WEBVTT - From the Vault: Beyond the Uncanny Valley 0:00:05.840 --> 0:00:07.840 Hey, welcome to Stuff to Blow Your Mind. My name 0:00:07.880 --> 0:00:10.640 is Robert Lamb and I'm Joe McCormick, and it's Saturday. 0:00:10.760 --> 0:00:12.840 Time to go into the vault. This time we're bringing 0:00:12.880 --> 0:00:16.480 you part two of our exploration of the Uncanny Valley. 0:00:16.600 --> 0:00:21.880 This episode was originally published April six. Uh, should we 0:00:21.960 --> 0:00:28.360 jump right in, Robert, Let's do it. Welcome to Stuff 0:00:28.360 --> 0:00:37.320 to Blow your Mind from how Stop works dot com. Hey, 0:00:37.479 --> 0:00:39.159 welcome to Stuff to Blow your Mind. My name is 0:00:39.240 --> 0:00:41.640 Robert Lamb and I'm Joe McCormick. In Today is going 0:00:41.720 --> 0:00:44.080 to be the second part of a two part series 0:00:44.120 --> 0:00:47.519 we're doing on the Uncanny Valley. Last time we ventured 0:00:47.640 --> 0:00:50.600 into the Uncanny Valley, so if you haven't heard that episode, 0:00:50.720 --> 0:00:52.440 you should go back and listen to that. First, we 0:00:52.479 --> 0:00:54.240 lay a lot of the groundwork for what we're gonna 0:00:54.280 --> 0:00:57.320 be talking about today, But we explored the origination of 0:00:57.360 --> 0:01:00.360 the concept of the Uncanny Valley, what it means means 0:01:00.480 --> 0:01:03.720 to to be in the Uncanny Valley, and some research 0:01:03.800 --> 0:01:08.000 on whether this valley actually exists or not. Today, I 0:01:08.040 --> 0:01:10.320 think we want to start off by looking at if, 0:01:10.360 --> 0:01:14.600 if it does exist, what might be some explanations for it. Indeed. Yeah, 0:01:14.640 --> 0:01:17.080 we're gonna we're gonna dive into it a bit more 0:01:17.120 --> 0:01:22.160 and move as the title, uh suggests, beyond the Uncanny Valley. 0:01:22.200 --> 0:01:24.160 But before we do that, I do want to talk 0:01:24.200 --> 0:01:27.400 about RoboCop. Of course you want to talk about RoboCop, 0:01:27.560 --> 0:01:30.000 because we talked about RoboCop pretty much every day. Yeah, 0:01:30.000 --> 0:01:33.040 it's it's it's an important film. Important films, I'll say, 0:01:33.160 --> 0:01:35.039 at least the at least the first two, arguably the 0:01:35.080 --> 0:01:37.360 third one to throwing the TV series if you like. 0:01:37.920 --> 0:01:42.039 But there was a TV series, Yeah, RoboCop TV series. 0:01:42.080 --> 0:01:43.200 It was one of those that would I think it 0:01:43.240 --> 0:01:47.240 would come on sci Fi or it came on various 0:01:47.280 --> 0:01:50.920 cable channels. I only have a vague awareness of it 0:01:51.520 --> 0:01:54.320 because it seemed to be a far lower key RoboCop 0:01:54.360 --> 0:01:58.560 type show. Okay, So in many of the studies that 0:01:58.600 --> 0:02:00.800 we talked about in the last episode, they were looking 0:02:00.840 --> 0:02:07.000 at largely three categories of of robots and humans and androids. 0:02:07.000 --> 0:02:10.120 So you had pure robots things that are just undeniably 0:02:10.240 --> 0:02:14.760 machines and we're mostly okay with. Then you have humans 0:02:15.040 --> 0:02:18.880 or or or perfect human replications. Okay, So you look 0:02:18.880 --> 0:02:20.880 at it, it either is a human or it's such 0:02:20.919 --> 0:02:25.000 a good representation of a human ideally that you would 0:02:25.040 --> 0:02:27.360 not think that it was a robot. Right. The third 0:02:27.440 --> 0:02:29.200 category here is where you're going to get into the 0:02:29.280 --> 0:02:33.320 danger zone, right, the human like robots, where you look 0:02:33.360 --> 0:02:35.160 at it and you say, I see what you're going 0:02:35.240 --> 0:02:39.600 for there, but it's creeping me out. So I think 0:02:39.600 --> 0:02:42.320 it's interesting to align this up with the Holy Trinity 0:02:42.320 --> 0:02:45.880 of robocops. Okay, so this will mostly make sense if 0:02:45.919 --> 0:02:48.200 you've seen the RoboCop films, but I feel like most 0:02:48.240 --> 0:02:50.440 people know what we're talking about here. First of all, 0:02:50.480 --> 0:02:53.799 you had the proto RoboCop and two oh nine. Oh, 0:02:53.919 --> 0:02:56.600 this is maybe the greatest scene in the first film 0:02:56.720 --> 0:02:59.880 is before we get a humanoid RoboCop, they just have 0:03:00.120 --> 0:03:03.760 this big drone object that is supposed to enforce the 0:03:03.840 --> 0:03:07.000 law and it ends up shooting somebody in the boardroom. Yeah, 0:03:07.040 --> 0:03:11.440 it's a walking law enforcement tank with a robot commanding 0:03:11.560 --> 0:03:14.560 robot voice, so that you look at it and there's 0:03:14.600 --> 0:03:17.400 no to not know this is not a friendly device. 0:03:17.639 --> 0:03:20.280 But it's not really humanoid at all, not really at all. 0:03:20.280 --> 0:03:23.200 It hit just walks on big, big legs that can't 0:03:23.200 --> 0:03:26.840 even navigate human stairs. But I would say, I have 0:03:26.919 --> 0:03:29.359 great affinity for ed to oh nine, Yeah, I mean, 0:03:29.639 --> 0:03:32.280 you know, without having to worry about it actually shooting me. 0:03:32.960 --> 0:03:36.000 It's kind of cute in a way. Now, then we 0:03:36.080 --> 0:03:42.480 have RoboCop itself himself, the classic RoboCop, the Peter Peter Weller, 0:03:43.240 --> 0:03:47.240 and uh, he is a cyborg or perhaps an android, 0:03:47.280 --> 0:03:50.760 depending on how you want to view the descriptions. So 0:03:50.840 --> 0:03:55.760 he's he has a relatable, living human face which is 0:03:55.760 --> 0:03:59.680 a fixed to to honor him in some some explanations, 0:03:59.800 --> 0:04:02.200 or has to make him more comfortable not only as 0:04:02.240 --> 0:04:05.640 a police killing machine but also a community law enforcement officer. 0:04:05.760 --> 0:04:09.280 So RoboCop, you know, moves around with very robotic movements, 0:04:09.600 --> 0:04:13.000 speaks in a very robotic voice, but his face is 0:04:13.120 --> 0:04:17.159 a living human face. So in a way like that, 0:04:17.160 --> 0:04:19.160 that seems like it might it just sort of you know, 0:04:19.200 --> 0:04:22.400 to read perhaps more into the original film than was intended. 0:04:22.680 --> 0:04:25.680 Perhaps this was a way to to get beyond the 0:04:25.760 --> 0:04:28.440 Uncanny Valley. We can't replicate the human face, We'll just 0:04:28.520 --> 0:04:31.160 get an actual human face from the dead cop and 0:04:31.200 --> 0:04:33.719 just plaster it up there. Oh but I'd say RoboCop 0:04:33.760 --> 0:04:35.839 with his mask on really does kind of get into 0:04:35.880 --> 0:04:38.880 the uncanny Valley and and Weller does some good work 0:04:39.120 --> 0:04:41.640 forcing us in there, Like I think they're the sort 0:04:41.680 --> 0:04:45.880 of going for it. Yeah, okay, Well that brings us 0:04:45.920 --> 0:04:50.160 to the next generation RoboCop two, which is not just 0:04:50.240 --> 0:04:53.080 the name of the second RoboCop movie, but also the 0:04:53.560 --> 0:04:57.400 model of RoboCop that replaced the original RoboCop. They think, Hey, 0:04:57.440 --> 0:04:59.479 what would happen if we put tom noon in there? 0:04:59.680 --> 0:05:03.880 That's right, So they have another essentially a walking tank 0:05:04.080 --> 0:05:06.440 kind of like d to O nine, except it's powered 0:05:06.440 --> 0:05:09.320 by the brain of a psychotic drug lord named Kane 0:05:09.800 --> 0:05:13.760 played by Tom Noonan. He's fabulous in it. Uh, but 0:05:13.839 --> 0:05:16.000 here's the here's the thing. It's walking around, it's killing 0:05:16.040 --> 0:05:18.320 everything with a gatling gun, but then it can pop 0:05:18.520 --> 0:05:20.800 a flat screen TV out of it out of the 0:05:21.000 --> 0:05:24.360 front of its body, and on that screen is a 0:05:24.400 --> 0:05:28.360 twisted uncanny lawnmower Man esque c g I face of 0:05:28.440 --> 0:05:33.000 tom Noonon. Yeah, so that one really leans into the 0:05:33.200 --> 0:05:36.440 Uncanny Valley. Well yeah, and this this does point out 0:05:36.480 --> 0:05:39.080 another thing, which is that there have been plenty of 0:05:39.680 --> 0:05:44.320 intentional realizations of the Uncanny Valley in film when when 0:05:44.360 --> 0:05:48.520 people are trying to create an unsettling, unpleasant, humanoid for 0:05:48.600 --> 0:05:51.200 story purposes, if it's supposed to be a villain, if 0:05:51.200 --> 0:05:53.960 it's supposed to make people uncomfortable, because that's its role 0:05:53.960 --> 0:05:56.080 in the plot. So one thing I kind of wish 0:05:56.080 --> 0:05:58.080 we'd done. I hadn't even thought about this is too 0:05:58.640 --> 0:06:02.080 if we could talk to somebody who has intentionally made 0:06:02.160 --> 0:06:05.200 things in the Uncanny Valley? What did they do on 0:06:05.400 --> 0:06:09.039 purpose to get it there? If your your job is 0:06:09.080 --> 0:06:12.480 to make a humanoid robot or an animated humanoid figure 0:06:12.760 --> 0:06:16.680 that intentionally pushes all the bad buttons and climbs as 0:06:16.680 --> 0:06:18.800 far down into the valley as it can, what do 0:06:18.880 --> 0:06:22.200 you do that would that would provide some really interesting 0:06:22.200 --> 0:06:25.560 insight into what it actually takes to get there. Well, 0:06:25.560 --> 0:06:27.359 you know, to come back to video games, something that 0:06:27.400 --> 0:06:29.960 comes up a lot. You see these videos going viral 0:06:30.040 --> 0:06:32.760 where it's like just a cut scene or a clip 0:06:32.800 --> 0:06:35.680 from the video game with a humanoid character in a 0:06:35.720 --> 0:06:39.200 more or less human situation, except something screwed up in 0:06:39.279 --> 0:06:41.800 the face is missing, so it's just two floating eyeballs 0:06:41.800 --> 0:06:45.640 and maybe a floating set of digital teeth. So the 0:06:45.680 --> 0:06:48.280 context is key there, Like this thing is acting as 0:06:48.279 --> 0:06:50.280 if it had a face, and it's in an environment 0:06:50.320 --> 0:06:53.440 where we're unsupposed to just roll with it, but clearly 0:06:53.520 --> 0:06:56.720 something's wrong. Yes, okay, Well, as we said in the 0:06:56.800 --> 0:06:58.919 last episode, we talked about the origin of the idea, 0:06:59.000 --> 0:07:01.600 We talked about evidence for and against the fact that 0:07:01.640 --> 0:07:05.919 the Uncanny Valley actually exists, to to the point that 0:07:06.000 --> 0:07:08.719 it actually does exist to some extent. Maybe not in 0:07:08.760 --> 0:07:11.880 the naive original sense everybody would say, where it's just 0:07:12.160 --> 0:07:17.400 related to the amount of realistic human nous in a figure, 0:07:17.440 --> 0:07:22.000 but has other dimensions as well. What causes our Uncanny 0:07:22.080 --> 0:07:25.360 Valley reaction? Obviously people do have this reaction they see 0:07:25.360 --> 0:07:30.000 a humanoid robot or a humanoid animated character, the Scorpion King, 0:07:30.160 --> 0:07:33.400 the Final Fantasy, the spirits within characters, whatever it is, 0:07:33.640 --> 0:07:38.280 and we get so, what causes it? Why do our 0:07:38.280 --> 0:07:42.480 brains react that way? Is it biological? Is it pure instinct? 0:07:42.600 --> 0:07:45.600 Is it a learned psychological reaction? Is it part of 0:07:45.600 --> 0:07:50.600 our culture? Is something coming from cognitive dissonance of some sort? Now, 0:07:50.640 --> 0:07:52.600 to go back to the origins of the idea with 0:07:52.840 --> 0:07:56.000 massive hero Mari in nineteen seventy and his original paper 0:07:56.040 --> 0:08:00.440 on the Uncanny Valley, Morey speculates that the Uncanny Valley 0:08:00.520 --> 0:08:03.880 might be a side effect of the self preservation instinct. 0:08:04.040 --> 0:08:07.200 In other words, it's a biological adaptation that helps us 0:08:07.240 --> 0:08:11.040 avoid disease and death. Uh. And he starts with the 0:08:11.080 --> 0:08:14.760 observation that when a normal, healthy person becomes sick and 0:08:14.800 --> 0:08:19.160 eventually dies, they basically tend to slide down from the 0:08:19.240 --> 0:08:22.000 second peak of the uncanny valley, what we were calling 0:08:22.000 --> 0:08:27.200 the realism peak, the reality peak, and slide down into 0:08:27.400 --> 0:08:30.800 the uncanny valley. It's like they become more like these 0:08:30.880 --> 0:08:34.360 upsetting puppets and robots. You know. They might uh suffer 0:08:34.480 --> 0:08:38.360 some kind of change in their in their the the 0:08:38.360 --> 0:08:42.160 appearance of their skin, of their eyes, of their facial expressions. 0:08:42.200 --> 0:08:45.960 Things begin to look off about them. And he writes, quote, 0:08:46.800 --> 0:08:49.840 the sense of eeriness is probably a form of instinct 0:08:49.880 --> 0:08:54.440 that protects us from proximal rather than distal sources of danger. 0:08:54.960 --> 0:08:59.640 Proximal sources of danger include corpses, members of different species, 0:09:00.000 --> 0:09:03.560 and other entities we can closely approach. Distal sources of 0:09:03.640 --> 0:09:08.200 danger include windstorms and floods. You know, this is interesting 0:09:08.200 --> 0:09:11.720 that the mention of other species. Um. I don't know 0:09:11.760 --> 0:09:15.360 how many films and documentaries I've seen of say, lions 0:09:15.440 --> 0:09:20.080 running around in the natural habitat, and it's almost never creepy. No, 0:09:20.200 --> 0:09:22.160 not not at all, and yet and then mostly a 0:09:22.240 --> 0:09:24.080 lot of times when I'm in a zoo, it's not creepy. 0:09:24.200 --> 0:09:26.439 But there have been times I take my son to 0:09:26.480 --> 0:09:29.400 the zoo a lot here in Atlanta, and there are 0:09:29.400 --> 0:09:31.880 times when we go down to the lion enclosure and 0:09:31.880 --> 0:09:33.760 we're the only people there at the zoo because we've 0:09:33.840 --> 0:09:37.040 arrived super early, and we just we're hanging out with 0:09:37.080 --> 0:09:39.400 a lion on the other side of the glass. You know, 0:09:39.400 --> 0:09:44.120 we're perfectly safe, but a deep uneasiness comes over me, 0:09:44.520 --> 0:09:48.559 washes over me, and uh, and I'm just confronted by 0:09:48.679 --> 0:09:52.240 the the the danger of this situation, Like there's a 0:09:52.320 --> 0:09:56.360 danger that that goes beyond any reason because I am 0:09:56.360 --> 0:10:00.000 in close proximity to a dangerous member of another species, 0:10:00.000 --> 0:10:04.320 a carnivorous um predator, tory animal that would, in under 0:10:04.320 --> 0:10:07.640 normal conditions, potentially kill me. Now there is I would 0:10:07.880 --> 0:10:12.160 bet a very different kind of sensation going on inside 0:10:12.200 --> 0:10:16.360 you when you're in proximity to a lion than when 0:10:16.400 --> 0:10:20.199 you see a creepy looking humanoid robot or a creepy 0:10:20.240 --> 0:10:25.000 looking animation human animation, right like it's it's probably exciting 0:10:25.040 --> 0:10:27.360 the difference between the when when we talked about the 0:10:27.360 --> 0:10:30.520 creepiness episode the creep the difference between the sort of 0:10:30.559 --> 0:10:35.360 like uncomfortable threat ambiguity versus sensing that there truly is 0:10:35.400 --> 0:10:37.880 a threat of some kind. Right, And you know, we 0:10:38.240 --> 0:10:41.400 spoke to illness here, you know, like certainly someone we 0:10:41.440 --> 0:10:43.439 talked about like what happens when a co worker walks 0:10:43.520 --> 0:10:46.640 up to you and they're they're sniveling or more a 0:10:46.679 --> 0:10:50.920 little bit pale, uh and they say, let's French. Well, 0:10:50.920 --> 0:10:53.720 well hopefully not, but yeah, there's a there's an at 0:10:53.760 --> 0:10:57.640 least an initial sense of, oh, what's wrong with this person? Um? 0:10:57.920 --> 0:10:59.760 I wonder if what they have as contagious, should they 0:10:59.800 --> 0:11:02.200 even be at work? I hope they cover their mouth 0:11:02.200 --> 0:11:04.600 when they sneeze, because we are going to be concerned 0:11:04.600 --> 0:11:06.840 on some level with catching whatever they have. If it's 0:11:06.960 --> 0:11:10.360 if it's it's transmittable, right and it I mean there 0:11:10.360 --> 0:11:13.440 are some ways in which we know that seeing illness 0:11:13.520 --> 0:11:18.240 and other people in sights a reaction in the disgust response, right. Yeah. 0:11:18.240 --> 0:11:19.960 And there's been actually a lot of research on the 0:11:20.000 --> 0:11:24.440 discussed the disgusted response. Darwin wrote about it discussed from 0:11:24.480 --> 0:11:29.000 an evolutionary standpoint, is all about disease avoidance, you know, 0:11:29.200 --> 0:11:33.560 like do not eat this, you know, stay away from Yeah, 0:11:33.600 --> 0:11:36.320 something's not right here. You might hurt yourself or get sick. 0:11:36.920 --> 0:11:41.200 So we're talking about sensuous disgust, So discuss that's tied 0:11:41.240 --> 0:11:44.680 to the senses to sense information that we're absorbing. Uh. 0:11:44.720 --> 0:11:47.560 This is deeply seated in the insula, the area of 0:11:47.559 --> 0:11:52.440 the brain that that malfunctions in patients with obsessive compulsive disorder, 0:11:52.920 --> 0:11:55.360 causing them to say, you know, wash and clean things 0:11:55.520 --> 0:12:00.400 endlessly or vacuum unrelenting lye. So the malfunk of that 0:12:00.520 --> 0:12:02.920 area gives us, you know, gives us a clue into 0:12:02.960 --> 0:12:07.680 its functionality. Uh. One interesting fact about disgusting smells, however, 0:12:08.120 --> 0:12:11.880 is that there's a drop off point for bad smell recognition, 0:12:12.080 --> 0:12:15.240 but not for good smells. So I think we've all 0:12:15.520 --> 0:12:18.640 encountered this where it's that like say, say you're in 0:12:18.679 --> 0:12:21.160 your your office and and you share your office with 0:12:21.200 --> 0:12:24.520 a cat box, and at some point the cat uh 0:12:24.880 --> 0:12:28.040 causes quite a stink in there, and you register it first, 0:12:28.040 --> 0:12:29.800 and like, geez, I should stop what I'm doing and 0:12:29.880 --> 0:12:32.480 clean out that cat box. But you keep working, and 0:12:32.520 --> 0:12:35.120 then after a while you don't smell it anymore. But 0:12:35.160 --> 0:12:37.720 then maybe you step outside to check the mail, or 0:12:37.880 --> 0:12:39.840 you go to the you know, the grocery store to 0:12:39.840 --> 0:12:41.960 pick something up, or you know, your partner comes home 0:12:42.280 --> 0:12:45.360 and when when you or they enter the room, you 0:12:45.440 --> 0:12:48.240 go cheese. What happened in here? Did the cat do 0:12:48.320 --> 0:12:51.520 something again? No, you're just re smelling with their Originally 0:12:51.640 --> 0:12:53.920 you forgot about it, right, Yeah, the the brain kind 0:12:53.920 --> 0:12:57.360 of decides, Look at this point, I assume you know 0:12:57.559 --> 0:13:00.440 that she that the cheese is nasty and you're not 0:13:00.480 --> 0:13:03.920 going to eat it um or or that yes, there 0:13:04.000 --> 0:13:06.199 is animal poop around here. It's it's done, it's part 0:13:06.240 --> 0:13:10.079 it's warned you. But the good smell will keep resonating 0:13:10.200 --> 0:13:13.400 because the good smell is probably saying, hey, there's something 0:13:13.400 --> 0:13:16.440 over here, delicious to eat with some fresh berries or whatnot, 0:13:16.720 --> 0:13:19.640 And it'll keep saying, hey, the berries are still here, 0:13:19.720 --> 0:13:22.400 why haven't you eaten them yet? This is there's not 0:13:22.440 --> 0:13:24.439 a lot of sugar out here. You should get at 0:13:24.480 --> 0:13:28.200 these berries while you have a chance. So the beautiful 0:13:28.240 --> 0:13:32.360 remains beautiful, The sweet smelling remains sweet smelling. But something 0:13:32.360 --> 0:13:35.320 that is disgusting even as you know, discussing as a 0:13:35.360 --> 0:13:38.160 foul smell we can grow accustomed to. Okay, so what's 0:13:38.200 --> 0:13:42.400 the analogy to the Uncanny Valley here? Well, I think 0:13:42.400 --> 0:13:45.640 the the analogy here is that if you have a 0:13:45.760 --> 0:13:51.319 disgusted response to the to visual sense information regarding an 0:13:51.360 --> 0:13:56.200 individual's appearance, a robots appearance, of a computer animated characters appearance, 0:13:57.040 --> 0:13:59.480 there could there could also be this disgust drop off 0:13:59.480 --> 0:14:01.640 point as you grow accustomed to it. Oh, and that's 0:14:01.679 --> 0:14:03.440 something that we have seen. We talked about a little 0:14:03.440 --> 0:14:06.640 bit in the last episode, and that uh, some people 0:14:06.720 --> 0:14:11.559 attest that when you spend time around these robots or uh, 0:14:11.600 --> 0:14:14.080 well it was mainly for the robots. The robots that 0:14:14.120 --> 0:14:18.079 at first seemed creepy, they stop bothering you become accustomed 0:14:18.120 --> 0:14:20.800 to them, they're not creepy anymore. I don't so much 0:14:20.920 --> 0:14:22.640 know if that's always going to be the case for 0:14:22.680 --> 0:14:27.560 creepy looking animated humans. Um, but who knows. Now. Another 0:14:27.600 --> 0:14:29.680 thing to keep in mind too, is that the discussed 0:14:29.680 --> 0:14:32.280 response is going to depend on a number of different factors, 0:14:32.880 --> 0:14:36.240 some of which are going to be tied to hormonal situations. So, 0:14:36.360 --> 0:14:39.720 for instance, pregnant women are more sensitive to discussed and 0:14:39.720 --> 0:14:43.800 this is linked to their elevated progesterone levels. So uh, 0:14:44.040 --> 0:14:45.840 And of course there's gonna be other factors beyond that 0:14:45.960 --> 0:14:49.120 for every individual. Yeah. So I think that there's certainly 0:14:49.160 --> 0:14:54.800 could be some amount of biological instinctual response going on 0:14:54.880 --> 0:14:57.680 in the uncanny value effect to the extent that it exists. 0:14:57.680 --> 0:14:59.880 But I think also, based on what we've seen so far, 0:15:00.160 --> 0:15:03.160 probably does not account for all of it. Uh. And 0:15:03.480 --> 0:15:06.920 I think another thing to consider would be going to 0:15:07.120 --> 0:15:12.080 more more complex sort of cognitive psychology, such as cognitive dissonance. Now, 0:15:12.520 --> 0:15:15.760 if you were to just ask me what I thought 0:15:15.880 --> 0:15:19.360 was the most likely answer before I got into the research, 0:15:20.160 --> 0:15:22.600 I would intuitively tend to think that the best answer 0:15:22.640 --> 0:15:27.000 for what causes the Uncanny value effect primarily is our 0:15:27.040 --> 0:15:31.240 inherent discomfort with category confusions. Uh. This is something that 0:15:31.280 --> 0:15:33.600 I think about a lot, in like the creation of 0:15:33.640 --> 0:15:36.120 monster mythology and stuff like that. We we don't like 0:15:36.200 --> 0:15:40.000 the feeling produced by by things that don't fit into 0:15:40.040 --> 0:15:44.080 our normal taxonomys for objects in the world and seem 0:15:44.160 --> 0:15:48.160 to violate our tax taxonomic ordering system. And this is 0:15:48.200 --> 0:15:51.160 why I think monsters are so often hybrids of existing 0:15:51.240 --> 0:15:55.120 things a bull's head on a man's body, things that 0:15:55.320 --> 0:16:00.200 defy our intuitive classification rules. They make us uncomfortable and 0:16:00.320 --> 0:16:03.520 cause a sense of unease, leading to this uncanny feeling. 0:16:04.400 --> 0:16:07.560 So that that's what I would have intuitively said, Yeah, 0:16:07.600 --> 0:16:10.920 that makes sense. Is it wolf or is it man? 0:16:10.280 --> 0:16:15.280 Or right? But negative affinity resulting from this difficulty and 0:16:15.320 --> 0:16:19.320 assigning the entity to a category is it robot or human? 0:16:19.840 --> 0:16:23.400 Despite my intuitive favor for this explanation, I think it 0:16:23.440 --> 0:16:26.360 looks like experimental evidence for this is not strong, and 0:16:26.400 --> 0:16:28.640 in fact, in some ways some of the studies we've 0:16:28.640 --> 0:16:31.760 looked at have somewhat invalidated this. For example, in the 0:16:31.840 --> 0:16:33.880 last episode at the end, I was talking about that 0:16:34.160 --> 0:16:38.560 study by matherin Rifling, and it did not find evidence 0:16:38.560 --> 0:16:42.280 of a strong correlation between here. Here's what you notice 0:16:42.320 --> 0:16:46.080 here the time it took people to rate the mechano 0:16:46.200 --> 0:16:50.960 humanoid qualities of a robot, and that robot pictures likability. 0:16:51.040 --> 0:16:54.480 So if you were talking it up to category confusion, 0:16:54.840 --> 0:16:57.400 you would probably think the robots that people took the 0:16:57.560 --> 0:17:00.440 longest time to figure out how to ray on the 0:17:00.480 --> 0:17:04.880 mechano humanoid scale, and those would be the least likable, right, 0:17:04.920 --> 0:17:09.520 Because they're the ones causing the most category confusion, right. 0:17:09.560 --> 0:17:11.000 But I mean you could also say that, well, you're 0:17:11.040 --> 0:17:13.360 not maybe having not a you're not having a visceral 0:17:13.440 --> 0:17:15.720 gut reaction to them either, you having to think it 0:17:15.760 --> 0:17:18.760 out and try and you know, analyze how you feel 0:17:18.800 --> 0:17:20.639 about it. Oh yeah, I'd say that's exactly what we 0:17:20.680 --> 0:17:23.119 don't like. I mean, we like to be able to 0:17:23.240 --> 0:17:27.080 viscerally categorize things. But but then the uncanny Valley is often, 0:17:27.680 --> 0:17:29.480 at least in terms of the way you're going to 0:17:29.560 --> 0:17:33.399 find it invoked by the average person. It's often discussed 0:17:33.400 --> 0:17:35.840 if it's as if it's a visceral reaction, the sort 0:17:35.880 --> 0:17:38.480 of i'll kill it with fire reaction that people might have, 0:17:38.760 --> 0:17:40.840 where I hate that saying yes, I'm not a fan 0:17:40.880 --> 0:17:43.960 as well, especially when it is applied to outside of 0:17:43.960 --> 0:17:48.040 of a fictional connotations. Right. But so they point out 0:17:48.080 --> 0:17:51.440 the authors here point out that they did not find this. Uh. 0:17:51.480 --> 0:17:54.760 They point out that it's just not a fact that 0:17:54.880 --> 0:17:57.400 the things that took longer to look at and decide 0:17:57.440 --> 0:18:01.360 where the least likable. Though while this is not statistically important, 0:18:01.680 --> 0:18:04.199 just as a point of curiosity, the single face that 0:18:04.240 --> 0:18:08.040 took the longest to rate on the mechanical versus humanoid quality. 0:18:08.440 --> 0:18:11.520 Was also just about the most disliked face in their 0:18:11.520 --> 0:18:15.400 whole collection of faces, but that was just like one outlier. Overall, 0:18:15.440 --> 0:18:19.680 this did not present as a general effect. Other studies 0:18:19.680 --> 0:18:22.280 have also looked into this and have failed to find 0:18:22.320 --> 0:18:25.679 solid evidence for category confusion as the primary driver of 0:18:25.720 --> 0:18:29.280 the negative affinity at the bottom of the uncanny valley. 0:18:29.320 --> 0:18:31.879 So it looks like my intuitions here I think are wrong. 0:18:32.600 --> 0:18:35.960 But there's something that's kind of related as an idea 0:18:36.000 --> 0:18:40.040 that's been explored, and that is the idea of perceptual mismatch. 0:18:40.640 --> 0:18:44.480 So several authors have advocated the idea that this perceptual 0:18:44.600 --> 0:18:48.280 mismatch could be the primary cause of what we don't 0:18:48.320 --> 0:18:51.480 like about things that we would intuitively say fall into 0:18:51.520 --> 0:18:54.760 the uncanny valley. So one piece of research I want 0:18:54.760 --> 0:18:57.960 to mention is a sort of review by uh Cat 0:18:58.040 --> 0:19:01.439 Siri that is a review of empirical evidence on different 0:19:01.520 --> 0:19:05.960 uncanny valley hypotheses support for perceptual mismatch as one road 0:19:06.040 --> 0:19:08.520 to the valley of eerieness that gott to give it 0:19:08.560 --> 0:19:13.760 a different name, apparently in Frontiers in Psychology. And so 0:19:13.800 --> 0:19:17.399 in this study, the authors review present research and claim 0:19:17.520 --> 0:19:20.640 that experimental research attempting to show the Uncanny Valley has 0:19:20.640 --> 0:19:24.040 been inconsistent. They don't exactly say that the uncanny valley 0:19:24.080 --> 0:19:27.119 doesn't exist, but that it's not as simple as often 0:19:27.160 --> 0:19:29.240 believed to be. Something we've been saying for a while now. 0:19:29.320 --> 0:19:33.680 It's it's not that any manipulation of the variable of 0:19:33.800 --> 0:19:38.520 human likeness leads to Uncanny Valley effects. So, in other words, 0:19:38.600 --> 0:19:41.640 that the horizontal axis on the graph is more complicated 0:19:41.720 --> 0:19:44.520 than just the question of how realistically human is it. 0:19:44.600 --> 0:19:46.560 I've seen this come up enough now that I'm pretty 0:19:46.560 --> 0:19:50.160 convinced that that is not necessarily the only or even 0:19:50.200 --> 0:19:53.680 the primary factor here. But they still recognize that there 0:19:53.720 --> 0:19:55.520 is some kind of effect here. So they claim that 0:19:55.560 --> 0:19:59.000 there's evidence against the category confusion basis that we were 0:19:59.040 --> 0:20:01.520 just talking about. But they claim that there has been 0:20:01.520 --> 0:20:05.159 good evidence in support of the perceptual mismatch hypothesis. And 0:20:05.200 --> 0:20:07.040 I want to read what they say. They say quote 0:20:07.280 --> 0:20:11.440 taken together, the present review suggested that although not any 0:20:11.520 --> 0:20:14.600 kind of human likeness manipulation leads to the Uncanny Valley, 0:20:14.640 --> 0:20:18.840 the uncanny valley could be caused by more specific perceptual 0:20:18.920 --> 0:20:23.720 mismatch conditions. Such conditions could originate at least from inconsistent 0:20:23.880 --> 0:20:28.560 realism levels between individual features, and the examples they give 0:20:28.600 --> 0:20:33.240 would be like artificial eyes on a humanlike face, or 0:20:33.280 --> 0:20:36.840 the presence of atypical features such as a typically large 0:20:36.920 --> 0:20:41.119 eyes on an otherwise humanlike character. So what they're saying 0:20:41.160 --> 0:20:44.720 there is not necessarily that you can't tell whether it's 0:20:44.720 --> 0:20:47.320 a robot or a human, but that there have been 0:20:47.400 --> 0:20:50.840 multiple experiments that seem to show people are unsettled and 0:20:51.160 --> 0:20:55.439 made unhappy by things where the features on the face 0:20:55.760 --> 0:20:58.400 or the features of the figure as a whole are 0:20:58.600 --> 0:21:03.320 inconsistently realistic. Like, we're more okay with a robot that's 0:21:03.400 --> 0:21:08.000 consistently realistic at a certain level it maybe say, seventy 0:21:08.040 --> 0:21:11.800 percent than something that has eyes at nine d percent 0:21:11.960 --> 0:21:16.880 and skin at This reminds me, this is just coming 0:21:16.880 --> 0:21:18.000 off the top of my head. So I don't have 0:21:18.040 --> 0:21:21.240 the artist's name here, but there's there's an artist who's 0:21:21.240 --> 0:21:24.080 worth made the rounds where they took cartoon characters and 0:21:24.119 --> 0:21:27.600 they depicted them realistically. So it's Homer Simpson. Yes, Homer 0:21:27.640 --> 0:21:32.040 Simpson with like pores on his skin, you know, horrible. Yeah, 0:21:32.320 --> 0:21:33.879 So that comes to mind is a kind of a 0:21:33.880 --> 0:21:37.040 possible example of this. Yeah. I think that's a good explanation. 0:21:37.160 --> 0:21:39.479 So I want to get into my main takeaways from 0:21:39.520 --> 0:21:41.840 looking at the Uncanny Valley research so far. Maybe you 0:21:41.840 --> 0:21:44.280 can let me know what you think about this. I'd say, 0:21:44.359 --> 0:21:46.560 first of all, I think the Uncanny Valley is a 0:21:46.640 --> 0:21:49.880 real thing, but it's not as simple as Maury's original 0:21:49.960 --> 0:21:53.280 hypothesis would lead you to believe. Uh. First of all, 0:21:53.320 --> 0:21:56.960 people definitely do get creeped out by lots of almost 0:21:57.080 --> 0:22:01.120 human looking things, but it's not a necessarily just that 0:22:01.240 --> 0:22:06.000 the near failed human realism is what makes them unsettling. 0:22:06.000 --> 0:22:08.119 There are other things that appear to be making them 0:22:08.240 --> 0:22:12.600 unsettling though, that the near humanness plays some kind of role. 0:22:13.400 --> 0:22:15.040 And the other big thing is that there appear to 0:22:15.040 --> 0:22:19.359 be multiple dimensions to explain the phenomenon. Right, So, synthetic 0:22:19.440 --> 0:22:24.080 humanoid images, whether robotic or animated, offer multiple dimensions of 0:22:24.119 --> 0:22:27.439 attraction and revulsion. I think it's possible that there are 0:22:27.480 --> 0:22:31.720 some biologically triggered effects the appearance of health or disease, 0:22:31.880 --> 0:22:34.440 the appearance of life or death. But then I think 0:22:34.440 --> 0:22:38.080 there are possibly other things triggered by psychological cognitive dissonance, 0:22:38.119 --> 0:22:43.640 probably not category confusion, um, but but some good evidence 0:22:43.680 --> 0:22:46.639 for this idea of the perceptual mismatch being the cause. 0:22:47.440 --> 0:22:50.520 And then the final thing is that the Uncanny Valley 0:22:50.600 --> 0:22:54.480 effect is context dependent. How long have you been exposed 0:22:54.480 --> 0:22:57.800 to the image in what's setting? Is it part of 0:22:57.840 --> 0:23:00.400 a narrative or some other context in which you're being 0:23:00.480 --> 0:23:04.040 asked to suspend your disbelief or otherwise put yourself in 0:23:04.080 --> 0:23:07.280 a state of openness. Robert, what do you think about 0:23:07.280 --> 0:23:10.199 all this so far? You've got any disagreement? No, I 0:23:10.200 --> 0:23:12.800 mean I I feel like my view on it closely 0:23:12.880 --> 0:23:15.560 lines up with with with yours here basically that it's 0:23:15.560 --> 0:23:18.200 just that there is an effect going on, but it's 0:23:18.280 --> 0:23:22.120 far it's far more nuanced than simply oh these are 0:23:22.320 --> 0:23:25.280 these are the factors that make something fall into the 0:23:25.359 --> 0:23:29.240 Uncanny Valley, not just how realistically human. There's there's other 0:23:29.359 --> 0:23:32.240 stuff going on, right, and not just a mere hybridization. 0:23:33.119 --> 0:23:36.800 It was instantly thinking about um, the borg and UH 0:23:36.840 --> 0:23:40.600 and occasionally the sort of sexy boards that show up 0:23:40.920 --> 0:23:44.280 in the Star Trek universe. No, like, there's there's clearly 0:23:44.320 --> 0:23:48.800 category confusion going on there, but they're they're not depicted 0:23:48.880 --> 0:23:53.760 as particularly uncanny, Like was the board queen? Was she uncanny? 0:23:54.400 --> 0:23:56.400 Kind of? I don't know, I mean, even when I'm 0:23:56.400 --> 0:24:00.240 not a big fan of the way the board look well, 0:24:01.440 --> 0:24:04.119 but the outside of the board. You can also think of, 0:24:04.560 --> 0:24:08.280 you know, various sort of hybrid human creatures depicted in 0:24:08.359 --> 0:24:12.040 fantasy and fiction that that are created in such a 0:24:12.080 --> 0:24:15.639 way to be the alluring, uh, like they managed to 0:24:15.680 --> 0:24:21.120 fetishize the inhuman qualities of them the category of confusion. Yeah, 0:24:21.440 --> 0:24:23.760 uh yeah. It makes me think that there are certain 0:24:23.840 --> 0:24:27.600 qualities of the human appearance that, if altered, are much 0:24:27.640 --> 0:24:31.120 more significant in terms of our affects response than others. 0:24:31.800 --> 0:24:34.000 So it could be and I'm just making this up. 0:24:34.040 --> 0:24:36.560 I don't know if this is true, but that like, uh, 0:24:36.960 --> 0:24:41.320 getting the getting the size of the eyes wrong could 0:24:41.440 --> 0:24:45.600 quite easily lead to a disgust response and revulsion. But 0:24:45.680 --> 0:24:49.040 getting the size of the nose wrong wouldn't. Does that 0:24:49.080 --> 0:24:52.240 make sense? Yeah? Yeah? Or just thinking of eyes, like 0:24:52.359 --> 0:24:56.720 definitely making the eyes inappropriately large leads a creepy factor, 0:24:56.760 --> 0:24:59.440 and this is often employed. I instantly think of the 0:24:59.760 --> 0:25:03.240 Van Piate movie. What twenty thirty Days of Night? Is 0:25:03.240 --> 0:25:05.879 that the name of it? Oh? Something like that. Yeah, 0:25:05.880 --> 0:25:07.800 where they did some sort of digital effect to make ority. 0:25:07.840 --> 0:25:10.960 I don't remember the number, some number of days of PPD, 0:25:11.040 --> 0:25:14.119 days of night. Um. I didn't actually see if the 0:25:14.160 --> 0:25:18.560 trailers were we're certainly interesting. But likewise, if you just 0:25:18.600 --> 0:25:21.600 take an individual and have them wear blackout contact lenses 0:25:22.080 --> 0:25:24.560 like that is often sometimes that's played up for creepiness, 0:25:24.560 --> 0:25:26.159 but a lot of times it's played up for to 0:25:26.240 --> 0:25:30.359 be alluring. You'll have uh, male or female characters that 0:25:30.440 --> 0:25:34.119 are are otherwise dressed in some alluring fashion, but they 0:25:34.119 --> 0:25:36.040 have blacked out eyes and it's who It's kind of 0:25:36.080 --> 0:25:39.640 like supernatural, sexy, cool, as opposed to just like, oh 0:25:39.680 --> 0:25:43.120 my god, why why are your eyes pits of darkness? Yeah, 0:25:44.040 --> 0:25:46.119 but maybe that's just me. Okay, well, I think we 0:25:46.119 --> 0:25:48.040 should take a quick break and when we come back, 0:25:48.200 --> 0:25:57.159 we will go beyond the Uncanny Valley. All right, we're back, Okay, 0:25:57.160 --> 0:26:00.560 So Robert, I can recall discussions going back for years 0:26:00.680 --> 0:26:02.280 about whether we're going to make it out of the 0:26:02.359 --> 0:26:05.520 Uncanny Valley in the realm of robotics or animation, and 0:26:05.640 --> 0:26:08.359 I think from here on I want to focus primarily 0:26:08.440 --> 0:26:11.840 on animation, just to just to keep us focused and 0:26:11.880 --> 0:26:14.879 I think there are actually two separate questions here, assuming 0:26:14.920 --> 0:26:18.800 that the Uncanny Valley is to some extent to coherent idea. 0:26:18.880 --> 0:26:21.600 We've already explained all the ways in which it's obviously 0:26:21.600 --> 0:26:25.800 way more complicated than the naive popular culture culture understanding 0:26:25.840 --> 0:26:30.400 of it. But the two big questions. Number one, can 0:26:30.480 --> 0:26:34.520 we make realistic looking humanoid characters that aren't creepy? I 0:26:34.560 --> 0:26:37.000 think the answer here is yes. I think it's not 0:26:37.080 --> 0:26:39.840 a two dimensional graph. I think you can make things 0:26:39.880 --> 0:26:43.840 that aren't quite photo realistic but look realistic that aren't creepy. 0:26:44.080 --> 0:26:47.200 Current generation video games have been doing this, and as 0:26:47.200 --> 0:26:49.040 I mentioned in the last episode, I think that there 0:26:49.040 --> 0:26:52.679 are tricks to doing this. It's apparently in achieving like 0:26:52.720 --> 0:26:57.000 the right combination of realistic traits and unrealistic traits that 0:26:57.160 --> 0:26:59.399 maybe you would just land on by doing trial and 0:26:59.520 --> 0:27:03.080 error in design over time. You would never mistake these 0:27:03.160 --> 0:27:08.160 characters for photographs of real humans. But they're also not cartoony. 0:27:08.200 --> 0:27:12.520 They've got this feeling of really real ishness, if you 0:27:12.560 --> 0:27:15.560 know what I mean. But they've they've they've attained sort 0:27:15.600 --> 0:27:19.960 of a generally acceptable plateau of realistic affect. But they're 0:27:20.000 --> 0:27:24.639 not skewing into these different danger zones adjacent photo realism, 0:27:24.640 --> 0:27:27.600 where the shortcomings become creepy and off putting and we 0:27:27.640 --> 0:27:31.800 don't like it. Then there would be another question, and 0:27:31.880 --> 0:27:35.600 that's just can we make animated characters that are robustly 0:27:35.720 --> 0:27:39.000 indistinguishable from human like, can we get all the way 0:27:39.080 --> 0:27:42.600 up the other side of the mountain, up to the 0:27:42.880 --> 0:27:46.480 peak of reality? And last week, if you'd ask me, 0:27:46.560 --> 0:27:49.399 my personal answer would have been no, not yet. But 0:27:49.480 --> 0:27:51.879 I think that's actually not as clear cut as we 0:27:51.880 --> 0:27:54.800 would first guess, because you think that perhaps many of 0:27:54.840 --> 0:27:58.640 the humans we see on TV are actually digital creations. 0:27:58.840 --> 0:28:03.200 Oh I know, for a act that that who's that guy, 0:28:03.440 --> 0:28:06.320 that Jimmy Fallon guy? Oh yeah, Jimmy Fallon might be 0:28:06.480 --> 0:28:11.199 a computer generated that is not a person. He has 0:28:11.240 --> 0:28:15.320 been generated by a computer that's in Palo Alto, California. 0:28:16.040 --> 0:28:18.400 It's a super computer. I mean, it's a really good computer. 0:28:18.520 --> 0:28:20.760 But yeah, well, I often feel the same way about 0:28:21.080 --> 0:28:24.360 Michael Fastbender, And granted it's complicated by the fact that 0:28:24.400 --> 0:28:27.320 he has he has a thing for playing androids recently, 0:28:27.760 --> 0:28:29.720 but at times you're just like, no, he's just a 0:28:29.720 --> 0:28:32.199 little too handsome. There's something in human about this this 0:28:32.280 --> 0:28:35.480 man's handsomeness in his charm. So I want to talk 0:28:35.520 --> 0:28:38.480 about one thing that is that has given me pause 0:28:38.560 --> 0:28:41.080 on this subject. And it's going back to what we 0:28:41.120 --> 0:28:44.080 talked about at the beginning of the last episode Rogue one. Yes, 0:28:44.720 --> 0:28:47.320 so back to the c G I Grand mof Tarken 0:28:47.720 --> 0:28:49.680 when I saw When I first saw Rogue one, I 0:28:49.720 --> 0:28:51.440 liked a lot of things about the movie, but I 0:28:51.480 --> 0:28:54.000 did not really like the the c g I. Tarken. 0:28:54.560 --> 0:28:58.959 The the almost Peter Cushing was very, very good, and 0:28:59.000 --> 0:29:03.200 I really mean that, I'm shockingly good, but still not 0:29:03.320 --> 0:29:06.440 quite real to me, still kind of distracting because of 0:29:06.440 --> 0:29:10.080 how slightly off it was. I would not have mistaken 0:29:10.080 --> 0:29:13.280 it for a real person. But the other day I 0:29:13.320 --> 0:29:15.920 was talking in the office to Holly fry Who it's 0:29:15.960 --> 0:29:19.959 It's Holly, a Star Wars fan. She's so, she's one 0:29:19.960 --> 0:29:21.959 of the hosts of Stuff You Missed in History Class, 0:29:22.040 --> 0:29:24.960 one of our one of our podcasts in the podcast 0:29:25.000 --> 0:29:27.760 family here. Yeah, she has hands down the most Star 0:29:27.800 --> 0:29:30.360 Wars knowledgeable person in the office in an office full 0:29:30.400 --> 0:29:33.920 of nerds. Yeah, I should point out that Holly has 0:29:34.440 --> 0:29:37.240 slash had like a golden ticket to go see Rogue 0:29:37.240 --> 0:29:40.680 one anytime she wanted to at the theater. What are 0:29:40.720 --> 0:29:43.240 you serious? And serious? This is? Yeah, yeah, legit, where'd 0:29:43.240 --> 0:29:45.320 that come from? I don't know, I don't know. I'm 0:29:45.320 --> 0:29:47.600 not I'm not at that level of a fandom where 0:29:47.600 --> 0:29:50.200 I'm even offered such things. I know. So if if 0:29:50.240 --> 0:29:52.280 you want to have a funny experience, go up to 0:29:52.320 --> 0:29:56.400 Holly and just like, ask some really obscure random question 0:29:56.480 --> 0:29:59.240 about Star Wars where you think it could not be 0:29:59.480 --> 0:30:03.080 possible that there's an actual answer to this, like oh, 0:30:03.120 --> 0:30:06.160 that stormtrooper on the left, where did he? What planet 0:30:06.200 --> 0:30:08.520 was he born on? Holly will have an answer. She'll 0:30:08.560 --> 0:30:10.960 be like, oh, that was actually addressed in dialogue in 0:30:11.000 --> 0:30:14.479 the Greek dub of this episode of Clone Wars. So 0:30:14.600 --> 0:30:17.400 Holly has amazing Star Wars knowledge. She's super fun to 0:30:17.400 --> 0:30:21.040 talk to about the Star Wars universe. But anyway, Holly 0:30:21.080 --> 0:30:23.160 pointed out that while a lot of people like me 0:30:23.280 --> 0:30:25.440 were saying that the c G I. Tarkan was a 0:30:25.480 --> 0:30:29.080 few pixels short of escaping the valley, then again, there 0:30:29.080 --> 0:30:31.960 were plenty of people, including some older people that she knew, 0:30:32.440 --> 0:30:35.400 who couldn't tell that it wasn't a real person they 0:30:35.440 --> 0:30:39.000 literally they couldn't tell. Well, I'm worried about my viewing 0:30:39.120 --> 0:30:41.640 upcoming viewing of the film because I have I've been 0:30:42.240 --> 0:30:46.160 preconditioned to have a certain response to the to the 0:30:46.600 --> 0:30:49.640 Tarkan bot here. Oh, I'm sorry that we've had this 0:30:49.680 --> 0:30:52.080 discussion before you were able to see the movie for yourself. 0:30:52.440 --> 0:30:55.480 It'll be interesting being preconditioned. Will I go into it, 0:30:56.120 --> 0:30:59.520 you know, expecting an abomination and like, but the level 0:30:59.520 --> 0:31:02.120 of detail will overwhelm me and it won't matter. Or 0:31:02.280 --> 0:31:04.280 Am I going to go into there and nothing is 0:31:04.280 --> 0:31:06.600 gonna fool me because I'm gonna be looking for the cracks? 0:31:07.760 --> 0:31:10.720 You know, I can, I can put myself in a 0:31:10.760 --> 0:31:15.960 mindset where I think it's possible. I might not have 0:31:16.120 --> 0:31:18.920 known that it was a c g I effect if 0:31:18.960 --> 0:31:21.160 I wasn't familiar with what to look for, Like if 0:31:21.200 --> 0:31:23.240 I didn't watch a lot of movies that had c 0:31:23.360 --> 0:31:25.960 g I effects in them, and if I didn't complain 0:31:26.040 --> 0:31:28.880 about c G I a lot. I'm sorry, guilty is charged, 0:31:29.080 --> 0:31:32.280 I'm guilty of that. Uh. If I wasn't aware that 0:31:32.320 --> 0:31:35.719 Peter Cushing was dead, um, if it wasn't if I 0:31:35.760 --> 0:31:38.040 wasn't sort of prepared to see a lot of high 0:31:38.040 --> 0:31:40.200 tech c g I by virtue of the fact that 0:31:40.240 --> 0:31:43.240 I'm sitting in a theater for a Lucasfilm movie, all 0:31:43.280 --> 0:31:45.520 those things. If you took away all the context and 0:31:45.600 --> 0:31:49.520 my pre knowledge, I'm very might possibly have fallen for it. 0:31:49.560 --> 0:31:51.800 I think I might have just it might have just 0:31:51.840 --> 0:31:54.080 gone past me. If I was absorbed in the story, 0:31:54.160 --> 0:31:56.560 I might have thought, Yeah, it's kind of strange looking dude, 0:31:56.600 --> 0:31:59.880 but it's just a dude. And I think it gets 0:32:00.040 --> 0:32:03.760 better or worse, depending on your perspective. So I have 0:32:04.120 --> 0:32:09.040 in October news piece from BBC Asia here about a 0:32:09.160 --> 0:32:14.080 character called Saia, a computer animated character created by the 0:32:14.240 --> 0:32:17.600 Japanese husband and wife graphic design team Terry Yuki and 0:32:17.640 --> 0:32:21.200 Yuka Ishikawa. And I mentioned this one in particular because 0:32:21.240 --> 0:32:23.160 before we did this episode, I went and I looked 0:32:23.200 --> 0:32:26.040 up what are considered a lot of the most realistic 0:32:26.080 --> 0:32:30.320 c g I character creations, the most impressive animations. This 0:32:30.360 --> 0:32:32.480 one came up, and I think this was the most 0:32:32.520 --> 0:32:37.160 impressive to me. It's probably the most photo real computer 0:32:37.280 --> 0:32:40.680 generated human I've come across so far. So Siah is 0:32:40.680 --> 0:32:44.360 supposed to be a seventeen year old Japanese student and 0:32:44.400 --> 0:32:46.920 The creators have been working on her design for a 0:32:46.960 --> 0:32:50.120 couple of years now, and as versions of Saia have 0:32:50.200 --> 0:32:52.920 been posted on the Internet, people have widely reacted with 0:32:53.000 --> 0:32:55.880 comments like, I can't believe that's not a real person. 0:32:57.040 --> 0:32:59.640 And I kind of have to agree. I'm looking at 0:32:59.680 --> 0:33:01.880 these these pictures of her. There are a couple of 0:33:01.880 --> 0:33:05.800 different generations of her design up and the most recent 0:33:05.840 --> 0:33:09.920 one just looks like a photograph of a person. Yeah, 0:33:10.040 --> 0:33:13.120 I I can't tell that that is not a person. 0:33:13.280 --> 0:33:18.000 I have no recourse to critical faculties in my mind 0:33:18.080 --> 0:33:20.440 that would say no, here's where you can tell that 0:33:20.440 --> 0:33:23.360 that's not a real person. Now. At the same time, 0:33:23.360 --> 0:33:24.840 I do have to come back to a comment I 0:33:24.880 --> 0:33:27.640 made earlier that that this is also a it's a 0:33:27.760 --> 0:33:30.960 it's a it's a pretty face, it's a very standard 0:33:31.000 --> 0:33:34.600 face like this is this is leading lady material, Whereas 0:33:34.960 --> 0:33:37.000 I think it gets more problematic when you look at 0:33:37.280 --> 0:33:41.120 character actor type figures such Peter Cushing, because they have 0:33:41.320 --> 0:33:44.360 such a distinctive face. Yes, and Peter Cushing. So this 0:33:44.440 --> 0:33:46.560 is one thing that helps, I think, is that it's 0:33:46.640 --> 0:33:51.240 a young character who has very smooth features. Uh, Peter 0:33:51.320 --> 0:33:53.800 Cushing has a lot of cracks and crags, a lot 0:33:53.800 --> 0:33:55.880 of wrinkles, and I think that that may actually be 0:33:56.360 --> 0:33:59.640 simply having more texture on your face could make it 0:33:59.720 --> 0:34:03.000 much more difficult to make a photo real copy of you. 0:34:03.360 --> 0:34:06.960 That that's entirely possible. But to get back to Saia, 0:34:07.080 --> 0:34:10.880 So in October of last year, the Artist's debuted the 0:34:10.920 --> 0:34:14.880 first animated clip of Siah, which they created using motion 0:34:14.920 --> 0:34:18.880 capture technology, and they debuted at a Japanese consumer electronics show. 0:34:19.640 --> 0:34:23.400 I watched this footage and I think Morey's distinction about 0:34:23.440 --> 0:34:27.920 having different standards for motion and still images does apply 0:34:28.080 --> 0:34:31.319 because while with the still image, I can't tell that's 0:34:31.360 --> 0:34:34.719 not a real person, with the short animated clip, I can. 0:34:34.880 --> 0:34:36.680 I can tell it's not a real person, but it's 0:34:36.680 --> 0:34:42.600 still very very impressive, not as absolutely photo real as 0:34:42.640 --> 0:34:46.000 the still images. But I don't know. I wonder to 0:34:46.080 --> 0:34:50.320 what extent this gap is just um that motion animation 0:34:50.400 --> 0:34:53.480 is a bigger technical project, it takes more investment and 0:34:53.560 --> 0:34:56.520 money and all that, um, And to what extent the 0:34:56.520 --> 0:34:59.879 gap is within the viewers mind essentially, to what ext 0:35:00.000 --> 0:35:02.520 and it's caused by the fact that the climb out 0:35:02.520 --> 0:35:06.560 of the uncanny valley is steeper if you're moving. Now. 0:35:06.600 --> 0:35:08.440 I know we're not talking about robots here, but this, 0:35:08.480 --> 0:35:10.879 of course, this brings up the thought that as where 0:35:10.960 --> 0:35:13.680 as we attempt to conquer this in the realm of 0:35:13.880 --> 0:35:18.080 humanoid robotics, you're gonna inevitably have situations where, oh, it 0:35:18.080 --> 0:35:20.200 looks just like a person if it's walking down the street, 0:35:20.640 --> 0:35:23.640 but if it climbs stairs. There you go. It's not 0:35:23.680 --> 0:35:26.960 necessarily going to add to oh nine the lights in 0:35:27.000 --> 0:35:29.600 the air, but maybe there would be something telling like, oh, 0:35:29.680 --> 0:35:31.520 it looks like a human most of the time. But 0:35:31.560 --> 0:35:36.880 they're gonna be certain movements, certain environmental reactions that are 0:35:36.920 --> 0:35:39.560 just not going to hold up to scrutiny, right, So, 0:35:39.960 --> 0:35:42.240 I don't know. Looking at these things, looking at Tarkan 0:35:42.520 --> 0:35:46.000 in Rogue one, looking at Siah, I think it's clear 0:35:46.040 --> 0:35:49.680 that we're getting closer and closer to really bridging the 0:35:49.719 --> 0:35:55.400 gap on indistinguishable photo real humanity in computer animation. Whether 0:35:55.440 --> 0:35:58.040 you'd call that an uncanny valley or not, obviously this 0:35:58.120 --> 0:36:01.239 is a related but maybe different an issue. What we're 0:36:01.280 --> 0:36:05.040 definitely drawing near is that peak of reality where there 0:36:05.040 --> 0:36:09.040 are synthetically generated images of humans that you can't tell 0:36:09.200 --> 0:36:12.960 from the real thing, and since things in the Uncanny 0:36:13.040 --> 0:36:15.239 Valley are creepy, I think we usually just assume that 0:36:15.400 --> 0:36:19.200 overcoming it is a good thing, right, like designs getting better? Uh. 0:36:19.239 --> 0:36:21.160 In't it kind of cool that we can generate these 0:36:21.160 --> 0:36:25.200 photo real images without without actually having to photograph someone. 0:36:26.040 --> 0:36:28.799 But I'm not so sure that's a good thing. I 0:36:28.840 --> 0:36:31.160 think we should maybe think about the implications of this, 0:36:31.320 --> 0:36:34.960 like what would happen in the world where human simulations, 0:36:35.520 --> 0:36:39.960 especially computer animation, can reliably climb up that second peak. 0:36:40.719 --> 0:36:42.480 So we are going to take a quick break and 0:36:42.520 --> 0:36:49.440 when we come back, we will go beyond the Uncanny Valley. 0:36:50.080 --> 0:36:52.120 All right, we're back. You know, I'm glad you brought 0:36:52.160 --> 0:36:54.000 up this idea of you know, is it a bad thing? 0:36:54.040 --> 0:36:56.120 Is it a good thing? It does remind me of 0:36:56.160 --> 0:36:58.080 a fabulous book that came out a few years back. 0:36:58.080 --> 0:36:59.960 I think I just referenced it in our Sex Spots 0:37:00.040 --> 0:37:04.600 episode titled The wind Up Girl by Paolo baglapi Um. 0:37:04.640 --> 0:37:08.680 It's a near future science fiction tale, just really a 0:37:08.680 --> 0:37:12.000 wonderful novel, very fun, but that the wind Up Girl 0:37:12.280 --> 0:37:16.920 in question is a essentially a sex Spot character. It 0:37:16.960 --> 0:37:19.319 goes you know, that ends up rebelling and you have 0:37:19.400 --> 0:37:22.319 sort of a typical narrative with her, but they call 0:37:22.440 --> 0:37:26.439 her a wind up girl because she's she's she's very 0:37:26.440 --> 0:37:29.600 convincing as a humanoid, except that her skin pours are 0:37:29.640 --> 0:37:34.320 too small and she has an intentionally herkey jerky movement 0:37:34.400 --> 0:37:38.000 to uh to as she walks around. That they did 0:37:38.200 --> 0:37:41.080 so that she could not be mistaken as a person. 0:37:41.920 --> 0:37:44.279 So that so that apparently, like all the I think 0:37:44.320 --> 0:37:46.920 they were called the new people, it's in some cases 0:37:47.120 --> 0:37:50.120 so that the new people could not be mistaken for 0:37:50.200 --> 0:37:53.200 the old people. Whoa It reminds me of how they 0:37:53.200 --> 0:37:57.920 had to add artificial sounds to electric cars for safety purposes. 0:37:58.120 --> 0:38:00.840 Because the cars are too quiet, they can really sneak 0:38:00.920 --> 0:38:02.759 up on you from behind, so they had to make 0:38:02.840 --> 0:38:07.600 them rumble a little bit. Yeah, I think that's an 0:38:07.600 --> 0:38:11.640 app comparison. So here's a question for you, Robert, what 0:38:11.840 --> 0:38:16.120 is the gold standard of evidence that somebody did something? 0:38:17.280 --> 0:38:20.440 Imagine you're on a jury. I'm the defendant. I've been 0:38:20.440 --> 0:38:23.560 accused of offering a cash bribe to a police officer 0:38:23.680 --> 0:38:25.880 if she'll let me borrow her gun for five minutes. 0:38:26.600 --> 0:38:29.359 She says, I did it. I plead not guilty. What 0:38:29.360 --> 0:38:32.240 what's the best evidence to convince you that I really 0:38:32.280 --> 0:38:36.640 did that. Well, it's not human memory, because if we've 0:38:36.680 --> 0:38:39.560 touched on many times before, human memory is falliable and 0:38:39.800 --> 0:38:42.360 and it's a legitimate problem when it comes to to 0:38:42.719 --> 0:38:47.120 eyewitness testimony. But when the eyewitness is a video camera, 0:38:47.200 --> 0:38:50.680 a digital camera that has a long photographic evidence as 0:38:50.719 --> 0:38:53.400 well to certain degrees, like this has been held up 0:38:53.400 --> 0:38:56.520 as the gold standard. I mean, assuming the footage is 0:38:56.560 --> 0:39:00.680 clear enough that the individual's face is visible, all of that. Uh. 0:39:00.840 --> 0:39:03.759 Like even in our science fiction, right, we have so 0:39:03.800 --> 0:39:06.960 many examples of like on Star Trek, again, there would 0:39:06.960 --> 0:39:09.920 be scenes where Picard would command it we zoom in 0:39:09.960 --> 0:39:12.920 and enhance and it was never questioned that there were 0:39:12.920 --> 0:39:15.560 any problems with the enhancing of the image. It was 0:39:15.600 --> 0:39:17.239 just something that was done. It's like, oh, yeah, the 0:39:17.280 --> 0:39:19.120 image is enhanced, and now we see exactly who the 0:39:19.160 --> 0:39:21.640 killer is. Huh. So we should look at this this 0:39:21.760 --> 0:39:24.920 booming new research field. I shouldn't say booming, I just 0:39:24.960 --> 0:39:29.680 mean there are some papers on it, okay, called facial reenactment. 0:39:30.360 --> 0:39:32.480 So this employs some of the same techniques that you 0:39:32.520 --> 0:39:36.479 would see used in uh in studios. If if people 0:39:36.520 --> 0:39:39.040 are doing motion capture for c g I characters In 0:39:39.080 --> 0:39:43.000 movies and video games. You have an actor performer who 0:39:43.040 --> 0:39:45.719 puts on special gear and a special environment surrounded by 0:39:45.800 --> 0:39:49.439 lights and cameras, and the performer acts out motions. These 0:39:49.480 --> 0:39:53.359 motions are captured from multiple angles and different lighting conditions, 0:39:53.760 --> 0:39:56.719 and then they're translated by a computer into the motions 0:39:56.760 --> 0:39:58.600 of a c g I character. You could make a 0:39:58.680 --> 0:40:01.239 c g I me that was doing all the same 0:40:01.280 --> 0:40:05.440 things I did with my body. But what if instead 0:40:05.440 --> 0:40:09.000 of a c g I character you used captured motion 0:40:09.080 --> 0:40:13.880 to manipulate existing video or images of a real person, 0:40:14.040 --> 0:40:17.800 not a c g I character. This technology is already 0:40:17.800 --> 0:40:20.520 in development today, and one example is the research being 0:40:20.560 --> 0:40:23.440 done under the heading as I said, of facial reenactment. 0:40:23.440 --> 0:40:26.520 There are a couple of papers along with accompanying video 0:40:26.600 --> 0:40:29.839 demonstrations by a group of researchers based out of Stanford, 0:40:29.960 --> 0:40:33.279 out of the Max Planck Institute for Informatics and the 0:40:33.400 --> 0:40:37.920 University of Erlangen Nuremberg, and in their own words quote, 0:40:38.040 --> 0:40:41.040 we present a method for the real time transfer of 0:40:41.160 --> 0:40:45.279 facial expressions from an actor in a source video to 0:40:45.400 --> 0:40:49.120 an actor in a target video, thus enabling the ad 0:40:49.160 --> 0:40:53.080 hoc control of the facial expressions of the target actor. 0:40:54.239 --> 0:40:55.840 And so if you haven't seen video of this, you 0:40:55.840 --> 0:40:59.600 should look it up trying facial reenactment video. If you 0:40:59.680 --> 0:41:03.359 have efficient sample video of your target, you can use 0:41:03.400 --> 0:41:08.240 a regular camera to project new facial expressions, including mouth 0:41:08.320 --> 0:41:12.520 movements which form the shapes of words, onto your your 0:41:12.560 --> 0:41:15.200 target in your video. So I could take video of 0:41:15.320 --> 0:41:19.160 Robert talking if I had all this technology, if I 0:41:19.160 --> 0:41:21.800 could take video of Robert talking, and then I could 0:41:21.840 --> 0:41:26.040 film myself saying Halloween five is the best entry in 0:41:26.080 --> 0:41:30.640 the Halloween franchise, and then map that onto Robert's face 0:41:30.719 --> 0:41:33.719 to make his lips say those words, to make his 0:41:33.880 --> 0:41:37.279 face move along with my face as it's being recorded. 0:41:37.719 --> 0:41:40.879 And in the demonstrations of this it looks nearly photo real. 0:41:40.960 --> 0:41:44.279 They do it with with public figures, making them move 0:41:44.320 --> 0:41:47.839 their faces around, move their lips to say things. In 0:41:47.880 --> 0:41:50.880 some cases, I think the average observer already would not 0:41:51.040 --> 0:41:54.759 be able to tell the difference in this video and 0:41:54.800 --> 0:41:57.160 in fact that a similar thing appears to be happening 0:41:57.160 --> 0:41:59.839 with voice. This thing you might have read about last 0:41:59.880 --> 0:42:02.200 year of this thing Adobe Voco, where they came out 0:42:02.239 --> 0:42:05.040 with this announcement that Adobe is working on software where 0:42:05.120 --> 0:42:07.680 you can take a twenty minute sample of your audio 0:42:07.800 --> 0:42:10.040 to learn from. If you've got a recording of somebody 0:42:10.120 --> 0:42:12.840 talking for twenty minutes, I can take it, make a 0:42:12.880 --> 0:42:15.879 recording of you saying things you never said in your 0:42:15.880 --> 0:42:19.800 own voice. I wanna. I seem to remember that before 0:42:19.800 --> 0:42:22.920 his death, Roger Ebert was involved with the project with 0:42:22.960 --> 0:42:25.360 some of this technology. The idea of being that, of 0:42:25.360 --> 0:42:29.279 course he had lost his his availability to talk due 0:42:29.320 --> 0:42:34.400 to illness, but there was so much Roger Ebert audio 0:42:34.719 --> 0:42:36.560 out there from all of his years as a as 0:42:36.600 --> 0:42:39.400 a as a film critic and a TV personality that 0:42:39.440 --> 0:42:41.920 they had this they had, they had everything they needed 0:42:42.280 --> 0:42:45.799 to enable him to say anything new. Yeah, and that 0:42:45.800 --> 0:42:49.480 that explores the totally non nefarious aspect of this. I mean, 0:42:49.520 --> 0:42:51.640 I I don't think people who are pursuing these lines 0:42:51.680 --> 0:42:54.279 of research are just trying to create a world where 0:42:54.320 --> 0:42:57.680 we can fake video evidence of things. But wouldn't that 0:42:57.680 --> 0:42:59.960 be a wonderful thing for somebody who lost their capacity 0:43:00.040 --> 0:43:02.759 for speech they had recordings of their voice to be 0:43:02.800 --> 0:43:05.840 able to create a text to speech voice box that 0:43:05.880 --> 0:43:10.600 could speak with their own voice. That's amazing, that's kind 0:43:10.600 --> 0:43:15.240 of beautiful. But there are these other ways of looking 0:43:15.239 --> 0:43:17.879 at this, and the authors also that they point out that, 0:43:18.080 --> 0:43:20.719 you know, they they in their own defense, they're like, look, 0:43:20.760 --> 0:43:22.680 we're not trying to create a world where people can 0:43:22.719 --> 0:43:25.319 fake video. We also try to show how you can 0:43:25.360 --> 0:43:28.840 detect altered video. So that's another thing they're trying to 0:43:28.880 --> 0:43:31.440 explore and make public because you know, they're not the 0:43:31.480 --> 0:43:34.520 only people pursuing this research. Obviously, people all over the 0:43:34.520 --> 0:43:38.200 place are doing stuff like this, and stuff like this 0:43:38.239 --> 0:43:39.960 has been in the use, has been in use in 0:43:40.000 --> 0:43:42.879 the movie industry for years. Yeah, I mean it makes 0:43:42.880 --> 0:43:44.440 me think that what you would need to go for 0:43:44.680 --> 0:43:47.319 is the equivalent of a water mark. I don't know 0:43:47.360 --> 0:43:50.200 exactly what that watermark would be and what form it 0:43:50.239 --> 0:43:53.279 would take, but it does make me think, well, we're 0:43:53.320 --> 0:43:55.440 going to reach a point where any kind of footage 0:43:55.440 --> 0:43:59.680 has to have the watermark of authenticity, otherwise doubt will 0:43:59.719 --> 0:44:04.520 be asked upon it. Yeah, I'm concerned about the idea 0:44:04.560 --> 0:44:07.600 of living in a world where you can make very 0:44:07.680 --> 0:44:12.960 convincing looking fake video evidence of things, and not just 0:44:13.080 --> 0:44:16.160 because of the specific example of somebody can make a 0:44:16.280 --> 0:44:18.840 video of me or somebody I like, you know, saying 0:44:18.920 --> 0:44:21.480 or doing something that they didn't do. It's not just 0:44:21.560 --> 0:44:25.680 the specifics, it's the general degrading of our trust in 0:44:25.719 --> 0:44:29.960 the ability to look at things and know that they're true. Yeah. 0:44:30.000 --> 0:44:32.680 I mean, we look at the current news cycle and 0:44:32.680 --> 0:44:36.120 there's been a lot of discussion about the reliability of information, 0:44:36.239 --> 0:44:40.799 of so called fake news, the idea that you read, 0:44:40.920 --> 0:44:43.520 do you reach this point where nobody knows what to 0:44:43.560 --> 0:44:46.480 trust anymore? Set of just trusting nothing? And you already 0:44:46.520 --> 0:44:49.440 know this because you don't trust any weird looking picture 0:44:49.560 --> 0:44:51.759 of somebody you see, right, because you know what can 0:44:51.800 --> 0:44:54.920 be done with photoshop. We're already there with still images. 0:44:55.080 --> 0:44:59.600 Photoshop is is basically in many contexts become a verb 0:45:01.040 --> 0:45:04.359 for for the the distortion of truth. But what what 0:45:04.440 --> 0:45:08.880 if we had photoshop to undermine moving video evidence to 0:45:08.920 --> 0:45:13.359 the same extent that photoshop has undermined still images. I mean, 0:45:13.400 --> 0:45:15.560 this is what a lot of these viral fake news 0:45:15.560 --> 0:45:17.920 stories are based on, is a photoshopped image. If you 0:45:17.960 --> 0:45:20.520 go to Snopes or something and you look at what 0:45:20.600 --> 0:45:22.759 a lot a lot of what they're debunking is, it's 0:45:22.840 --> 0:45:25.640 just an image that claims to be real of somebody 0:45:25.680 --> 0:45:28.480 doing something, and they have to track down where it 0:45:28.520 --> 0:45:33.040 came from Yeah, it's difficult to imagine what that what 0:45:33.080 --> 0:45:35.840 would slash will be like when we reach that point 0:45:35.880 --> 0:45:39.640 to where there's where where video even digital footage is 0:45:39.680 --> 0:45:41.520 no longer the gold standard it was. I mean, I 0:45:41.600 --> 0:45:46.880 think it's actually a very important project to maintain a 0:45:47.040 --> 0:45:51.920 version of the Uncanny Valley, to to help people find 0:45:51.960 --> 0:45:55.839 a way to all to separate real video evidence from 0:45:55.840 --> 0:45:58.920 fake video evidence of things, to understand that there are 0:45:58.920 --> 0:46:01.800 things you can look for that separate real moving imagery 0:46:01.920 --> 0:46:05.400 from falsified or synthetic moving imagery. Now, one way you 0:46:05.400 --> 0:46:08.960 can approach this is to do what the authors of 0:46:08.960 --> 0:46:11.279 this research I was talking about do is they say, look, 0:46:11.480 --> 0:46:15.160 here are things that are are signs that video has 0:46:15.200 --> 0:46:18.719 been manipulated. And that's one thing. And maybe there will 0:46:18.760 --> 0:46:20.720 be a lot of experts on this in the future, 0:46:20.800 --> 0:46:23.040 Like it could be a whole field of people who 0:46:23.080 --> 0:46:27.919 are just there to have expertise in authenticating purportedly real 0:46:28.040 --> 0:46:31.960 video of you doing things or not. Uh. Then on 0:46:32.000 --> 0:46:35.719 the other hand, we could hope that there is in 0:46:35.880 --> 0:46:40.400 fact an adaptive response in our discernment in general. And 0:46:40.400 --> 0:46:42.800 this is where I want to go to the concept 0:46:42.840 --> 0:46:47.279 of an of the uncanny Wall. So paper in the 0:46:47.320 --> 0:46:51.640 International Journal of Arts and Technology, authored by Tinwell Grimshawn 0:46:51.719 --> 0:46:56.280 Williams offers this interesting counter hypothesis to the Uncanny Valley. 0:46:56.320 --> 0:46:59.240 And I want to emphasize again we've sort of shifted 0:46:59.239 --> 0:47:01.920 back and forth betwe in the Uncanny Valley itself in 0:47:02.000 --> 0:47:05.960 terms of what causes negative affinity, and then over slightly 0:47:06.160 --> 0:47:08.960 just to the issue of nearing photo realism or not. 0:47:09.120 --> 0:47:12.239 So keep in mind the difference in those subjects. But 0:47:13.920 --> 0:47:17.680 they proposed this idea of the uncanny wall. To put 0:47:17.680 --> 0:47:23.239 it succinctly, they proposed that quote, increasing technological sophistication in 0:47:23.320 --> 0:47:27.880 the creation of realism for humanlike virtual characters is matched 0:47:28.480 --> 0:47:33.040 by increasing technological discernment on the part of the viewer. 0:47:33.800 --> 0:47:38.320 In other words, as humanoid characters become more real, our 0:47:38.400 --> 0:47:42.799 standards for what looks realistic go up. And I do 0:47:42.880 --> 0:47:46.799 think they're just anecdotally personally that I think there's some 0:47:46.880 --> 0:47:48.920 support for this, and I kind of hope this is 0:47:48.920 --> 0:47:52.520 true so we can avoid this world where all video 0:47:52.560 --> 0:47:55.960 evidence is in question. Because I immediately think, right now, 0:47:56.000 --> 0:48:00.480 I've got I've got antennae for photoshopped images, unlike had 0:48:00.520 --> 0:48:04.040 ten years ago. I think stuff that would look obviously 0:48:04.160 --> 0:48:07.479 photoshop to me today would have fooled me ten years ago. 0:48:07.760 --> 0:48:11.480 I think I've simply adapted. And another thing is it 0:48:11.520 --> 0:48:13.560 makes me flashback to the early days of c g 0:48:13.680 --> 0:48:17.040 I in movies and like the thirty two bit video 0:48:17.160 --> 0:48:21.759 game era, or think about like PlayStation one games and 0:48:21.880 --> 0:48:24.800 back then. I remember looking at games for the original 0:48:24.840 --> 0:48:29.920 PlayStation and thinking, wow, that looks so real. And you 0:48:29.960 --> 0:48:31.960 try and play him now and it's painful. Yeah, you 0:48:32.000 --> 0:48:36.799 can't blocky polygons. People's faces have all these sharp corners. 0:48:36.960 --> 0:48:40.880 It's it's hilarious. Uh, there was some kind of geometrical 0:48:41.120 --> 0:48:44.360 nightmare world everything was taking place in where there's just 0:48:45.080 --> 0:48:48.240 lots of sharp angles. But at the time it looked 0:48:48.360 --> 0:48:50.839 so real to me. And another fun trick is go 0:48:50.920 --> 0:48:53.839 back and read movie reviews for movies with bad c 0:48:53.960 --> 0:48:58.359 g I from the nineties. Professional movie reviewers at the time, 0:48:58.400 --> 0:49:02.160 we're often praising the X. One example is like the 0:49:02.239 --> 0:49:05.960 Mortal Kombat movie, the original Mortal Kombat movie. Remember this. 0:49:06.280 --> 0:49:08.799 You can find reviews at the time where people are like, well, 0:49:08.840 --> 0:49:13.279 the stories then an immature but dazzling special effects. Now 0:49:13.360 --> 0:49:17.799 even mentioning those special effects conjures a kind of delirious hilarity. 0:49:17.880 --> 0:49:20.160 You just start laughing when you think about the c 0:49:20.280 --> 0:49:23.480 G I in Mortal Kombat. But that being said, the 0:49:23.480 --> 0:49:27.640 GOA puppet was above reproach, really good. It did have 0:49:27.719 --> 0:49:31.799 kind of nasty beady eyes, kind of crypt keeper ask. Yeah, 0:49:31.960 --> 0:49:34.400 oh yeah, it was like a very It was like 0:49:34.440 --> 0:49:40.120 a bloated, buffed up crypt kicker. But yeah, it's so 0:49:40.280 --> 0:49:43.840 ugly it provokes uncontrollable laughter. But at the time people 0:49:43.880 --> 0:49:47.680 were like, dazzling, looks amazing. Uh. So it makes me 0:49:47.719 --> 0:49:51.320 think that I hope that there is something to this, 0:49:51.320 --> 0:49:55.360 this idea that these authors have that as things continue 0:49:55.400 --> 0:49:59.120 to chase photo realism, as synthetic imagery of humans gets 0:49:59.160 --> 0:50:01.799 closer and closer to the real thing, we just get 0:50:01.880 --> 0:50:06.920 more and more attuned to the minute problems with them 0:50:06.960 --> 0:50:11.279 with them and never really get fully fooled. Well. I 0:50:11.320 --> 0:50:14.640 have two thoughts here, One on the whole water mark thing. 0:50:14.640 --> 0:50:17.080 Maybe it would be some sort of a Bitcoin type 0:50:17.120 --> 0:50:21.120 of authentication system that would be in place. Uh. The 0:50:21.560 --> 0:50:23.720 other is maybe you have to go beyond the real 0:50:24.080 --> 0:50:26.239 maybe safe for a head of state to appear in 0:50:26.239 --> 0:50:29.600 a video and it'd be authentic. They have to appear 0:50:29.800 --> 0:50:33.880 as an as a computer generated avatar so advanced than 0:50:33.960 --> 0:50:36.920 it is that it is beyond the ability of any 0:50:37.120 --> 0:50:41.360 like non state production or company to create like something 0:50:41.800 --> 0:50:45.359 I'm I'm something that at this point in what would 0:50:45.360 --> 0:50:47.719 be their past, I cannot even conceive of, like a 0:50:48.000 --> 0:50:51.160 I don't know, like a five dimensional unfolding c g 0:50:51.440 --> 0:50:55.200 I god being, because how he's going to fake that? 0:50:55.239 --> 0:50:57.480 You can fake a person, but good luck faking the 0:50:57.760 --> 0:51:01.680 fifth dimension hole of avatar or a fish new Okay, 0:51:01.680 --> 0:51:05.640 like like the algorithms for faking a person with a 0:51:05.640 --> 0:51:11.040 lot of photo and and audio visual cues to sample from. 0:51:11.080 --> 0:51:12.600 If you've got a lot of footage out there, you 0:51:12.600 --> 0:51:16.280 could simulate that person, But you couldn't simulate this brand 0:51:16.280 --> 0:51:20.680 new creation that is it requires you know, supercomputers to generate. 0:51:20.800 --> 0:51:22.600 And yeah, and it was probably built from the bottom 0:51:22.719 --> 0:51:26.439 up with completely alien physiology and movements. Just a thought. 0:51:26.520 --> 0:51:28.799 That's kind of a crazy idea, but I like it. Well, 0:51:28.840 --> 0:51:31.480 that's what I'm here for with the crazy ideas. I 0:51:31.520 --> 0:51:33.960 don't know, have you got anything else, Robert, any anything 0:51:34.000 --> 0:51:35.799 else you can think of to save us from the 0:51:35.800 --> 0:51:39.719 future of uh synthetic human imagery? Oh? You know, I 0:51:39.719 --> 0:51:41.480 could sit around here all day and talk about c 0:51:41.600 --> 0:51:45.240 g I, monsters and UH and Uncanny Valley and films 0:51:45.239 --> 0:51:47.520 and video games and whatnot. But you don't have to 0:51:47.760 --> 0:51:49.640 what to save that for another time. Maybe save some 0:51:49.719 --> 0:51:52.960 of it for trailer talk, which will If you're listening 0:51:52.960 --> 0:51:54.799 to this on a Thursday, hopefully you can tune in 0:51:54.800 --> 0:51:58.680 tomorrow around eleven am on our Facebook page. I like 0:51:58.719 --> 0:52:00.719 our Facebook page. While you're at it, follow us there, 0:52:00.760 --> 0:52:04.000 but tune into a little discussion of trailers that are 0:52:04.040 --> 0:52:07.760 associated with the Uncanny Valley. Oh yeah, And in the meantime, 0:52:07.960 --> 0:52:09.719 heading over to stuff to Blow your Mind dot com. 0:52:09.719 --> 0:52:12.560 That's the mothership. That's we will find all of our 0:52:12.640 --> 0:52:15.960 blog posts, our podcast, our videos, and links out to 0:52:16.000 --> 0:52:18.600 our various social media accounts such that Facebook page or 0:52:18.719 --> 0:52:21.799 the Twitter page, the Instagram account, you name it. And 0:52:21.880 --> 0:52:23.680 if you want to get in touch with us directly, 0:52:23.760 --> 0:52:26.760 as always, you can email us at blow the Mind 0:52:27.000 --> 0:52:38.880 at how stuff works dot com for more on this 0:52:39.080 --> 0:52:41.600 and thousands of other topics. Is that how stuff Works 0:52:41.600 --> 0:53:02.200 dot com has atance by the bar