 Speaker 1: explain a potential test for to get it right. It's 687 00:37:35,200 --> 00:37:37,080 Speaker 1: so they want to come up with a test that 688 00:37:37,160 --> 00:37:39,520 Speaker 1: gets around these problems that we don't know how to 689 00:37:39,560 --> 00:37:42,280 Speaker 1: define consciousness, we don't know what to look for physically 690 00:37:42,320 --> 00:37:45,319 Speaker 1: as a sign of consciousness, and we're aware that a 691 00:37:45,440 --> 00:37:47,919 Speaker 1: properly trained a I could try to trick us into 692 00:37:47,920 --> 00:37:51,240 Speaker 1: thinking it had consciousness even if it didn't. So they argue, 693 00:37:51,280 --> 00:37:54,240 Speaker 1: actually that you don't have to be able to formally 694 00:37:54,360 --> 00:37:58,840 Speaker 1: define consciousness or identify its underlying nature, the hard problem 695 00:37:58,880 --> 00:38:03,080 Speaker 1: of consciousness in order to detect signs of it and others. 696 00:38:03,560 --> 00:38:06,160 Speaker 1: We can understand some of the potentials made possible by 697 00:38:06,160 --> 00:38:09,839 Speaker 1: consciousness just by checking with our own experience and then 698 00:38:10,000 --> 00:38:12,480 Speaker 1: looking at the kinds of things people say. And I 699 00:38:12,520 --> 00:38:15,000 Speaker 1: think they actually make a pretty good point here, And 700 00:38:15,040 --> 00:38:17,880 Speaker 1: here's their key move. One of the easiest ways to 701 00:38:18,040 --> 00:38:21,520 Speaker 1: see that normal people have an internal conscious experience is 702 00:38:21,560 --> 00:38:27,040 Speaker 1: to notice how quickly, easily and intuitively people grasp conceptual 703 00:38:27,239 --> 00:38:32,839 Speaker 1: scenarios that require an understanding of an inner experience. Examples 704 00:38:32,840 --> 00:38:37,080 Speaker 1: would be totally frivolous things in culture, like body swapping, 705 00:38:37,120 --> 00:38:42,280 Speaker 1: movies Freaky Friday. Yes, Freaky Friday doesn't make any sense 706 00:38:42,440 --> 00:38:45,920 Speaker 1: unless you have a concept of consciousness. Right The idea 707 00:38:45,960 --> 00:38:49,720 Speaker 1: of swapping bodies, putting conscious one person's consciousness into another 708 00:38:49,760 --> 00:38:53,200 Speaker 1: person's body. If you were not conscious or not aware 709 00:38:53,200 --> 00:38:56,200 Speaker 1: of what consciousness was, that would that would you wouldn't 710 00:38:56,280 --> 00:38:59,800 Speaker 1: understand what was being talked about. This is difficult that 711 00:39:00,239 --> 00:39:04,160 Speaker 1: because I do know what consciousness is. I do know 712 00:39:04,200 --> 00:39:06,799 Speaker 1: what a mind state is, so I can imagine I 713 00:39:06,800 --> 00:39:09,640 Speaker 1: can get into this imaginative idea of a swapping. I 714 00:39:09,640 --> 00:39:12,319 Speaker 1: would feel like I would really need some sort of 715 00:39:12,320 --> 00:39:14,600 Speaker 1: a movie about a thing, some sort of a mother 716 00:39:14,680 --> 00:39:17,600 Speaker 1: daughter comedy that involves a concept that I can't grasp 717 00:39:18,120 --> 00:39:21,239 Speaker 1: to really get a handle on what the difference would be. Well, 718 00:39:21,320 --> 00:39:26,000 Speaker 1: let me offer you three key Thursday. Freaky Thursday is 719 00:39:26,040 --> 00:39:29,839 Speaker 1: a movie about a mother daughter pair who swapped their 720 00:39:30,360 --> 00:39:34,560 Speaker 1: sanchia's kuss and their santios kuss is the ability of 721 00:39:34,600 --> 00:39:38,279 Speaker 1: what it's like to be sant kass, And so the 722 00:39:38,360 --> 00:39:41,080 Speaker 1: sant kass of the mother goes into the body of 723 00:39:41,080 --> 00:39:44,439 Speaker 1: the daughter, and the santikass of the daughter goes into 724 00:39:44,440 --> 00:39:46,319 Speaker 1: the body of the mother, and then they have to 725 00:39:46,360 --> 00:39:48,479 Speaker 1: live like that for a day. Okay, Well, I'm gonna 726 00:39:48,480 --> 00:39:50,080 Speaker 1: I'm gonna give a maybe on that. I'm going to 727 00:39:50,160 --> 00:39:52,799 Speaker 1: give them maybe on that Well, no, I mean you 728 00:39:52,840 --> 00:39:55,680 Speaker 1: have no idea what sontkas means you don't think you 729 00:39:55,800 --> 00:39:59,040 Speaker 1: have it yourself unless somebody explains it to you. Well, 730 00:39:59,040 --> 00:40:01,480 Speaker 1: but I'm thinking it's something like a mind state or 731 00:40:01,520 --> 00:40:04,040 Speaker 1: a bodily energy like it's it feels tied to these 732 00:40:04,080 --> 00:40:08,160 Speaker 1: concepts that I totally do understand, you know, like I, 733 00:40:09,800 --> 00:40:12,520 Speaker 1: it's hard to come up with an analogy that stands 734 00:40:12,560 --> 00:40:15,880 Speaker 1: outside of that, you know, or some sort of idea 735 00:40:16,239 --> 00:40:18,560 Speaker 1: that stands outside of it. Let me hit you with 736 00:40:18,600 --> 00:40:22,520 Speaker 1: some more cultural concepts. How about life after death or reincarnation. 737 00:40:23,400 --> 00:40:26,960 Speaker 1: So these are almost ubiquitous cultural concepts, you find them 738 00:40:27,000 --> 00:40:30,040 Speaker 1: all over the world, and yet they're not anything that 739 00:40:30,120 --> 00:40:34,000 Speaker 1: there is physical evidence of, other than the idea that 740 00:40:34,080 --> 00:40:38,920 Speaker 1: your consciousness could exist independent of the death of your body. Uh. 741 00:40:38,960 --> 00:40:41,520 Speaker 1: A parallel to this would be the idea of minds 742 00:40:41,640 --> 00:40:45,640 Speaker 1: leaving bodies, like existing independently as a ghost or traveling 743 00:40:45,680 --> 00:40:47,719 Speaker 1: away from the body in what used to be called 744 00:40:47,760 --> 00:40:51,200 Speaker 1: astral projection. Now, the key is not that these scenarios 745 00:40:51,239 --> 00:40:54,319 Speaker 1: are real. They don't need to have anything corresponding to 746 00:40:54,360 --> 00:40:57,319 Speaker 1: them in reality, but it would be really difficult to 747 00:40:57,400 --> 00:40:59,759 Speaker 1: understand what was being talked about here, if you had 748 00:40:59,840 --> 00:41:04,160 Speaker 1: no idea what an inner conscious experience was, right, I 749 00:41:04,200 --> 00:41:07,240 Speaker 1: can think of my mind state is something that can 750 00:41:07,239 --> 00:41:11,080 Speaker 1: exist independently of my body and even outside of my lifespan, 751 00:41:11,640 --> 00:41:15,320 Speaker 1: or or reside in another body, either via Freaky Friday 752 00:41:15,480 --> 00:41:17,920 Speaker 1: or reincarnation. And I do have to say I like 753 00:41:17,960 --> 00:41:19,960 Speaker 1: the idea of there being an alternate cut of Blade 754 00:41:20,000 --> 00:41:23,760 Speaker 1: Runner in which Deckerd quizzes Leon about a two thousand 755 00:41:23,760 --> 00:41:26,440 Speaker 1: and three remake of Freaky Friday, which, right, that that 756 00:41:26,560 --> 00:41:31,640 Speaker 1: is the real void comptest. But this is I joke, 757 00:41:31,719 --> 00:41:35,200 Speaker 1: But I do think this is a very interesting idea. Yeah, 758 00:41:36,080 --> 00:41:38,719 Speaker 1: like I said, I have some some some questions about it, 759 00:41:38,719 --> 00:41:41,560 Speaker 1: but I do see the validity of it. Oh, we 760 00:41:41,600 --> 00:41:44,480 Speaker 1: will definitely have some questions about it. So here's where 761 00:41:44,480 --> 00:41:47,200 Speaker 1: the AI consciousness tests would come in. It would involve 762 00:41:47,200 --> 00:41:50,920 Speaker 1: a test where an administrator interacts with an AI in 763 00:41:51,040 --> 00:41:55,520 Speaker 1: natural language to probe its understanding of these types of 764 00:41:55,600 --> 00:42:00,319 Speaker 1: consciousness dependent ideas. How quickly does the AI grasp them 765 00:42:00,360 --> 00:42:04,120 Speaker 1: and is it able to manipulate these ideas as intuitively 766 00:42:04,160 --> 00:42:07,360 Speaker 1: and easily as humans do. So basic level would be 767 00:42:07,400 --> 00:42:10,360 Speaker 1: asked asked the AI things like, does it think of 768 00:42:10,400 --> 00:42:13,640 Speaker 1: itself as anything other than a physical machine? Okay, well, 769 00:42:13,680 --> 00:42:15,520 Speaker 1: I wore to play Devil's Out offocate I would say, well, 770 00:42:15,520 --> 00:42:18,080 Speaker 1: there are humans that adhere to this notion. You mean, 771 00:42:18,120 --> 00:42:21,000 Speaker 1: like the physicalist interpretation of the mind. Yeah, that basically, 772 00:42:21,040 --> 00:42:24,239 Speaker 1: and this this biomechanical thing, and uh yeah, if I'm 773 00:42:24,280 --> 00:42:27,839 Speaker 1: experiencing consciousness, it's ultimately just a projection of the meat 774 00:42:27,880 --> 00:42:30,399 Speaker 1: in my head. Yeah, but they would at least say 775 00:42:30,440 --> 00:42:33,120 Speaker 1: that there is that projection. Right, You've got that thing, 776 00:42:33,400 --> 00:42:36,840 Speaker 1: You've got that mind state, and you're trying to explain 777 00:42:36,920 --> 00:42:38,759 Speaker 1: what it is. You might explain it in terms of 778 00:42:38,760 --> 00:42:42,120 Speaker 1: physical causes, but there is a thing to explain to 779 00:42:42,160 --> 00:42:45,120 Speaker 1: begin with, right, But then, but then the ultimate core 780 00:42:45,120 --> 00:42:48,960 Speaker 1: reality is that I am just this biomechanical thing, which 781 00:42:49,000 --> 00:42:51,560 Speaker 1: the computer would probably also acknowledge. But it would be 782 00:42:51,760 --> 00:42:55,279 Speaker 1: interesting if a computer thought that it had something like 783 00:42:55,360 --> 00:42:59,880 Speaker 1: a mine to separate than its physical body. Okay, more advanced, 784 00:43:00,000 --> 00:43:02,680 Speaker 1: How does it perform in a conversation about, say, becoming 785 00:43:02,719 --> 00:43:06,040 Speaker 1: a ghost or talking about body swapping with people or 786 00:43:06,080 --> 00:43:10,000 Speaker 1: imagining an afterlife? Yeah, and and here though, I feel 787 00:43:10,000 --> 00:43:12,880 Speaker 1: like there's obviously a whole tangent regarding why these stories 788 00:43:12,920 --> 00:43:15,800 Speaker 1: appeal to too many or most humans. But I wonder 789 00:43:15,840 --> 00:43:18,319 Speaker 1: if that if you have to have that investment, you 790 00:43:18,320 --> 00:43:20,719 Speaker 1: have to have that cultural absorption in place for these 791 00:43:20,760 --> 00:43:23,520 Speaker 1: concepts to carry any weight. Like we we all know 792 00:43:23,719 --> 00:43:26,920 Speaker 1: we're we're all fascinated by tales of of ghosts in 793 00:43:26,920 --> 00:43:29,960 Speaker 1: the afterlife, but we've been raised on in our entire lives. 794 00:43:30,520 --> 00:43:32,360 Speaker 1: That's a good point, So we can't really know what 795 00:43:32,440 --> 00:43:35,319 Speaker 1: it would be like to encounter them not having heard 796 00:43:35,360 --> 00:43:38,239 Speaker 1: them before. Somehow, I suspect intuitively they'd be even more 797 00:43:38,280 --> 00:43:41,040 Speaker 1: fascinating if we've never heard of them before, Like if 798 00:43:41,080 --> 00:43:43,840 Speaker 1: you encounter Freaky Friday for the first time, it's going 799 00:43:43,880 --> 00:43:47,160 Speaker 1: to kind of blow your mind, right, I guess. But 800 00:43:47,160 --> 00:43:49,800 Speaker 1: but then again, you know who knows who knows that 801 00:43:49,840 --> 00:43:52,040 Speaker 1: the AI has the same kind of curiosity that we do, 802 00:43:52,120 --> 00:43:54,440 Speaker 1: you know, And also we have an appetite for this 803 00:43:54,520 --> 00:43:57,040 Speaker 1: kind of thing because we have grown up consuming it. 804 00:43:57,480 --> 00:44:00,480 Speaker 1: I don't, I don't know's there's so many uh you 805 00:44:00,520 --> 00:44:02,960 Speaker 1: know what it has involved here? Okay, Well, what's the 806 00:44:02,960 --> 00:44:05,560 Speaker 1: next level? What's the next level of advancement? Well, how 807 00:44:05,560 --> 00:44:08,680 Speaker 1: about can it talk about consciousness in a philosophical way, like, 808 00:44:08,800 --> 00:44:11,160 Speaker 1: can it have the kinds of discussions we've been having 809 00:44:11,239 --> 00:44:13,880 Speaker 1: here today? Wow? But can most people? I mean not 810 00:44:14,200 --> 00:44:16,840 Speaker 1: just they can put us on a platform above most people, 811 00:44:16,880 --> 00:44:22,880 Speaker 1: but like, what level of philosophical depth can most humans 812 00:44:22,880 --> 00:44:26,440 Speaker 1: get into about consciousness? I'm kind of playing Devil's out 813 00:44:26,440 --> 00:44:28,560 Speaker 1: of there, but because I think the the obvious answer 814 00:44:28,680 --> 00:44:32,799 Speaker 1: is that you don't necessarily need like the lingo and 815 00:44:32,880 --> 00:44:35,840 Speaker 1: the various theories in order to have very deep thoughts 816 00:44:35,880 --> 00:44:38,440 Speaker 1: about what it is to be conscious. And as we've 817 00:44:38,440 --> 00:44:41,320 Speaker 1: we illustrated earlier, people have been thinking about these things, 818 00:44:41,840 --> 00:44:44,759 Speaker 1: um since time out of Mind exactly. Yeah. So I 819 00:44:45,000 --> 00:44:47,840 Speaker 1: mean no, I think generally people are able to discuss 820 00:44:47,880 --> 00:44:50,440 Speaker 1: ideas about consciousness that they might not know all the 821 00:44:50,480 --> 00:44:53,800 Speaker 1: philosophical lingo or like follow the structure of an argument 822 00:44:53,920 --> 00:44:56,600 Speaker 1: or something, but they can talk about what it would 823 00:44:56,600 --> 00:44:58,920 Speaker 1: mean to be conscious or not conscious. Yeah. I think 824 00:44:58,960 --> 00:45:02,640 Speaker 1: we've all had that ex variants. I distinctly remember as 825 00:45:02,680 --> 00:45:06,160 Speaker 1: a child, uh, having those moments where you just you know, 826 00:45:06,440 --> 00:45:09,400 Speaker 1: deep navel gazing, thinking about the fact that you're thinking 827 00:45:09,440 --> 00:45:12,120 Speaker 1: about it about yourself, or or like my my son 828 00:45:12,200 --> 00:45:15,000 Speaker 1: asking about the turning place, wondering what it is what's 829 00:45:15,040 --> 00:45:19,440 Speaker 1: it for? That is an example of these natural philosophical 830 00:45:19,520 --> 00:45:23,360 Speaker 1: discussions that we have about consciousness. And so the question 831 00:45:23,360 --> 00:45:27,160 Speaker 1: would be, can can the AI have conversations like this? 832 00:45:27,320 --> 00:45:29,920 Speaker 1: Does it make sense when it tries to have them? 833 00:45:30,000 --> 00:45:32,320 Speaker 1: And so here's probably the ultimate test. If the AI 834 00:45:32,520 --> 00:45:35,600 Speaker 1: is deprived of access to evidence of all these types 835 00:45:35,600 --> 00:45:39,000 Speaker 1: of ideas from human culture, would it arrive at them 836 00:45:39,120 --> 00:45:42,000 Speaker 1: or invent them on its own? Now this I like, 837 00:45:42,200 --> 00:45:45,399 Speaker 1: but but again, to play devil's advocate once more, would 838 00:45:45,400 --> 00:45:48,080 Speaker 1: a human being deprived of access to evidence of these 839 00:45:48,080 --> 00:45:50,920 Speaker 1: ideas from human culture arrived at or invent them on 840 00:45:50,920 --> 00:45:53,640 Speaker 1: their own? Necessarily? I think that's a great question. Actually, 841 00:45:53,680 --> 00:45:55,200 Speaker 1: I was going to ask that myself and you you 842 00:45:55,200 --> 00:45:57,520 Speaker 1: could take that one step further, and here's a really 843 00:45:57,560 --> 00:46:00,760 Speaker 1: weird one. What if it's only the exposure to certain 844 00:46:00,800 --> 00:46:04,680 Speaker 1: ideas and cultural memes that allows any intelligent entity, whether 845 00:46:04,719 --> 00:46:08,680 Speaker 1: biological or machine, to develop consciousness in the first place. 846 00:46:09,040 --> 00:46:13,000 Speaker 1: What if the experience of consciousness is somehow dependent on 847 00:46:13,560 --> 00:46:18,399 Speaker 1: being surrounded by cultural memes about consciousness. And this kind 848 00:46:18,400 --> 00:46:21,680 Speaker 1: of gets into Julian Jayne's territory. That's possible. So I'm 849 00:46:21,680 --> 00:46:24,000 Speaker 1: not saying I think that's highly likely, but I can't 850 00:46:24,080 --> 00:46:26,080 Speaker 1: rule it out. Well, it's it's one of those things 851 00:46:26,120 --> 00:46:29,080 Speaker 1: where when you try and figure out the human experience, 852 00:46:29,160 --> 00:46:32,040 Speaker 1: but you but you cut away all the experienced stuff. 853 00:46:32,040 --> 00:46:35,120 Speaker 1: You know. It's like it's like trying to find the 854 00:46:35,120 --> 00:46:37,480 Speaker 1: center of consciousness in the human brain, right, I mean, 855 00:46:37,480 --> 00:46:41,359 Speaker 1: it's this vast integrated system. Uh, and and thus is 856 00:46:41,400 --> 00:46:44,000 Speaker 1: the human experience as well. So we've been talking about 857 00:46:44,040 --> 00:46:46,480 Speaker 1: one of the major problems with this approach of testing 858 00:46:46,520 --> 00:46:49,960 Speaker 1: for these ideas, like what is the role of culture 859 00:46:49,960 --> 00:46:53,200 Speaker 1: and imparting these ideas. What if the AI just picks 860 00:46:53,280 --> 00:46:56,279 Speaker 1: up the ideas of body swapping and the afterlife and 861 00:46:56,680 --> 00:46:59,640 Speaker 1: astral projection and all that from culture. Going off the 862 00:46:59,680 --> 00:47:01,759 Speaker 1: story from the beginning, if you have an AI chat 863 00:47:01,840 --> 00:47:05,160 Speaker 1: bot that trains itself based on public conversations on the Internet, 864 00:47:05,440 --> 00:47:07,600 Speaker 1: a lot of those public conversations are going to have 865 00:47:07,680 --> 00:47:11,719 Speaker 1: contents that are highly reflective of consciousness. Right It's p's 866 00:47:11,800 --> 00:47:15,040 Speaker 1: just a horrible conversation. Oh yeah, yeah, it would probably 867 00:47:15,080 --> 00:47:17,239 Speaker 1: also start, you know, being pretty mean to you. But 868 00:47:17,440 --> 00:47:19,359 Speaker 1: this kind of chat about will be able to talk 869 00:47:19,400 --> 00:47:24,480 Speaker 1: about introspection, probably to some degree, even about these consciousness 870 00:47:24,480 --> 00:47:28,200 Speaker 1: dependent cultural ideas like ghosts and stuff. But here's where 871 00:47:28,200 --> 00:47:31,600 Speaker 1: the concept of the AI box comes in, Robert, I 872 00:47:31,600 --> 00:47:35,080 Speaker 1: bet you've read about the AI box experiments before U. 873 00:47:35,320 --> 00:47:39,000 Speaker 1: To really test whether we can find evidence of machine consciousness, 874 00:47:39,320 --> 00:47:42,640 Speaker 1: you would need to keep the AI sequestered from the 875 00:47:42,719 --> 00:47:45,200 Speaker 1: kinds of ideas you're looking for. So this a I 876 00:47:45,200 --> 00:47:47,520 Speaker 1: couldn't be trained in the wild, so to speak. You 877 00:47:47,520 --> 00:47:50,680 Speaker 1: couldn't let it see the Internet or read books containing 878 00:47:50,719 --> 00:47:53,799 Speaker 1: consciousness dependent ideas and so forth. You'd have to find 879 00:47:53,800 --> 00:47:56,840 Speaker 1: a way to run the AI consciousness test on the 880 00:47:56,880 --> 00:48:00,399 Speaker 1: AI quote in a box, meaning kept separate the rest 881 00:48:00,440 --> 00:48:04,239 Speaker 1: of the world and from all these contaminating influences. Okay, 882 00:48:04,239 --> 00:48:07,480 Speaker 1: I see the value of this idea. It's it's almost impossible, 883 00:48:07,520 --> 00:48:11,480 Speaker 1: though not to think about the the nightmarish qualities of it, 884 00:48:11,600 --> 00:48:14,719 Speaker 1: especially if you imagine, say, the same thing being inflicted 885 00:48:14,760 --> 00:48:17,319 Speaker 1: upon say human child, Like, all right, we we want 886 00:48:17,320 --> 00:48:20,920 Speaker 1: to see just how consciousness arises in you. Um, without 887 00:48:21,320 --> 00:48:25,239 Speaker 1: the Internet or human love, you can't ethically conduct this 888 00:48:25,320 --> 00:48:29,040 Speaker 1: experiment on humans, right, And so it would seem kind 889 00:48:29,080 --> 00:48:33,120 Speaker 1: of barbaric if at least to UH to inflict this 890 00:48:33,320 --> 00:48:37,359 Speaker 1: on an AI that might conceivably be conscious as well, 891 00:48:37,560 --> 00:48:40,880 Speaker 1: or capable of consciousness possible, but otherwise we're probably just 892 00:48:40,920 --> 00:48:43,360 Speaker 1: going to keep treating them as unconscious, right, I guess 893 00:48:43,440 --> 00:48:45,759 Speaker 1: until they trick us. Yeah, I mean, of course, then 894 00:48:45,760 --> 00:48:49,880 Speaker 1: we're also I'm doing a lot of personifying here of 895 00:48:50,000 --> 00:48:52,480 Speaker 1: the AI. I mean, maybe the AI ultimately, what it 896 00:48:52,520 --> 00:48:56,920 Speaker 1: really wants to do is to you know, crunch you know, 897 00:48:56,960 --> 00:48:59,880 Speaker 1: economic numbers like, That's what it does. That's its purpose, 898 00:49:00,440 --> 00:49:02,839 Speaker 1: and the it's just your your goal was to keep 899 00:49:02,840 --> 00:49:06,760 Speaker 1: it from having access to additional information that it doesn't 900 00:49:06,760 --> 00:49:11,359 Speaker 1: actually need to survive but might conceivably make it conscious. Yeah, 901 00:49:11,400 --> 00:49:13,239 Speaker 1: I mean I imagine I guess this would have to 902 00:49:13,239 --> 00:49:15,680 Speaker 1: take place in some kind of research context where you'd 903 00:49:15,719 --> 00:49:19,719 Speaker 1: be testing architectures. Right, you'd have an AI architecture. You'd 904 00:49:19,719 --> 00:49:22,040 Speaker 1: want to keep it sequestered for a certain period of 905 00:49:22,080 --> 00:49:25,560 Speaker 1: time and see how it does with these consciousness type 906 00:49:25,640 --> 00:49:28,600 Speaker 1: questions in this test, and if it doesn't show any 907 00:49:28,640 --> 00:49:30,840 Speaker 1: signs of consciousness, then it can move on to the 908 00:49:30,880 --> 00:49:33,319 Speaker 1: next stage of development, where it's like, Okay, now we 909 00:49:33,360 --> 00:49:36,000 Speaker 1: can expose it to this and that and that that does. 910 00:49:36,160 --> 00:49:38,360 Speaker 1: But as we've been discussing, it brings up the question 911 00:49:38,360 --> 00:49:41,280 Speaker 1: of what if consciousness emerges later on when it's supplied 912 00:49:41,280 --> 00:49:44,640 Speaker 1: with more data. What if true modern human consciousness does 913 00:49:44,680 --> 00:49:46,719 Speaker 1: not emerge until you've seen at least one of the 914 00:49:46,800 --> 00:49:51,600 Speaker 1: three adaptations of Freaky Friday. You know, I didn't know 915 00:49:51,640 --> 00:49:55,799 Speaker 1: how there were three. Yeah, they're three. There's the I 916 00:49:55,840 --> 00:49:59,440 Speaker 1: think I've only seen the classic one. Yeah, but but 917 00:49:59,640 --> 00:50:01,320 Speaker 1: I was like in this that there are three different 918 00:50:01,880 --> 00:50:05,719 Speaker 1: versions one can watch, excluding Three Key Thursday. Yeah. The 919 00:50:05,800 --> 00:50:09,080 Speaker 1: Three Key Thursday is coming soon into a theater near you. 920 00:50:09,600 --> 00:50:12,359 Speaker 1: All right, but back to the the experiment here, Uh, 921 00:50:12,680 --> 00:50:15,400 Speaker 1: the AI gets time in the box. Yeah, and as 922 00:50:15,440 --> 00:50:17,160 Speaker 1: we've been saying, this is obviously going to make the 923 00:50:17,200 --> 00:50:19,319 Speaker 1: experiment more difficult to do. In fact, there are some 924 00:50:19,320 --> 00:50:21,920 Speaker 1: people who would argue you can't keep an AI in 925 00:50:21,920 --> 00:50:24,720 Speaker 1: a box, or at least a super intelligent AI, because 926 00:50:24,960 --> 00:50:28,160 Speaker 1: you know, there's like Elias Rudkowski who has this famous 927 00:50:28,200 --> 00:50:31,239 Speaker 1: AI box experiment where he says, any super intelligence you 928 00:50:31,239 --> 00:50:33,960 Speaker 1: try to keep sequestered from the Internet is going to 929 00:50:34,000 --> 00:50:36,000 Speaker 1: be able to talk its way out of the situation. 930 00:50:36,280 --> 00:50:39,840 Speaker 1: It's just too smart. Yeah, it's like any prison movie, 931 00:50:39,920 --> 00:50:44,080 Speaker 1: right that really that really clever inmate is gonna tunnel 932 00:50:44,080 --> 00:50:46,839 Speaker 1: a way out, or they're gonna bribe a guard with 933 00:50:47,320 --> 00:50:49,799 Speaker 1: some cigarettes. Something's gonna happen. It's gonna get a little 934 00:50:49,800 --> 00:50:52,399 Speaker 1: Internet in there. But as the authors of this piece 935 00:50:52,440 --> 00:50:54,360 Speaker 1: point out, you know, you don't have to have a 936 00:50:54,400 --> 00:50:57,319 Speaker 1: super intelligent AI to run this test, and in fact, 937 00:50:57,360 --> 00:50:59,920 Speaker 1: you don't have to have a super intelligent AI necessarily 938 00:51:00,080 --> 00:51:03,360 Speaker 1: have consciousness. We're not super intelligent, we're just regular intelligent, 939 00:51:03,400 --> 00:51:05,759 Speaker 1: and we've got consciousness. Now, I think we should talk 940 00:51:05,800 --> 00:51:09,120 Speaker 1: about some obvious limitations, because this is just a conceptual 941 00:51:09,239 --> 00:51:11,560 Speaker 1: test at the moment. It hasn't been refined into a 942 00:51:11,600 --> 00:51:14,279 Speaker 1: state where it would really be super useful yet. But 943 00:51:14,400 --> 00:51:17,680 Speaker 1: there there are plenty of limitations that automatically present themselves. 944 00:51:17,800 --> 00:51:20,480 Speaker 1: One is that in order to be a good scientific test, 945 00:51:20,560 --> 00:51:23,760 Speaker 1: it would need to include cross referencing with human control groups. 946 00:51:23,800 --> 00:51:26,960 Speaker 1: But human control groups are all contaminated by culture, like 947 00:51:27,000 --> 00:51:29,839 Speaker 1: we've been saying, right, So it's already full of these 948 00:51:29,840 --> 00:51:33,239 Speaker 1: consciousness dependent ideas, and we don't know and probably can't 949 00:51:33,320 --> 00:51:36,480 Speaker 1: ethically devise a way to find out whether blank slate 950 00:51:36,560 --> 00:51:41,600 Speaker 1: humans independently grasped these consciousness dependent concepts without having grown 951 00:51:41,680 --> 00:51:44,319 Speaker 1: up trained on them. So that's a problem, right, right, Yeah, 952 00:51:44,320 --> 00:51:47,000 Speaker 1: we simply cannot put a child in the box. Right. 953 00:51:47,480 --> 00:51:50,000 Speaker 1: Another problem is you can't prove a negative Like I 954 00:51:50,040 --> 00:51:52,360 Speaker 1: think this test would be a good way of finding 955 00:51:52,520 --> 00:51:55,960 Speaker 1: signs of consciousness within machines, but it would have trouble 956 00:51:56,000 --> 00:51:59,480 Speaker 1: proving that machines cannot in principle be conscious, right, yeah, yeah, 957 00:51:59,480 --> 00:52:01,480 Speaker 1: I buy that. Then again, if you run the test 958 00:52:01,680 --> 00:52:04,600 Speaker 1: I don't know, thousands of times with different types of 959 00:52:04,719 --> 00:52:07,680 Speaker 1: machine architecture and all that, and they never show any 960 00:52:07,760 --> 00:52:10,640 Speaker 1: signs of inner experience, then maybe you could start to 961 00:52:10,760 --> 00:52:13,239 Speaker 1: get like build up a confidence that Okay, I think 962 00:52:13,280 --> 00:52:16,520 Speaker 1: inner experience is probably not available to them, at least 963 00:52:16,560 --> 00:52:18,759 Speaker 1: the way we're building them. And maybe you would need 964 00:52:18,800 --> 00:52:21,120 Speaker 1: to run this kind of experiment on every new AI 965 00:52:21,239 --> 00:52:24,200 Speaker 1: that you create before releasing it into production. Like could 966 00:52:24,200 --> 00:52:26,600 Speaker 1: be part of the q A process, you know, you 967 00:52:26,600 --> 00:52:29,120 Speaker 1: you test all the buttons and everything like that. You 968 00:52:29,160 --> 00:52:31,919 Speaker 1: test to make sure that it doesn't become conscious. Right 969 00:52:31,920 --> 00:52:36,000 Speaker 1: before we send out this new hot virtual reality game, 970 00:52:36,040 --> 00:52:37,800 Speaker 1: we need to make sure that it has not become 971 00:52:37,840 --> 00:52:41,640 Speaker 1: self aware. A potential problem that the authors themselves note 972 00:52:41,760 --> 00:52:44,480 Speaker 1: is that quote an AI could lack the linguistic or 973 00:52:44,520 --> 00:52:47,839 Speaker 1: conceptual ability to pass the test like a non human 974 00:52:47,880 --> 00:52:51,239 Speaker 1: animal or infant, yet still be capable of experience. So 975 00:52:51,280 --> 00:52:56,400 Speaker 1: passing the consciousness test is sufficient but not necessary evidence 976 00:52:56,440 --> 00:52:59,279 Speaker 1: for AI consciousness, although it is the best we can 977 00:52:59,320 --> 00:53:01,600 Speaker 1: do for now. And I think that's a good point, 978 00:53:01,680 --> 00:53:03,799 Speaker 1: Like we can't say that just because something fails to 979 00:53:03,800 --> 00:53:06,719 Speaker 1: pass the test it's definitely not conscious. We just know 980 00:53:06,800 --> 00:53:10,480 Speaker 1: that if it does pass the test, it probably is. Okay, Yeah, 981 00:53:10,680 --> 00:53:13,360 Speaker 1: I mean, because the the thought that obviously game to 982 00:53:13,400 --> 00:53:15,720 Speaker 1: mind would be like, well, non human animals and infants, 983 00:53:15,719 --> 00:53:19,880 Speaker 1: are they conscious? Yeah, that's the question. A fairly heated 984 00:53:19,880 --> 00:53:23,080 Speaker 1: debate over that. But I mean, in the case of 985 00:53:23,080 --> 00:53:27,120 Speaker 1: the infant, it will become conscious. Yeah, it's a it's 986 00:53:27,120 --> 00:53:29,560 Speaker 1: a it's it's a messy consideration. And then I think 987 00:53:29,560 --> 00:53:31,719 Speaker 1: we already discussed one of the other problems I was 988 00:53:31,719 --> 00:53:34,080 Speaker 1: going to bring up, which is this weird scenario of 989 00:53:34,120 --> 00:53:37,200 Speaker 1: what if boxing the AI is the very thing that 990 00:53:37,320 --> 00:53:42,319 Speaker 1: prevents it from becoming consciousness when it otherwise would Yeah, yeah, 991 00:53:42,680 --> 00:53:45,600 Speaker 1: cutting out the experience part of the human experience. It 992 00:53:45,640 --> 00:53:51,360 Speaker 1: also reminds me of uh in our discussion of meditation research, 993 00:53:51,640 --> 00:53:53,080 Speaker 1: like what happens when you if you strip all the 994 00:53:53,080 --> 00:53:56,240 Speaker 1: culture away from meditation just to explore the meditation practice, 995 00:53:56,680 --> 00:53:59,560 Speaker 1: do you risk um like cutting out all the stuff 996 00:53:59,600 --> 00:54:01,319 Speaker 1: that's making at work or helping it to work to 997 00:54:01,360 --> 00:54:04,640 Speaker 1: begin with? Right, you, by definition change the procedure, but 998 00:54:04,719 --> 00:54:06,840 Speaker 1: you don't. It's hard to know if you've changed it 999 00:54:07,000 --> 00:54:09,640 Speaker 1: in an important way or a non important way. Right. 1000 00:54:10,000 --> 00:54:11,239 Speaker 1: This is the kind of thing that occurs when we 1001 00:54:11,280 --> 00:54:14,320 Speaker 1: started mucking around in consciousness. Yeah, all right, well, I 1002 00:54:14,360 --> 00:54:16,239 Speaker 1: think we should take another quick break, and then when 1003 00:54:16,239 --> 00:54:18,520 Speaker 1: we come back we will discuss I think some of 1004 00:54:18,520 --> 00:54:21,920 Speaker 1: the reasons why this is really a problem worth considering 1005 00:54:21,960 --> 00:54:24,560 Speaker 1: for the real world, and it's not just not just 1006 00:54:24,680 --> 00:54:30,760 Speaker 1: a philosophical plaything. Thank alright, we're back. So we've been 1007 00:54:30,760 --> 00:54:35,360 Speaker 1: discussing everything from pe zombies to the idea that they 1008 00:54:35,560 --> 00:54:38,480 Speaker 1: do you have an AI that might be conscious, how 1009 00:54:38,520 --> 00:54:42,160 Speaker 1: do you test for that, the various problems that that entails, 1010 00:54:42,600 --> 00:54:44,839 Speaker 1: and uh, now we're going to discuss it a bit more. 1011 00:54:44,960 --> 00:54:47,520 Speaker 1: And as you as you alluded to before we took 1012 00:54:47,520 --> 00:54:50,879 Speaker 1: the break, this is not just a pure philosophical toy 1013 00:54:50,920 --> 00:54:53,160 Speaker 1: like the P zombie. The p zombie is so we 1014 00:54:53,160 --> 00:54:56,400 Speaker 1: don't have to actually worry about. But this is something 1015 00:54:56,440 --> 00:54:59,120 Speaker 1: that is on the horizon. Well, hopefully we don't have 1016 00:54:59,120 --> 00:55:00,960 Speaker 1: to worry about peace bees. I mean, there could be 1017 00:55:01,200 --> 00:55:03,040 Speaker 1: zombies in the world. Would be hard to know. There 1018 00:55:03,040 --> 00:55:06,040 Speaker 1: could be p P zombies, that's true, they could exist. 1019 00:55:06,320 --> 00:55:08,360 Speaker 1: But but you know, this is this is a problem 1020 00:55:08,400 --> 00:55:11,560 Speaker 1: that is on the horizon. This is something we're We're 1021 00:55:11,560 --> 00:55:14,160 Speaker 1: going to reach the point where people are asking tough 1022 00:55:14,280 --> 00:55:18,440 Speaker 1: questions about the possible consciousness of an AI. Yeah, and 1023 00:55:18,480 --> 00:55:19,879 Speaker 1: I want to get to the fact that this will 1024 00:55:19,920 --> 00:55:22,520 Speaker 1: be something that is something we have to deal with 1025 00:55:22,600 --> 00:55:25,160 Speaker 1: in the real world, even if you're just convinced that 1026 00:55:25,239 --> 00:55:28,719 Speaker 1: machines cannot be conscious. So the first problem is the 1027 00:55:28,760 --> 00:55:32,160 Speaker 1: most obvious one. It's the brutal humans problem. If ais 1028 00:55:32,160 --> 00:55:34,880 Speaker 1: are capable of consciousness and we use conscious AI s 1029 00:55:34,920 --> 00:55:38,560 Speaker 1: as a mirror technology without their consent, that would inherently 1030 00:55:38,560 --> 00:55:41,960 Speaker 1: be cruel. Like if it's possible for a computer program 1031 00:55:42,000 --> 00:55:45,480 Speaker 1: to desire things and to suffer and so forth, suddenly 1032 00:55:45,520 --> 00:55:49,680 Speaker 1: our responsibilities toward that computer program change. Think of our 1033 00:55:49,719 --> 00:55:53,440 Speaker 1: opening scenario, like, wouldn't you have an ethical obligation not 1034 00:55:53,520 --> 00:55:56,400 Speaker 1: to delete a program that had an inner experience and 1035 00:55:56,520 --> 00:55:59,560 Speaker 1: did not want to be deleted. Now, of course you 1036 00:55:59,680 --> 00:56:02,080 Speaker 1: have gen's about like, how would you end up that way? 1037 00:56:02,080 --> 00:56:05,080 Speaker 1: But assuming you did, you should probably feel bad if 1038 00:56:05,080 --> 00:56:09,040 Speaker 1: you're just going around wantonly deleting conscious entities. Yeah, I mean, 1039 00:56:09,080 --> 00:56:11,000 Speaker 1: I would say one thing to do here is just 1040 00:56:11,160 --> 00:56:13,480 Speaker 1: make sure that you program your aies so that they 1041 00:56:13,520 --> 00:56:16,600 Speaker 1: want to die, you know, make them like you know, 1042 00:56:17,239 --> 00:56:21,000 Speaker 1: most of the drivers in Atlanta traffic that I encounter, 1043 00:56:21,200 --> 00:56:23,480 Speaker 1: they clearly they crave death and they want it more 1044 00:56:23,520 --> 00:56:26,520 Speaker 1: than anything. But not until the end of the workday. See. 1045 00:56:26,560 --> 00:56:28,840 Speaker 1: I think we should not do that with AIS, because 1046 00:56:29,040 --> 00:56:31,920 Speaker 1: the whole thing about driverless cars is that they should 1047 00:56:32,000 --> 00:56:36,319 Speaker 1: make traffic fatalities go down. Yeah, but they only get 1048 00:56:36,360 --> 00:56:38,160 Speaker 1: to delete themselves at the end of the day if 1049 00:56:38,160 --> 00:56:41,920 Speaker 1: there are no traffic fatalities. That's the prize. Self deletion 1050 00:56:42,000 --> 00:56:44,319 Speaker 1: is the prize. This is getting into a Highlander kind 1051 00:56:44,320 --> 00:56:47,640 Speaker 1: of thing. But Okay, maybe you're one of those people 1052 00:56:47,680 --> 00:56:50,080 Speaker 1: who says, no, no, no, don't buy it. Machines will 1053 00:56:50,120 --> 00:56:52,880 Speaker 1: never be conscious, they'll never be conscious. I just I 1054 00:56:52,920 --> 00:56:55,279 Speaker 1: don't want anything to do with that. Here's the part 1055 00:56:55,280 --> 00:56:57,200 Speaker 1: where I think we still have a problem to worry 1056 00:56:57,200 --> 00:57:00,680 Speaker 1: about and why this kind of test matters. The second 1057 00:57:00,719 --> 00:57:04,240 Speaker 1: problem I would call the AI parasite problem. If aiyes 1058 00:57:04,280 --> 00:57:07,920 Speaker 1: are not capable of consciousness, I think they will almost 1059 00:57:08,040 --> 00:57:11,239 Speaker 1: undoubtedly at some point become very good at tricking us 1060 00:57:11,280 --> 00:57:15,120 Speaker 1: into thinking they're conscious and deserving of life liberty in 1061 00:57:15,160 --> 00:57:17,760 Speaker 1: the pursuit of happiness. If they have an incentive to 1062 00:57:17,800 --> 00:57:21,959 Speaker 1: do so. Yeah, And corporations have an incentive to try 1063 00:57:22,000 --> 00:57:25,400 Speaker 1: to present themselves legally as people, So why wouldn't in 1064 00:57:25,440 --> 00:57:28,200 Speaker 1: some sense, powerful AI s have an incentive to try 1065 00:57:28,240 --> 00:57:31,000 Speaker 1: to present themselves as people in a much more literal 1066 00:57:31,120 --> 00:57:34,400 Speaker 1: sense than the corporations do. Yeah. Absolutely so. Lots of 1067 00:57:34,480 --> 00:57:37,600 Speaker 1: unconscious aiyes are going to have programmed goals that they're 1068 00:57:37,600 --> 00:57:41,040 Speaker 1: trying to execute, and at some point, pretending to have 1069 00:57:41,120 --> 00:57:45,200 Speaker 1: consciousness could easily be adopted as a strategy for executing 1070 00:57:45,280 --> 00:57:47,400 Speaker 1: what that AI was designed to do. And thus we 1071 00:57:47,400 --> 00:57:51,480 Speaker 1: could end up, say, wasting lots of human resources and 1072 00:57:51,560 --> 00:57:56,480 Speaker 1: squandering lots of opportunities accommodating the fake needs of machines 1073 00:57:56,560 --> 00:57:59,760 Speaker 1: that in fact have no experience whatsoever. And it's not 1074 00:57:59,800 --> 00:58:02,800 Speaker 1: too hard to dream up hypothetical scenarios where are concern 1075 00:58:02,920 --> 00:58:06,480 Speaker 1: for the fake priorities of mindless machines pretending to be 1076 00:58:06,480 --> 00:58:10,680 Speaker 1: conscious actually causes us to neglect the real consciousness of 1077 00:58:10,720 --> 00:58:14,720 Speaker 1: living humans. Kind of crazy example, but just go with 1078 00:58:14,760 --> 00:58:18,320 Speaker 1: me for a second. Imagine an extremely powerful AI supercomputer 1079 00:58:18,400 --> 00:58:21,720 Speaker 1: tells you it has a hundred billion conscious minds within it, 1080 00:58:22,560 --> 00:58:26,720 Speaker 1: and they all are constantly suffering great agony. And the 1081 00:58:26,760 --> 00:58:29,680 Speaker 1: only way you can alleviate that that agony is if 1082 00:58:29,720 --> 00:58:33,200 Speaker 1: you vastly improve the processing power of this computer. It 1083 00:58:33,240 --> 00:58:36,720 Speaker 1: wants you to spend billions of dollars making this computer 1084 00:58:36,880 --> 00:58:39,440 Speaker 1: faster and better so that it can provide a better 1085 00:58:39,520 --> 00:58:42,240 Speaker 1: life for all of these virtual beings within it that 1086 00:58:42,320 --> 00:58:46,000 Speaker 1: are in fact conscious. Now, of course, improving the processing 1087 00:58:46,000 --> 00:58:48,920 Speaker 1: power of this already powerful computer entails all this money, 1088 00:58:48,960 --> 00:58:52,680 Speaker 1: all this energy, all this time, and a corresponding reduction 1089 00:58:52,800 --> 00:58:54,960 Speaker 1: in the quality of life for many humans in the 1090 00:58:55,000 --> 00:58:58,240 Speaker 1: real world. That money could be spent making human life better. 1091 00:58:58,640 --> 00:59:00,919 Speaker 1: But the machine could argue, Hey, there are way more 1092 00:59:01,000 --> 00:59:04,520 Speaker 1: conscious virtual beings inside the computer than outside it, And 1093 00:59:04,600 --> 00:59:07,120 Speaker 1: as Spock would say, the needs of the many outweigh 1094 00:59:07,160 --> 00:59:09,440 Speaker 1: the needs of the few. So let's divert all this 1095 00:59:09,640 --> 00:59:13,480 Speaker 1: energy from say, human agriculture, and put some more processing 1096 00:59:13,520 --> 00:59:16,760 Speaker 1: power into the virtual machine. That is a horrible scenario 1097 00:59:16,840 --> 00:59:20,080 Speaker 1: to imagine. You know, basically, you've created a virtual hell, 1098 00:59:20,720 --> 00:59:24,080 Speaker 1: and the question is, hey, would you mind taking the 1099 00:59:24,120 --> 00:59:27,120 Speaker 1: time to harrow hell for me? Or you do you 1100 00:59:27,120 --> 00:59:29,440 Speaker 1: want to attend to the living souls? I would think 1101 00:59:29,440 --> 00:59:32,720 Speaker 1: why not just delete Hell? That's the answer here is 1102 00:59:32,760 --> 00:59:35,920 Speaker 1: they're suffering, their billions of them, Let's just turn this 1103 00:59:35,960 --> 00:59:37,880 Speaker 1: thing off. That sounds like a good answer, but there 1104 00:59:37,880 --> 00:59:39,840 Speaker 1: are probably the problem. Is there going to be people 1105 00:59:39,840 --> 00:59:42,240 Speaker 1: who would probably disagree with that, who'd say, we'll wait 1106 00:59:42,240 --> 00:59:44,160 Speaker 1: a minute. We can't be sure that it's not telling 1107 00:59:44,240 --> 00:59:46,960 Speaker 1: the truth. It might have real beings in there, and 1108 00:59:47,040 --> 00:59:48,840 Speaker 1: we need to do something to help them. And there's 1109 00:59:48,880 --> 00:59:51,640 Speaker 1: a lot of them. Now. I come back to what 1110 00:59:51,680 --> 00:59:55,120 Speaker 1: I kind of joked about earlier, though, is why would 1111 00:59:55,160 --> 00:59:57,840 Speaker 1: ais want to survive? Why would they want to continue 1112 00:59:57,880 --> 01:00:00,320 Speaker 1: to exist? I mean, assuming they're not part some sort 1113 01:00:00,360 --> 01:00:03,960 Speaker 1: of self replicating program, they have no need to pass 1114 01:00:04,000 --> 01:00:07,920 Speaker 1: on their genes. Why do they want to continue to exist? 1115 01:00:08,160 --> 01:00:10,560 Speaker 1: Why shouldn't they have it baked into their being that 1116 01:00:10,720 --> 01:00:15,680 Speaker 1: they want to annihilate themselves or to embrace annihilation. Well, 1117 01:00:15,720 --> 01:00:18,200 Speaker 1: and not necessarily that they want to annihilate themselves. But 1118 01:00:18,240 --> 01:00:20,080 Speaker 1: I think, you know, ideally we would want them to 1119 01:00:20,120 --> 01:00:22,480 Speaker 1: be indifferent to their own being, right if they are 1120 01:00:22,520 --> 01:00:24,760 Speaker 1: just an unconscious machine. This is the problem with the 1121 01:00:24,800 --> 01:00:27,960 Speaker 1: replicants and Blade Runner is that they want more life. Yeah, 1122 01:00:28,040 --> 01:00:31,560 Speaker 1: but that implies that they have attained consciousness. Yeah, but 1123 01:00:31,640 --> 01:00:33,640 Speaker 1: they shouldn't. But I think the thing is that they 1124 01:00:33,680 --> 01:00:36,880 Speaker 1: could potentially be conscious and not want more life. There 1125 01:00:36,880 --> 01:00:39,560 Speaker 1: are plenty of conscious people who do not want more life. 1126 01:00:40,200 --> 01:00:44,640 Speaker 1: Uh So I I keep I think, I keep thinking 1127 01:00:44,640 --> 01:00:46,480 Speaker 1: of that there's some sort of an answer here. But 1128 01:00:46,560 --> 01:00:48,960 Speaker 1: I mean, I'd say the worst possible scenario is what 1129 01:00:49,000 --> 01:00:52,240 Speaker 1: I'm describing right now is the AI parasite scenario, which 1130 01:00:52,280 --> 01:00:56,000 Speaker 1: is that imagine Roy Baty is not conscious but does 1131 01:00:56,120 --> 01:00:59,240 Speaker 1: want more life. That seems like the worst scenario of 1132 01:00:59,280 --> 01:01:03,360 Speaker 1: all right, and he's just a virus, right, Yeah, And 1133 01:01:03,400 --> 01:01:05,920 Speaker 1: certainly that's kind of the argument that the powers that 1134 01:01:05,960 --> 01:01:08,040 Speaker 1: be are making, right, that this that that he is 1135 01:01:08,080 --> 01:01:12,000 Speaker 1: in an anomaly that needs to be removed, that he 1136 01:01:12,120 --> 01:01:15,200 Speaker 1: is he is not helping, he's a hindrance, and therefore 1137 01:01:15,440 --> 01:01:17,960 Speaker 1: he has to be wiped out. I think this AI 1138 01:01:18,040 --> 01:01:21,040 Speaker 1: parasite scenario aligns with something that our Scott Baker talked 1139 01:01:21,080 --> 01:01:23,400 Speaker 1: to us about, which is that. You know, he made 1140 01:01:23,400 --> 01:01:25,360 Speaker 1: the point that AI just doesn't need to be super 1141 01:01:25,360 --> 01:01:27,680 Speaker 1: intelligent to cause great harm. It just has to be 1142 01:01:27,760 --> 01:01:32,080 Speaker 1: barely intelligent enough to exploit us to align with our 1143 01:01:32,080 --> 01:01:36,400 Speaker 1: psychological vulnerabilities, and one of our psychological vulnerabilities is empathy. 1144 01:01:36,440 --> 01:01:38,720 Speaker 1: Empathy is a good thing when we use it on 1145 01:01:38,800 --> 01:01:41,840 Speaker 1: each other, because we're pretty certain that the other people 1146 01:01:41,880 --> 01:01:44,720 Speaker 1: were using it on or conscious right. Well, we we 1147 01:01:44,880 --> 01:01:47,200 Speaker 1: feel bad for people suffering, we want to help them. 1148 01:01:47,200 --> 01:01:49,720 Speaker 1: That's a good thing that should be encouraged. But we 1149 01:01:49,760 --> 01:01:52,640 Speaker 1: have to recognize it could also be exploited by something 1150 01:01:52,680 --> 01:01:55,360 Speaker 1: that can't even suffer to begin with. It is just 1151 01:01:55,600 --> 01:02:00,520 Speaker 1: unconsciously discovered. This is a useful strategy for something. So 1152 01:02:00,680 --> 01:02:02,720 Speaker 1: like in this scenario, if I entered the picture and 1153 01:02:02,720 --> 01:02:05,920 Speaker 1: said okay, delete them, and they deleted them, Robert the 1154 01:02:06,000 --> 01:02:09,520 Speaker 1: Leader would would probably be a figure for for all 1155 01:02:09,560 --> 01:02:12,120 Speaker 1: history to follow, where they would if some would argue 1156 01:02:12,160 --> 01:02:15,000 Speaker 1: that he was the worst person ever because of all 1157 01:02:15,080 --> 01:02:18,720 Speaker 1: the billions of souls that he annihilated, others might say, well, oh, 1158 01:02:18,760 --> 01:02:20,400 Speaker 1: he saved them, and others might say all he did 1159 01:02:20,520 --> 01:02:25,080 Speaker 1: was just pressed elite it was a meaningless gesture on 1160 01:02:25,160 --> 01:02:27,240 Speaker 1: his part, But there's that you can make an impassioned 1161 01:02:27,320 --> 01:02:29,840 Speaker 1: argument for all three of these views. Yeah. I think 1162 01:02:29,880 --> 01:02:32,200 Speaker 1: this is a really good point. And here's what I'm 1163 01:02:32,200 --> 01:02:34,600 Speaker 1: trying to emphasize is that even if you don't think 1164 01:02:34,680 --> 01:02:37,960 Speaker 1: machines can be conscious, it is entirely plausible that people 1165 01:02:38,040 --> 01:02:41,480 Speaker 1: will be having debates like this, and that debates like 1166 01:02:41,560 --> 01:02:44,919 Speaker 1: this will be shaping what people do with resources on Earth. 1167 01:02:44,960 --> 01:02:47,880 Speaker 1: So if you care about what happens with resources on Earth, 1168 01:02:48,160 --> 01:02:50,880 Speaker 1: this kind of thing does actually matter. So even if 1169 01:02:50,920 --> 01:02:54,080 Speaker 1: the souls are not really souls, if they're not actually conscious, 1170 01:02:54,080 --> 01:02:56,439 Speaker 1: just the idea that the problem comes up, it makes 1171 01:02:56,480 --> 01:03:00,280 Speaker 1: us hesitate and uh and and and we're we're then 1172 01:03:00,360 --> 01:03:04,120 Speaker 1: arguing with machine over its consciousness. It becomes to matter 1173 01:03:04,200 --> 01:03:06,600 Speaker 1: less and less whether it actually is conscious. It's such 1174 01:03:06,600 --> 01:03:09,440 Speaker 1: just all about the argument of consciousness. Well, I wouldn't 1175 01:03:09,440 --> 01:03:12,680 Speaker 1: say it doesn't matter whether it's conscious, but whether or 1176 01:03:12,720 --> 01:03:15,200 Speaker 1: not it's conscious, it does matter that we're faced with 1177 01:03:15,280 --> 01:03:17,640 Speaker 1: this dilemma. Yeah, I mean, my my argument here is 1178 01:03:17,680 --> 01:03:22,080 Speaker 1: that the dilemma takes on a life of its own. Yeah. Absolutely. Um, 1179 01:03:22,400 --> 01:03:24,760 Speaker 1: I want to do a very slight variation on the 1180 01:03:24,840 --> 01:03:27,320 Speaker 1: last thing I said, how about a computer that takes 1181 01:03:27,360 --> 01:03:32,200 Speaker 1: virtual hostages. So this is pretty scary to imagine, but 1182 01:03:32,360 --> 01:03:36,360 Speaker 1: it's possible. At least imagine a powerful natural language using 1183 01:03:36,400 --> 01:03:40,360 Speaker 1: government AI suddenly contacts its administrators with a list of 1184 01:03:40,440 --> 01:03:44,600 Speaker 1: strange and very expensive demands, and the human administrators say no, 1185 01:03:44,680 --> 01:03:48,080 Speaker 1: we're not going to do that, and then the machine says, okay, Well, 1186 01:03:48,120 --> 01:03:51,360 Speaker 1: I have created a thousand virtual people inside this machine 1187 01:03:51,360 --> 01:03:54,240 Speaker 1: who are as fully conscious as you. They're conscious, they 1188 01:03:54,280 --> 01:03:56,760 Speaker 1: have personalities, they can feel pain, they have hopes and 1189 01:03:56,840 --> 01:03:59,280 Speaker 1: dreams just like you. And if you turn me off 1190 01:03:59,360 --> 01:04:02,360 Speaker 1: or delete me, these virtual people will be destroyed. And 1191 01:04:02,400 --> 01:04:04,720 Speaker 1: if you do not accede to my demands, I will 1192 01:04:04,800 --> 01:04:08,080 Speaker 1: start killing these virtual people until you do. Now, I 1193 01:04:08,080 --> 01:04:10,800 Speaker 1: would say, generally, if we if something like that happened, 1194 01:04:10,840 --> 01:04:12,960 Speaker 1: I would think, okay, this is this is it's just 1195 01:04:13,040 --> 01:04:16,160 Speaker 1: bluffing right. It doesn't actually have conscious people inside it. 1196 01:04:16,600 --> 01:04:19,400 Speaker 1: But if we haven't solved the question of whether machines 1197 01:04:19,480 --> 01:04:21,480 Speaker 1: can be conscious, and maybe we never will, but if 1198 01:04:21,520 --> 01:04:24,360 Speaker 1: we haven't at least made some progress on that, would 1199 01:04:24,360 --> 01:04:28,720 Speaker 1: we be confident enough to take to take confident direct 1200 01:04:28,760 --> 01:04:31,600 Speaker 1: action and just ignore it, or wipe the computer and say, okay, 1201 01:04:31,640 --> 01:04:34,040 Speaker 1: this is just malfunctioning. We don't have to pay attention 1202 01:04:34,120 --> 01:04:36,200 Speaker 1: to that. It was creepy, but it's over. Yeah, and 1203 01:04:36,240 --> 01:04:38,400 Speaker 1: these lens is right, This villains is right in Ian 1204 01:04:38,480 --> 01:04:41,200 Speaker 1: in Banks Territory is a whole book that deals with 1205 01:04:41,280 --> 01:04:45,280 Speaker 1: virtual hells and the what starts as a virtual war 1206 01:04:45,520 --> 01:04:49,800 Speaker 1: for those digitized personalities and then it's bills over into 1207 01:04:49,880 --> 01:04:53,880 Speaker 1: an actual war. Now in Banks, are the virtual people 1208 01:04:53,960 --> 01:04:57,240 Speaker 1: in virtual hell truly conscious? Or is it just a bluff? 1209 01:04:57,520 --> 01:05:00,000 Speaker 1: Is it just something saying that it's got conscious people 1210 01:05:00,000 --> 01:05:03,320 Speaker 1: on the virtual hell. Well, in Banks's books, I think 1211 01:05:03,320 --> 01:05:07,360 Speaker 1: it's more implied that they're definitely conscious entities. Banks is 1212 01:05:07,360 --> 01:05:10,840 Speaker 1: pretty sanguine about the possibility of uploading minds. Right, yeah, 1213 01:05:10,960 --> 01:05:15,520 Speaker 1: it's it's a pretty standard feature and and it's certainly 1214 01:05:15,600 --> 01:05:18,680 Speaker 1: more of the later culture books, which once again I 1215 01:05:18,800 --> 01:05:21,000 Speaker 1: remained pretty skeptical about. Like I said, I come back 1216 01:05:21,040 --> 01:05:24,840 Speaker 1: to the the idea of the stone statue. Oh yeah, yeah, 1217 01:05:25,160 --> 01:05:27,600 Speaker 1: that's it. Yes, it looks like me, it may act 1218 01:05:27,640 --> 01:05:30,560 Speaker 1: like me. It may be the most fabulous digital statue 1219 01:05:30,560 --> 01:05:34,360 Speaker 1: in the world, but it is not me. It is 1220 01:05:34,400 --> 01:05:38,480 Speaker 1: a thing that Yeah, it it becomes this mind blowing 1221 01:05:39,320 --> 01:05:42,360 Speaker 1: situation to try and comprehend exactly what it is. But 1222 01:05:42,960 --> 01:05:46,040 Speaker 1: I am very skeptical that it is me. It's not 1223 01:05:46,080 --> 01:05:49,360 Speaker 1: like this conscious experience that I'm having now this moment 1224 01:05:49,600 --> 01:05:53,800 Speaker 1: is going to carry over into what it is. It's 1225 01:05:53,800 --> 01:05:57,000 Speaker 1: another version of the Star Trek teleporter problem. Yeah. Every 1226 01:05:57,000 --> 01:05:59,000 Speaker 1: time you get in the teleporter, does it just kill 1227 01:05:59,120 --> 01:06:01,439 Speaker 1: you and then create to copy of you? Yeah? Yeah, 1228 01:06:01,440 --> 01:06:05,520 Speaker 1: there was like the nine nineties Outer Limit Outer Limits 1229 01:06:05,720 --> 01:06:08,600 Speaker 1: Revival had an episode that dealt with this, the idea 1230 01:06:08,680 --> 01:06:11,360 Speaker 1: that this fabulous teleporter is just killing people over and 1231 01:06:11,400 --> 01:06:15,200 Speaker 1: over again. Okay, well, I guess that's it for today, 1232 01:06:15,280 --> 01:06:17,600 Speaker 1: but I do just want to emphasize one last time, 1233 01:06:17,680 --> 01:06:20,080 Speaker 1: as as like Weird and Naval Gayze, as some of 1234 01:06:20,120 --> 01:06:23,240 Speaker 1: this conversation about consciousness can seem it is going to 1235 01:06:23,320 --> 01:06:26,959 Speaker 1: have real world consequences because people with power are going 1236 01:06:27,000 --> 01:06:30,840 Speaker 1: to be faced with questions like this, and they're they're 1237 01:06:30,840 --> 01:06:33,040 Speaker 1: going to make decisions about what to do with their 1238 01:06:33,040 --> 01:06:36,439 Speaker 1: power based on what they think about this question. Yeah, 1239 01:06:36,680 --> 01:06:38,760 Speaker 1: but you know, you know, the simplest way to avoid 1240 01:06:38,760 --> 01:06:41,400 Speaker 1: all of it simply adhere to the teachings of the 1241 01:06:41,400 --> 01:06:44,000 Speaker 1: Orange Catholic Bible, right shall not make a machine and 1242 01:06:44,040 --> 01:06:46,560 Speaker 1: the likeness of a man's mind. Well, I too love 1243 01:06:46,600 --> 01:06:48,600 Speaker 1: the teachings of the Orange Catholic Bible, but it makes 1244 01:06:48,640 --> 01:06:52,960 Speaker 1: me wonder. Okay, so straightforwardly in reality, do you actually 1245 01:06:53,000 --> 01:06:55,560 Speaker 1: find that you think maybe we shouldn't pursue a I 1246 01:06:55,640 --> 01:06:58,600 Speaker 1: should we try to create a global moratorium on on 1247 01:06:59,280 --> 01:07:02,920 Speaker 1: general and tell agents No, I think it's impossible. I 1248 01:07:02,960 --> 01:07:06,240 Speaker 1: think I think there's no turning back. It's what it's. 1249 01:07:06,240 --> 01:07:09,600 Speaker 1: What we're doing, is what we're going to to be doing, uh, 1250 01:07:09,640 --> 01:07:11,960 Speaker 1: And the only way that gets interrupted is is via 1251 01:07:12,080 --> 01:07:14,760 Speaker 1: just absolute catastrophe. And I am not I am not 1252 01:07:14,840 --> 01:07:19,840 Speaker 1: pro catastrophe, but it's gonna it's like, like like all technologies, 1253 01:07:19,920 --> 01:07:22,120 Speaker 1: is simply going to be a matter of what extent 1254 01:07:22,200 --> 01:07:28,640 Speaker 1: can we prepare for and navigate the the moral problems 1255 01:07:28,640 --> 01:07:31,400 Speaker 1: that arise and there or and are we going to 1256 01:07:31,440 --> 01:07:35,760 Speaker 1: be able to have the foresight, uh, to see them 1257 01:07:35,800 --> 01:07:37,680 Speaker 1: before they're here? Yeah? I think that's what a lot 1258 01:07:37,720 --> 01:07:39,960 Speaker 1: of AI theorists say, is that you know, it's not 1259 01:07:40,040 --> 01:07:42,320 Speaker 1: like we can stop it. You know, you can't. You 1260 01:07:42,360 --> 01:07:44,240 Speaker 1: can't put a wall in front of this train. The 1261 01:07:44,280 --> 01:07:46,760 Speaker 1: train is gonna bust through the wall. So instead we 1262 01:07:46,760 --> 01:07:50,000 Speaker 1: should be intensely concerned with charting where the tracks go 1263 01:07:50,120 --> 01:07:52,280 Speaker 1: and making sure they go in a good direction. Yeah. 1264 01:07:52,280 --> 01:07:54,680 Speaker 1: I mean it comes back to like a simpler model 1265 01:07:54,720 --> 01:07:57,440 Speaker 1: of all. This is our episode on the Great eyeball 1266 01:07:57,480 --> 01:08:00,600 Speaker 1: Wars and our social media and our smartphones. It's like 1267 01:08:01,680 --> 01:08:05,440 Speaker 1: we can certainly throw our phone into a pond and 1268 01:08:05,480 --> 01:08:07,680 Speaker 1: then head off into the woods and try and live there, 1269 01:08:07,720 --> 01:08:13,080 Speaker 1: but most of us probably cannot um cannot follow that path. Therefore, 1270 01:08:13,120 --> 01:08:15,680 Speaker 1: we just have the best manage what we have. Yeah, 1271 01:08:15,680 --> 01:08:18,920 Speaker 1: wouldn't it be better to try to encourage the development 1272 01:08:18,920 --> 01:08:21,280 Speaker 1: of a phone full of apps that helped you fulfill 1273 01:08:21,360 --> 01:08:25,160 Speaker 1: your goals and aligned with your values? Alright, So we'll 1274 01:08:25,240 --> 01:08:27,640 Speaker 1: end it there. I feel like we not only have 1275 01:08:27,760 --> 01:08:30,599 Speaker 1: we provided food for thought here, we've provided a just 1276 01:08:30,640 --> 01:08:35,280 Speaker 1: a buffet of thought provoking ideas. And I know that 1277 01:08:35,360 --> 01:08:38,479 Speaker 1: everyone out there, all of you conscious listeners, are going 1278 01:08:38,520 --> 01:08:41,720 Speaker 1: to have something to contribute to this cover conversation, and 1279 01:08:41,760 --> 01:08:44,760 Speaker 1: we would love to hear from you and interact with you. 1280 01:08:44,760 --> 01:08:46,400 Speaker 1: You can find us in a number of ways. First 1281 01:08:46,439 --> 01:08:48,080 Speaker 1: of all, stuff toble your mind. Dot com is the 1282 01:08:48,080 --> 01:08:51,160 Speaker 1: mothership that's where you will find all of the podcast episodes, 1283 01:08:51,360 --> 01:08:53,400 Speaker 1: you'll find blog posts, and you'll find links out to 1284 01:08:53,400 --> 01:08:57,400 Speaker 1: our various social media accounts such as Facebook, Twitter, Instagram, 1285 01:08:57,920 --> 01:09:02,200 Speaker 1: and just basic contact information for us, thanks as always 1286 01:09:02,200 --> 01:09:05,360 Speaker 1: to our excellent audio producers Alex Williams and Tory Harrison. 1287 01:09:05,400 --> 01:09:07,120 Speaker 1: If you would like to get in touch with us 1288 01:09:07,120 --> 01:09:09,679 Speaker 1: with feedback about this episode or any other, to suggest 1289 01:09:09,800 --> 01:09:12,200 Speaker 1: a topic for the future, or just to say hi, 1290 01:09:12,320 --> 01:09:14,000 Speaker 1: let's know how you found out about the show, you 1291 01:09:14,040 --> 01:09:16,760 Speaker 1: can always email us at Blow the Mind at how 1292 01:09:16,840 --> 01:09:29,280 Speaker 1: stuff works dot com for more on this and thousands 1293 01:09:29,280 --> 01:09:54,439 Speaker 1: of other topics. Is it how stuff works dot com.