WEBVTT - Beyond the Uncanny Valley 0:00:03.080 --> 0:00:06.160 Welcome to Stuff to Blow Your Mind from how stoffworks 0:00:06.200 --> 0:00:14.640 dot com. Hey, welcome to Stuff to Blow your Mind. 0:00:14.680 --> 0:00:17.000 My name is Robert Lamb and I'm Joe McCormick in. 0:00:17.040 --> 0:00:19.159 Today is going to be the second part of a 0:00:19.200 --> 0:00:22.520 two part series we're doing on the Uncanny Valley. Last time, 0:00:22.760 --> 0:00:25.560 we ventured into the Uncanny Valley, so if you haven't 0:00:25.560 --> 0:00:28.200 heard that episode, you should go back and listen to that. First, 0:00:28.240 --> 0:00:29.920 we lay a lot of the groundwork for what we're 0:00:29.920 --> 0:00:33.000 gonna be talking about today, But we explored the origination 0:00:33.159 --> 0:00:36.159 of the concept of the Uncanny Valley, what it means 0:00:36.360 --> 0:00:39.800 to be in the Uncanny Valley, and some research on 0:00:39.920 --> 0:00:44.120 whether this valley actually exists or not. Today, I think 0:00:44.120 --> 0:00:46.400 we want to start off by looking at if if 0:00:46.440 --> 0:00:50.520 it does exist, what might be some explanations for it. Indeed, Yeah, 0:00:50.560 --> 0:00:53.000 we're gonna we're gonna dive into it a bit more 0:00:53.000 --> 0:00:57.280 and move, as the title, uh suggests, beyond the Uncanny Valley. 0:00:58.080 --> 0:01:00.120 But before we do that, I do want to talk 0:01:00.160 --> 0:01:03.280 about RoboCop. Of course you want to talk about RoboCop, 0:01:03.440 --> 0:01:05.880 because we talked about RoboCop pretty much every day. Yeah, 0:01:05.880 --> 0:01:08.920 it's it's it's an important film important films, I'll say 0:01:09.080 --> 0:01:11.160 at least at least the first to arguably the third 0:01:11.160 --> 0:01:14.480 one to throwing the TV series if you like. But wait, 0:01:14.520 --> 0:01:18.280 there was a TV series RoboCop. It was one of 0:01:18.319 --> 0:01:20.319 those that would I think it would come on Sci 0:01:20.360 --> 0:01:25.240 Fi or it came on various cable channels. I only 0:01:25.240 --> 0:01:28.080 have a vague awareness of it because it seemed to 0:01:28.120 --> 0:01:33.600 be a far lower key RoboCop type show. So in 0:01:33.640 --> 0:01:35.120 many of the studies that we talked about in the 0:01:35.240 --> 0:01:40.080 last episode, they were looking at largely three categories of 0:01:40.080 --> 0:01:43.520 of robots and humans and androids. So you had pure 0:01:43.600 --> 0:01:48.200 robots things that are just undeniably machines and we're mostly 0:01:48.240 --> 0:01:52.400 okay with. Then you have humans or or or perfect 0:01:52.440 --> 0:01:55.440 human replications. Okay, so you look at it, it either 0:01:55.600 --> 0:01:58.000 is a human or it's such a good representation of 0:01:58.040 --> 0:02:01.640 a human ideally that you would not think that it 0:02:01.720 --> 0:02:04.240 was a robot. Right. The third category here is where 0:02:04.280 --> 0:02:06.840 you're gonna get into the danger zone, right, the human 0:02:07.000 --> 0:02:09.919 like robots where you look at it and you say, 0:02:09.960 --> 0:02:12.920 I see what you're going for there, but it's creeping 0:02:12.960 --> 0:02:16.640 me out. So I think it's interesting to aligne this 0:02:16.760 --> 0:02:20.760 up with the Holy Trinity of Robocops. Okay, so this 0:02:20.760 --> 0:02:23.200 will mostly make sense if you've seen the RoboCop films, 0:02:23.240 --> 0:02:24.880 but I feel like most people know what we're talking 0:02:24.919 --> 0:02:27.480 about here. First of all, you had the the proto 0:02:27.560 --> 0:02:30.880 RoboCop add two oh nine. Oh. This is maybe the 0:02:30.919 --> 0:02:33.600 greatest scene in the first film is before we get 0:02:33.639 --> 0:02:38.000 a humanoid RoboCop, they just have this big drone object 0:02:38.120 --> 0:02:40.560 that is supposed to enforce the law and it ends 0:02:40.600 --> 0:02:43.720 up shooting somebody in the boardroom. Yeah. It's a walking 0:02:43.840 --> 0:02:49.240 law enforcement tank with a robot commanding robot voice so 0:02:49.280 --> 0:02:50.840 that you look at it and there's no to not 0:02:51.600 --> 0:02:54.600 this is not a friendly device. But it's not really 0:02:54.639 --> 0:02:56.720 humanoid at all, not really at all. It hit just 0:02:56.800 --> 0:03:00.600 walks on big, big legs that can't even navigate and stairs. 0:03:01.720 --> 0:03:03.799 But I would say I have great affinity for ed 0:03:03.880 --> 0:03:06.600 to oh nine. Yeah, I mean, you know, without having 0:03:06.600 --> 0:03:09.240 to worry about it actually shooting me. Uh, it's kind 0:03:09.240 --> 0:03:13.160 of cute in a way. Now then we have RoboCop 0:03:13.880 --> 0:03:19.680 itself himself, the classic RoboCop, thee Peter Weller, and uh 0:03:19.720 --> 0:03:23.560 he is a cyborg or perhaps an android, depending on 0:03:23.600 --> 0:03:27.360 how you want to view the descriptions. So he's he 0:03:27.480 --> 0:03:32.080 has a relatable, living human face which is a fixed 0:03:32.160 --> 0:03:36.200 to to honor him in some some explanations, or perhaps 0:03:36.200 --> 0:03:38.760 to make him more comfortable not only as a police 0:03:38.840 --> 0:03:42.840 killing machine but also a community law enforcement officer. So RoboCop, 0:03:42.960 --> 0:03:45.960 you know, moves around with very robotic movements, speaks in 0:03:45.960 --> 0:03:49.520 a very robotic voice, but his face is a living 0:03:49.600 --> 0:03:53.560 human face. So in a way like that, that seems 0:03:53.560 --> 0:03:55.120 like it might it just sort of you know, to 0:03:55.240 --> 0:03:58.320 read perhaps more into the original film than was intended. 0:03:58.560 --> 0:04:01.560 Perhaps this was a way to to get beyond the 0:04:01.680 --> 0:04:04.360 Uncanny Valley. We can't replicate the human face. We'll just 0:04:04.400 --> 0:04:07.040 get an actual human face from the dead cop and 0:04:07.080 --> 0:04:09.600 just plaster it up there. Oh but I'd say RoboCop 0:04:09.640 --> 0:04:11.720 with his mask on really does kind of get into 0:04:11.760 --> 0:04:14.760 the Uncanny Valley, and and Weller does some good work 0:04:15.000 --> 0:04:17.160 forcing us in there. Like I think that they're the 0:04:17.320 --> 0:04:21.599 sort of going for it. Yeah, Okay, Well that brings 0:04:21.680 --> 0:04:25.800 us to the next generation RoboCop two, which is not 0:04:25.920 --> 0:04:28.440 just the name of the second RoboCop movie, but also 0:04:28.920 --> 0:04:33.279 the model of RoboCop that replaced the original RoboCop. They think, Hey, 0:04:33.320 --> 0:04:35.400 what would happen if we put tom noon in in there, 0:04:35.640 --> 0:04:39.760 that's right. So they have another essentially a walking tank 0:04:39.960 --> 0:04:42.320 kind of like Ed Too O nine, except it's powered 0:04:42.320 --> 0:04:45.200 by the brain of a psychotic drug lord named Kane 0:04:45.680 --> 0:04:49.640 played by Tom Noonan. He's fabulous in it. Uh. But 0:04:49.720 --> 0:04:51.880 here's the here's the thing. It's walking around, it's killing 0:04:51.920 --> 0:04:54.200 everything with a gatling gun. But then it can pop 0:04:54.400 --> 0:04:56.680 a flat screen TV out of it out of the 0:04:56.880 --> 0:05:00.240 front of its body, and on that screen is a 0:05:00.320 --> 0:05:04.240 twisted uncanny lawnmower man esque c g I face of 0:05:04.360 --> 0:05:08.680 Tom Noonan. Yeah. So so that one really leans into 0:05:08.720 --> 0:05:12.200 the Uncanny Valley. Well yeah, and this this does point 0:05:12.200 --> 0:05:14.640 out another thing, which is that there have been plenty 0:05:14.839 --> 0:05:19.919 of intentional realizations of the Uncanny Valley in film. When 0:05:20.000 --> 0:05:24.240 when people are trying to create an unsettling, unpleasant humanoid 0:05:24.320 --> 0:05:26.919 for story purposes, if it's supposed to be a villain, 0:05:26.960 --> 0:05:29.520 if it's supposed to make people uncomfortable, because that's its 0:05:29.640 --> 0:05:31.680 role in the plot. So one thing I kind of 0:05:31.720 --> 0:05:33.800 wish we'd done. I hadn't even thought about this is 0:05:33.839 --> 0:05:37.520 too if we could talk to somebody who has intentionally 0:05:37.680 --> 0:05:40.800 made things in the Uncanny Valley, what did they do 0:05:40.960 --> 0:05:44.320 on purpose to get it there. If your your job 0:05:44.839 --> 0:05:48.039 is to make a humanoid robot or an animated humanoid 0:05:48.080 --> 0:05:52.360 figure that intentionally pushes all the bad buttons and climbs 0:05:52.400 --> 0:05:54.560 as far down into the valley as it can, what 0:05:54.600 --> 0:05:57.600 do you do that would that would provide some really 0:05:57.600 --> 0:06:01.440 interesting insight into what it actually takes to get there. Well, 0:06:01.480 --> 0:06:03.240 you know, to come back to video games, something that 0:06:03.279 --> 0:06:05.839 comes up a lot. You see these videos going viral 0:06:05.920 --> 0:06:08.640 where it's like just a cut scene or a clip 0:06:08.680 --> 0:06:11.560 from the video game with a humanoid character in a 0:06:11.600 --> 0:06:15.120 more or less human situation, except something screwed up in 0:06:15.160 --> 0:06:17.680 the face is missing, so it's just two floating eyeballs 0:06:17.680 --> 0:06:21.560 and maybe a floating set of digital t So the 0:06:21.560 --> 0:06:24.160 context is key there, Like this thing is acting as 0:06:24.200 --> 0:06:26.160 if it had a face, and it's in an environment 0:06:26.200 --> 0:06:29.359 where we're unsupposed to just roll with it, but clearly 0:06:29.400 --> 0:06:32.600 something's wrong. Yes, okay, Well, as we said in the 0:06:32.680 --> 0:06:34.799 last episode, we talked about the origin of the idea, 0:06:34.880 --> 0:06:37.320 We talked about some evidence for and against the fact 0:06:37.360 --> 0:06:41.640 that the Uncanny Valley actually exists, to to the point 0:06:41.720 --> 0:06:44.440 that it actually does exist to some extent, maybe not 0:06:44.520 --> 0:06:47.520 in the naive original sense everybody would say, where it's 0:06:47.680 --> 0:06:52.719 just related to the amount of realistic human nous in 0:06:52.760 --> 0:06:56.480 a figure, but has other dimensions as well. What causes 0:06:56.680 --> 0:07:00.799 our Uncanny Valley reaction? Obviously people do have this reaction 0:07:00.880 --> 0:07:04.800 they see a humanoid robot or a humanoid animated character, 0:07:04.920 --> 0:07:08.600 the Scorpion King, the final Fantasy, the spirits within characters, 0:07:08.640 --> 0:07:13.160 whatever it is, and we get so, what causes it? 0:07:13.360 --> 0:07:16.960 Why do our brains react that way? Is it biological? 0:07:17.160 --> 0:07:20.760 Is it pure instinct? Is it a learned psychological reaction? 0:07:20.920 --> 0:07:23.800 Is it part of our culture? Is something coming from 0:07:23.880 --> 0:07:26.960 cognitive dissonance of some sort? Now, to go back to 0:07:27.040 --> 0:07:29.720 the origins of the idea with massive hero Mari in 0:07:29.800 --> 0:07:33.480 nineteen seventy and his original paper on the Uncanny Valley, 0:07:33.640 --> 0:07:37.280 Morey speculates that the Uncanny Valley might be a side 0:07:37.280 --> 0:07:40.760 effect of the self preservation instinct. In other words, it's 0:07:40.760 --> 0:07:46.080 a biological adaptation that helps us avoid disease and death. Uh. 0:07:46.120 --> 0:07:48.520 And he starts with the observation that when a normal, 0:07:48.640 --> 0:07:53.200 healthy person becomes sick and eventually dies, they basically tend 0:07:53.240 --> 0:07:57.200 to slide down from the second peak of the Uncanny Valley. 0:07:57.240 --> 0:08:01.000 What we were calling the realism peak, the reality peak, 0:08:01.280 --> 0:08:05.320 and slide down into the uncanny valley. It's like they 0:08:05.360 --> 0:08:08.680 become more like these upsetting puppets and robots. You know, 0:08:08.760 --> 0:08:12.680 they might uh suffer some kind of change in their 0:08:13.320 --> 0:08:16.280 in their the appearance of their skin, of their eyes, 0:08:16.360 --> 0:08:20.760 of their facial expressions. Things begin to look off about them. 0:08:20.840 --> 0:08:24.600 And he writes, quote, the sense of eeriness is probably 0:08:24.640 --> 0:08:28.280 a form of instinct that protects us from proximal rather 0:08:28.360 --> 0:08:33.640 than distal sources of danger. Proximal sources of danger include corpses, 0:08:33.800 --> 0:08:38.160 members of different species, and other entities we can closely approach. 0:08:38.440 --> 0:08:42.720 Distal sources of danger include windstorms and floods. You know, 0:08:43.240 --> 0:08:47.280 this is interesting the dimension of other species. Um, I 0:08:47.320 --> 0:08:50.840 don't know how many films and documentaries I've seen of say, 0:08:50.840 --> 0:08:54.640 lions running around in the natural habitat, and it's almost 0:08:54.679 --> 0:08:57.520 never creepy, No, not not at all, and yet and 0:08:57.559 --> 0:08:59.280 then mostly a lot of times when I'm in a zoo, 0:08:59.280 --> 0:09:01.920 it's not creepy. But there have been times I take 0:09:01.960 --> 0:09:03.600 my son to the zoo a lot here in Atlanta, 0:09:03.760 --> 0:09:06.520 and there are times when we go down to the 0:09:06.800 --> 0:09:08.959 lion enclosure and we're the only people there at the 0:09:09.040 --> 0:09:12.319 zoo because we've arrived super early and we just we're 0:09:12.400 --> 0:09:14.600 hanging out with a lion on the other side of 0:09:14.600 --> 0:09:18.120 the glass. You know, we're perfectly safe, but a deep 0:09:18.320 --> 0:09:22.760 uneasiness comes over me, washes over me, and uh, and 0:09:23.120 --> 0:09:27.400 I'm just confronted by the the the danger of the situation, 0:09:27.480 --> 0:09:31.240 Like there's a danger that that goes beyond any reason 0:09:31.679 --> 0:09:34.760 because I am in close proximity to a dangerous member 0:09:34.880 --> 0:09:39.679 of another species of carnivorous predator, tory animal that would, 0:09:39.800 --> 0:09:42.920 in under normal conditions, potentially kill me. Now there is 0:09:43.080 --> 0:09:47.320 I would bet a very different kind of sensation going 0:09:47.400 --> 0:09:51.200 on inside you when you're in proximity to a lion 0:09:51.679 --> 0:09:55.040 than when you see a creepy looking humanoid robot or 0:09:55.160 --> 0:10:00.160 a creepy looking animation human animation. Right, Like it's it's 0:10:00.160 --> 0:10:02.880 probably exciting. The difference between the when when we talked 0:10:02.880 --> 0:10:06.040 about the creepiness episode the creed, the difference between the 0:10:06.040 --> 0:10:10.640 sort of like uncomfortable threat ambiguity versus sensing that there 0:10:10.640 --> 0:10:13.680 truly is a threat of some kind, right, And you know, 0:10:13.720 --> 0:10:16.760 we spoke to illness here, you know, and like certainly 0:10:16.760 --> 0:10:18.600 someone we talked about like what happens when a co 0:10:18.640 --> 0:10:22.160 worker walks up to you and they're they're sniveling or 0:10:22.320 --> 0:10:26.800 more a little bit pale. Uh, and they say, let's French. Well, 0:10:26.800 --> 0:10:29.600 well hopefully not, but yeah, there's a there's an at 0:10:29.679 --> 0:10:33.760 least an initial sense of, oh, what's wrong with this person? Um, 0:10:33.800 --> 0:10:35.679 I wonder if what they have as contagious, should they 0:10:35.720 --> 0:10:38.080 even be at work? I hope they cover their mouth 0:10:38.080 --> 0:10:40.480 when they sneeze, because we are going to be concerned 0:10:40.480 --> 0:10:42.720 on some level with catching whatever they have. If it's 0:10:43.040 --> 0:10:46.360 it's it's transmittable, right and it I mean there are 0:10:46.400 --> 0:10:49.480 some ways in which we know that seeing illness and 0:10:49.520 --> 0:10:54.120 other people in sights a reaction in the disgust response, right, yeah, 0:10:54.120 --> 0:10:56.600 And there's been actually a lot of research on the 0:10:56.600 --> 0:11:00.360 the disgust response. Darwin wrote about it. Discuss Us from 0:11:00.360 --> 0:11:04.880 an evolutionary standpoint, is all about disease avoidance, you know, 0:11:05.080 --> 0:11:09.440 like do not eat this, you know, stay away from Yeah, 0:11:09.480 --> 0:11:12.200 something's not right here. You might hurt yourself or get sick. 0:11:12.800 --> 0:11:17.120 So we're talking about sensuous discuss So discuss it's tied 0:11:17.160 --> 0:11:20.560 to the senses to sense information that we're absorbing. Uh. 0:11:20.600 --> 0:11:23.440 This is deeply seated in the insula the area of 0:11:23.440 --> 0:11:28.320 the brain that that malfunctions in patients with obsessive compulsive disorder, 0:11:28.800 --> 0:11:31.240 causing them to say, you know, wash and clean things 0:11:31.400 --> 0:11:36.280 endlessly or vacuum unrelenting lee. So the malfunction of that 0:11:36.400 --> 0:11:38.800 area gives us, you know, gives us a clue into 0:11:38.840 --> 0:11:43.559 its functionality. Uh. One interesting fact about disgusting smells, however, 0:11:44.000 --> 0:11:47.760 is that there's a drop off point for bad smell recognition, 0:11:47.960 --> 0:11:51.160 but not for good smells. So I think we've all 0:11:51.400 --> 0:11:54.520 encountered this where it's that like, say, say you're in 0:11:54.559 --> 0:11:57.040 your your office and and you share your office with 0:11:57.080 --> 0:12:00.480 a cat box, and at some point the cat uh 0:12:00.760 --> 0:12:03.880 causes quite a stink in there, and you register it first, 0:12:03.920 --> 0:12:05.679 and like, jeez, I should stop what I'm doing and 0:12:05.760 --> 0:12:08.360 clean out that cat box, but you keep working and 0:12:08.400 --> 0:12:11.000 then after a while you don't smell it anymore. But 0:12:11.080 --> 0:12:13.640 then maybe you step outside to check the mail, or 0:12:13.760 --> 0:12:15.719 you go to the you know, the grocery store to 0:12:15.720 --> 0:12:17.880 pick something up, or you know, your partner comes home 0:12:18.160 --> 0:12:22.079 and when when you or they enter the room, you go, cheese, 0:12:22.120 --> 0:12:24.880 what happened in here? Did the cat do something? Again? No, 0:12:24.920 --> 0:12:28.880 you're just re smelling there. Originally you forgot about it, right, Yeah, 0:12:28.880 --> 0:12:32.160 the the brain kind of decides. Look at this point, 0:12:32.240 --> 0:12:35.160 I assume you know that she that the the cheese 0:12:35.240 --> 0:12:37.960 is nasty and you're not going to eat it um 0:12:38.120 --> 0:12:41.120 or or that yes, there is animal poop around here. 0:12:41.160 --> 0:12:43.720 It's it's done, it's part. It's warned you. But the 0:12:43.840 --> 0:12:47.040 good smell will keep resonating because the good smell is 0:12:47.440 --> 0:12:50.400 probably saying, hey, there's something over here delicious to eat 0:12:50.679 --> 0:12:54.320 with some fresh berries or whatnot. And it'll keep saying, hey, 0:12:54.360 --> 0:12:57.120 the berries are still here. Why haven't you eaten them yet? 0:12:57.400 --> 0:12:59.319 This is there's not a lot of sugar out here. 0:12:59.480 --> 0:13:01.920 You should it at these berries? Why you have a chance. 0:13:03.160 --> 0:13:07.440 So the beautiful remains beautiful, The sweet smelling remains sweet smelling. 0:13:07.679 --> 0:13:11.000 But something that is disgusting, even as you know, discussing 0:13:11.000 --> 0:13:13.720 as a foul smell, we can grow accustomed to. Okay, 0:13:13.720 --> 0:13:18.000 So what's the analogy to the Uncanny Valley here? Well, 0:13:18.040 --> 0:13:21.040 I think the analogy here is that if you have 0:13:21.440 --> 0:13:27.200 a discussed response to to visual sense information regarding an 0:13:27.240 --> 0:13:32.080 individual's appearance, a robots appearance, a computer animated characters appearance, 0:13:32.920 --> 0:13:35.360 there could there could also be this disgust drop off 0:13:35.400 --> 0:13:37.560 point as you grow accustomed to it. Oh, and that's 0:13:37.559 --> 0:13:39.320 something that we have seen. We talked about a little 0:13:39.320 --> 0:13:42.560 bit in the last episode, and that, uh, some people 0:13:42.600 --> 0:13:47.240 attest that when you spend time around these robots or uh, 0:13:47.480 --> 0:13:49.960 well it was mainly for the robots. The robots that 0:13:50.000 --> 0:13:53.400 at first seemed creepy, they they stop bothering. You become 0:13:53.440 --> 0:13:56.320 accustomed to them, they're not creepy anymore. I don't so 0:13:56.400 --> 0:13:58.400 much know if that's always going to be the case 0:13:58.440 --> 0:14:03.120 for creepy looking animated humans. Um, but who knows. Now. 0:14:03.160 --> 0:14:05.040 Another thing to keep in mind too, is that the 0:14:05.080 --> 0:14:07.319 discussed response is going to depend on a number of 0:14:07.360 --> 0:14:10.040 different factors, some of which are going to be tied 0:14:10.400 --> 0:14:14.240 to hormonal situations. So, for instance, pregnant women are more 0:14:14.280 --> 0:14:17.000 sensitive to discussed and this is linked to their elevated 0:14:17.000 --> 0:14:20.760 progesterone levels. So uh, and of course there's gonna be 0:14:20.760 --> 0:14:24.000 other factors beyond that for every individual. Yeah. So I 0:14:24.040 --> 0:14:28.400 think that there's certainly could be some amount of biological 0:14:29.000 --> 0:14:32.280 instinctual response going on in the uncanny value effect, to 0:14:32.320 --> 0:14:34.840 the extent that it exists. But I think also based 0:14:34.880 --> 0:14:36.800 on what we've seen so far, that probably does not 0:14:36.880 --> 0:14:40.000 account for all of it. Uh, And I think another 0:14:40.040 --> 0:14:44.080 thing to consider would be going to more more complex 0:14:44.120 --> 0:14:48.520 sort of cognitive psychology, such as cognitive dissonance. Now, if 0:14:48.600 --> 0:14:51.840 you were to just ask me what I thought was 0:14:51.920 --> 0:14:55.240 the most likely answer before I got into the research, 0:14:56.040 --> 0:14:58.480 I would intuitively tend to think that the best answer 0:14:58.520 --> 0:15:02.880 for what causes the uncanny value effect primarily is our 0:15:02.960 --> 0:15:07.120 inherent discomfort with category confusions. Uh. This is something that 0:15:07.160 --> 0:15:09.480 I think about a lot, in like the creation of 0:15:09.520 --> 0:15:12.000 monster mythology and stuff like that. We we don't like 0:15:12.120 --> 0:15:15.880 the feeling produced by by things that don't fit into 0:15:15.920 --> 0:15:19.960 our normal taxonomys for objects in the world and seem 0:15:20.040 --> 0:15:24.080 to violate our tax taxonomic ordering system. And this is 0:15:24.080 --> 0:15:27.040 why I think monsters are so often hybrids of existing 0:15:27.120 --> 0:15:31.000 things a bull's head on a man's body, things that 0:15:31.240 --> 0:15:36.080 defy our intuitive classification rules. They make us uncomfortable and 0:15:36.200 --> 0:15:39.400 cause a sense of unease, leading to this uncanny feeling. 0:15:40.280 --> 0:15:43.480 So that that's what I would have intuitively said, Yeah, 0:15:43.480 --> 0:15:45.160 that makes sense. Is it wolf or is it man? 0:15:45.320 --> 0:15:49.200 Is it or is it man? Right? But negative affinity 0:15:49.280 --> 0:15:53.680 resulting from this difficulty and assigning the entity to a category. 0:15:54.040 --> 0:15:57.680 Is it robot or human. Despite my intuitive favor for 0:15:57.720 --> 0:16:00.800 this explanation, I think it looks like your mental evidence 0:16:00.840 --> 0:16:03.080 for this is not strong. And in fact, in some 0:16:03.120 --> 0:16:05.640 ways some of the studies we've looked at have somewhat 0:16:05.680 --> 0:16:08.720 invalidated this. For example, in the last episode at the end, 0:16:08.760 --> 0:16:11.720 I was talking about that study by Mather and Rifling, 0:16:12.520 --> 0:16:15.720 and it did not find evidence of a strong correlation 0:16:16.120 --> 0:16:19.400 between here. Here's what you notice here, the time it 0:16:19.480 --> 0:16:23.920 took people to rate the mechano humanoid qualities of a robot, 0:16:24.640 --> 0:16:28.760 and that robot pictures likability. So if you were talking 0:16:28.760 --> 0:16:31.800 it up to category confusion, you would probably think the 0:16:31.960 --> 0:16:35.280 robots that people took the longest time to figure out 0:16:35.280 --> 0:16:39.160 how to rate on the mechano humanoid scale, and those 0:16:39.200 --> 0:16:41.600 would be the least likable, right, because they're the ones 0:16:41.640 --> 0:16:45.840 causing the most category confusion. Right. But I mean you 0:16:45.840 --> 0:16:47.760 could also say that, well, you're not maybe having not 0:16:47.880 --> 0:16:50.640 a you're not having a visceral gut reaction to them either. 0:16:50.680 --> 0:16:53.480 You haven't to think it out and try and you know, 0:16:53.560 --> 0:16:55.720 analyze how you feel about it. Oh yeah, I'd say 0:16:55.720 --> 0:16:58.280 that's exactly what we don't like. I mean, we like 0:16:58.400 --> 0:17:01.200 to be able to viscerally category arise things. But but 0:17:01.240 --> 0:17:04.600 then the uncanny valley is often, at least in terms 0:17:04.640 --> 0:17:06.359 of the way you're going to find it invoked by 0:17:06.359 --> 0:17:09.919 the average person. It's often discussed if it's as if 0:17:09.960 --> 0:17:12.479 it's a visceral reaction, the sort of i'll kill it 0:17:12.480 --> 0:17:15.359 with fire reaction people might have where I hate that 0:17:15.440 --> 0:17:17.720 saying yes, I'm not a fan as well, especially when 0:17:17.720 --> 0:17:21.760 it is applied to outside of of of fictional connotations. Right. 0:17:22.320 --> 0:17:24.960 But so, they point out The authors here point out 0:17:25.000 --> 0:17:28.040 that they did not find this. Uh. They point out 0:17:28.080 --> 0:17:31.359 that it's just not a fact that the things that 0:17:31.440 --> 0:17:34.480 took longer to look at and decide where the least likable. 0:17:34.560 --> 0:17:38.000 Though while this is not statistically important, just as a 0:17:38.040 --> 0:17:41.040 point of curiosity, the single face that took the longest 0:17:41.119 --> 0:17:44.840 to rate on the mechanical versus humanoid quality was also 0:17:45.040 --> 0:17:48.080 just about the most disliked face in their whole collection 0:17:48.160 --> 0:17:51.280 of faces, But that was just like one outlier overall, 0:17:51.320 --> 0:17:55.560 this did not present as a general effect. Other studies 0:17:55.560 --> 0:17:58.160 have also looked into this and have failed to find 0:17:58.200 --> 0:18:01.320 solid evidence for category confused and as the primary driver 0:18:01.480 --> 0:18:05.160 of the negative affinity at the bottom of the uncanny valley. 0:18:05.200 --> 0:18:07.800 So it looks like my intuitions here I think are wrong. 0:18:08.480 --> 0:18:11.840 But there's something that's kind of related as an idea 0:18:11.880 --> 0:18:15.920 that's been explored, and that is the idea of perceptual mismatch. 0:18:16.520 --> 0:18:20.360 So several authors have advocated the idea that this perceptual 0:18:20.480 --> 0:18:24.160 mismatch could be the primary cause of what we don't 0:18:24.200 --> 0:18:27.400 like about things that we would intuitively say fall into 0:18:27.400 --> 0:18:30.680 the uncanny valley. So one piece of research I want 0:18:30.680 --> 0:18:33.840 to mention is a sort of review by uh Cat 0:18:33.920 --> 0:18:37.320 Siri that is a review of empirical evidence on different 0:18:37.400 --> 0:18:41.879 uncanny valley hypotheses support for perceptual mismatch as one road 0:18:41.960 --> 0:18:44.399 to the valley of eeriness that gott to give it 0:18:44.440 --> 0:18:49.639 a different name, apparently in Frontiers in Psychology. And so 0:18:49.680 --> 0:18:53.280 in this study, the author's review present research and claim 0:18:53.400 --> 0:18:56.520 that experimental research attempting to show the uncanny valley has 0:18:56.520 --> 0:18:59.880 been inconsistent. They don't exactly say that the uncanny valley 0:19:00.000 --> 0:19:03.000 doesn't exist, but that it's not as simple as often 0:19:03.040 --> 0:19:05.120 believed to be. Something we've been saying for a while now. 0:19:05.200 --> 0:19:09.600 It's it's not that any manipulation of the variable of 0:19:09.680 --> 0:19:14.399 human likeness leads to Uncanny Valley effects. So, in other words, 0:19:14.520 --> 0:19:17.520 that the horizontal axis on the graph is more complicated 0:19:17.600 --> 0:19:20.399 than just the question of how realistically human is it. 0:19:20.480 --> 0:19:22.440 I've seen this come up enough now that I'm pretty 0:19:22.440 --> 0:19:26.040 convinced that that is not necessarily the only or even 0:19:26.080 --> 0:19:29.560 the primary factor here. But they still recognize that there 0:19:29.640 --> 0:19:31.399 is some kind of effect here. So they claim that 0:19:31.440 --> 0:19:34.879 there's evidence against the category confusion basis that we were 0:19:34.920 --> 0:19:37.400 just talking about. But they claim that there has been 0:19:37.400 --> 0:19:41.040 good evidence in support of the perceptual mismatch hypothesis, And 0:19:41.080 --> 0:19:42.919 I want to read what they say. They say quote 0:19:43.160 --> 0:19:47.280 taken together, The present review suggested that although not any 0:19:47.400 --> 0:19:50.520 kind of human likeness manipulation leads to the uncanny Valley, 0:19:50.520 --> 0:19:54.720 the uncanny valley could be caused by more specific perceptual 0:19:54.800 --> 0:19:59.600 mismatch conditions. Such conditions could originate at least from inconsistent 0:19:59.760 --> 0:20:04.160 real is um levels between individual features, and the examples 0:20:04.160 --> 0:20:07.480 they give would be like artificial eyes on a human 0:20:07.520 --> 0:20:11.479 like face, or the presence of atypical features such as 0:20:11.560 --> 0:20:16.320 a typically large eyes on an otherwise humanlike character. So 0:20:16.359 --> 0:20:19.840 what they're saying there is not necessarily that you can't 0:20:19.880 --> 0:20:22.600 tell whether it's a robot or a human, but that 0:20:22.800 --> 0:20:25.359 there have been multiple experiments that seem to show people 0:20:25.400 --> 0:20:30.520 are unsettled and made unhappy by things where the features 0:20:30.560 --> 0:20:33.400 on the face or the features of the figure as 0:20:33.400 --> 0:20:38.239 a whole are inconsistently realistic. Like we're more okay with 0:20:38.280 --> 0:20:42.840 a robot that's consistently realistic at a certain level at 0:20:42.880 --> 0:20:46.720 maybe say seventy percent, than something that has eyes at 0:20:46.800 --> 0:20:52.439 nine percent and skin attent. This reminds me. This is 0:20:52.480 --> 0:20:53.560 just coming off the top of my head. So I 0:20:53.600 --> 0:20:56.560 don't have the artist's name here, but there's there's an 0:20:56.600 --> 0:20:59.480 artist who's worth made the rounds where they took cartoon characters, 0:21:00.040 --> 0:21:03.960 they depicted them realistically. So it's Simpson. Yes, Homer Simpson 0:21:04.040 --> 0:21:07.280 with like pores on his skin, you know, like horrible. Yeah, 0:21:08.200 --> 0:21:09.680 So that comes to mind. It is a kind of 0:21:09.680 --> 0:21:12.159 a possible example of this. Yeah, I think that's a 0:21:12.160 --> 0:21:14.600 good explanation. So I want to get into my main 0:21:14.680 --> 0:21:17.400 takeaways from looking at the Uncanny Valley research so far. 0:21:17.440 --> 0:21:19.240 Maybe you can let me know what you think about this. 0:21:19.760 --> 0:21:22.200 I'd say, first of all, I think the Uncanny Valley 0:21:22.280 --> 0:21:24.919 is a real thing, but it's not as simple as 0:21:24.960 --> 0:21:28.880 Maury's original hypothesis would lead you to believe. Uh. First 0:21:28.920 --> 0:21:32.080 of all, people definitely do get creeped out by lots 0:21:32.119 --> 0:21:36.680 of almost human looking things, But it's not necessarily just 0:21:36.840 --> 0:21:41.879 that the near failed human realism is what makes them unsettling. 0:21:41.880 --> 0:21:45.040 There are other things that appear to be making them unsettling, though, 0:21:45.080 --> 0:21:49.359 that the near humanness plays some kind of role. And 0:21:49.400 --> 0:21:51.000 the other big thing is that there appear to be 0:21:51.080 --> 0:21:56.280 multiple dimensions to explain the phenomenon. Right, So, synthetic humanoid images, 0:21:56.280 --> 0:22:01.359 whether robotic or animated, offer multiple dimensions of attraction and revulsion. 0:22:01.800 --> 0:22:05.400 I think it's possible that there are some biologically triggered effects, 0:22:05.760 --> 0:22:08.840 the appearance of health or disease, the appearance of life 0:22:08.920 --> 0:22:11.360 or death. But then I think there are possibly other 0:22:11.400 --> 0:22:17.040 things triggered by psychological cognitive dissonance. Probably not category confusion, um, 0:22:17.200 --> 0:22:20.320 but but some good evidence for this idea of the 0:22:20.359 --> 0:22:24.440 perceptual mismatch being the cause. And then the final thing 0:22:24.680 --> 0:22:29.200 is that the Uncanny Valley effect is context dependent. How 0:22:29.240 --> 0:22:32.440 long have you been exposed to the image in what's setting? 0:22:33.160 --> 0:22:35.440 Is it part of a narrative or some other context 0:22:35.560 --> 0:22:38.320 in which you're being asked to suspend your disbelief or 0:22:38.359 --> 0:22:42.560 otherwise put yourself in a state of openness. Robert, what 0:22:42.560 --> 0:22:44.239 do you think about all this so far? You've got 0:22:44.280 --> 0:22:47.760 any disagreement? No, I mean I feel like my view 0:22:47.760 --> 0:22:50.680 on it closely lines up with with with yours here 0:22:50.720 --> 0:22:53.640 basically that it's just that there is an effect going on, 0:22:53.720 --> 0:22:57.600 but it's far it's far more nuanced than simply oh 0:22:57.640 --> 0:23:00.760 these are these are the factors that make something fall 0:23:00.840 --> 0:23:04.600 into the uncanny valley, not just how realistically human. There's 0:23:04.640 --> 0:23:07.080 there's other stuff going on, right, and not just a 0:23:07.119 --> 0:23:12.120 mere hybridization. I was instantly thinking about um, the borg 0:23:12.200 --> 0:23:15.360 and uh and occasionally, you know the sort of sexy 0:23:15.400 --> 0:23:19.119 boards that show up in the Star Trek universe, like 0:23:19.160 --> 0:23:23.720 there's there's clearly category confusion going on there, but they're 0:23:23.760 --> 0:23:27.720 they're not depicted as particularly uncanny, like it was the 0:23:27.760 --> 0:23:31.280 board Queen. Was she uncanny kind of? I don't know. 0:23:31.440 --> 0:23:33.040 I mean, even when I'm not a big fan of 0:23:33.080 --> 0:23:38.800 the way the board look well, but the outside of 0:23:38.800 --> 0:23:41.199 the board. You can also think of, you know, various 0:23:41.240 --> 0:23:45.360 sort of hybrid human creatures depicted in fantasy and fiction 0:23:45.840 --> 0:23:48.719 that that are created in such a way to the 0:23:48.720 --> 0:23:53.240 the alluring. Uh, like they managed to fetishize the inhuman 0:23:53.320 --> 0:23:57.840 qualities of them the category confusion. Yeah, uh yeah. It 0:23:57.880 --> 0:24:01.119 makes me think that there are certain qualities of the 0:24:01.200 --> 0:24:04.639 human appearance that, if altered, are much more significant in 0:24:04.720 --> 0:24:08.359 terms of our affects response than others. So it could 0:24:08.440 --> 0:24:10.399 be I'm just making this up. I don't know if 0:24:10.440 --> 0:24:14.479 this is true, but that like getting the getting the 0:24:14.640 --> 0:24:18.560 size of the eyes wrong could quite easily lead to 0:24:18.600 --> 0:24:22.600 a disgust response and revulsion, but getting the size of 0:24:22.640 --> 0:24:26.359 the nose wrong wouldn't. Does that make sense? Yeah? Yeah, 0:24:26.480 --> 0:24:29.480 or just thinking of eyes like definitely, making the eyes 0:24:30.080 --> 0:24:33.720 inappropriately large leads a creepy factor, and this is often employed. 0:24:34.440 --> 0:24:38.320 I instantly think of the vampire movie. What twenty thirty 0:24:38.400 --> 0:24:40.760 days of Night is that the name of it? Oh, 0:24:40.920 --> 0:24:42.480 something like that, Yeah, where they did some sort of 0:24:42.520 --> 0:24:44.720 digital effect to make ory. I don't remember the number, 0:24:45.440 --> 0:24:48.600 some number of days of PPD days of night. Um, 0:24:48.920 --> 0:24:53.119 I didn't actually see if the trailers were we're certainly interesting. 0:24:53.560 --> 0:24:55.520 But likewise, if you just take an individual and have 0:24:55.600 --> 0:24:58.840 them wear blackout contact lenses, yeah, like that is often 0:24:58.960 --> 0:25:00.800 sometimes that's played up for creepiness, but a lot of 0:25:00.800 --> 0:25:04.080 times it's played up for to be alluring. You'll have uh, 0:25:04.480 --> 0:25:08.520 male or female characters that are are otherwise dressed in 0:25:08.560 --> 0:25:11.040 some alluring fashion, but they have blacked out eyes and 0:25:11.080 --> 0:25:14.560 it's who it's kind of like supernatural, sexy cool, as 0:25:14.560 --> 0:25:16.639 opposed to just like, oh my god, why why are 0:25:16.640 --> 0:25:21.719 your eyes pits of darkness? Yeah, but maybe that's just me. Okay, Well, 0:25:21.760 --> 0:25:23.320 I think we should take a quick break and when 0:25:23.320 --> 0:25:30.760 we come back, we will go beyond the Uncanny Valley. 0:25:31.400 --> 0:25:34.560 All right, we're back, Okay, So Robert, I can recall 0:25:34.680 --> 0:25:37.440 discussions going back for years about whether we're going to 0:25:37.520 --> 0:25:39.560 make it out of the Uncanny Valley in the realm 0:25:39.600 --> 0:25:42.399 of robotics or animation, And I think from here on 0:25:42.880 --> 0:25:45.960 I want to focus primarily on animation, just just to 0:25:46.040 --> 0:25:49.000 keep us focused. And I think there are actually two 0:25:49.000 --> 0:25:52.960 separate questions here, assuming that the Uncanny Valley is to 0:25:53.119 --> 0:25:55.800 some extent to go heerent idea. We've already explained all 0:25:55.840 --> 0:25:58.800 the ways in which it's obviously way more complicated than 0:25:58.840 --> 0:26:03.800 the naive popular culture culture understanding of it. But the 0:26:03.880 --> 0:26:07.760 two big questions. Number one, can we make realistic looking 0:26:07.840 --> 0:26:11.240 humanoid characters that aren't creepy I think the answer here 0:26:11.280 --> 0:26:14.560 is yes, I think it's not a two dimensional graph. 0:26:14.640 --> 0:26:16.879 I think you can make things that aren't quite photo 0:26:16.920 --> 0:26:21.160 realistic but look realistic that aren't creepy. Current generation video 0:26:21.200 --> 0:26:23.639 games have been doing this, and as I mentioned in 0:26:23.640 --> 0:26:26.119 the last episode, I think that their tricks to doing 0:26:26.160 --> 0:26:30.159 this it's apparently in achieving like the right combination of 0:26:30.240 --> 0:26:33.760 realistic traits and unrealistic traits that maybe you would just 0:26:33.920 --> 0:26:37.119 land on by doing trial and error in design over time. 0:26:37.600 --> 0:26:41.600 You would never mistake these characters for photographs of real humans. 0:26:42.080 --> 0:26:45.720 But they're also not cartoony. They've got this feeling of 0:26:45.920 --> 0:26:49.919 really real ishness, if you know what I mean. But 0:26:50.000 --> 0:26:53.480 they've they've they've attained sort of a generally acceptable plateau 0:26:53.680 --> 0:26:57.520 of realistic affect, but they're not skewing into these different 0:26:57.600 --> 0:27:01.600 danger zones adjacent photo real is him where the shortcomings 0:27:01.640 --> 0:27:04.200 become creepy and off putting and we don't like it. 0:27:05.800 --> 0:27:08.600 Then there would be another question, and that's just can 0:27:08.680 --> 0:27:13.720 we make animated characters that are robustly indistinguishable from human like? 0:27:13.760 --> 0:27:16.240 Can we get all the way up the other side 0:27:16.480 --> 0:27:20.760 of the mountain, up to the peak of reality? And 0:27:20.920 --> 0:27:23.520 last week, if you'd asked me, my personal answer would 0:27:23.520 --> 0:27:26.240 have been no, not yet. But I think that's actually 0:27:26.280 --> 0:27:29.240 not as clear cut as we would first guess, because 0:27:29.280 --> 0:27:31.520 you think that perhaps many of the humans we see 0:27:31.520 --> 0:27:35.439 on TV are actually digital creations. Oh I know for 0:27:35.520 --> 0:27:40.600 a fact that that who's that guy, that Jimmy Fallon guy. 0:27:40.720 --> 0:27:45.359 Oh yeah, Jimmy Fallon might be a computer generated that 0:27:45.480 --> 0:27:48.159 is not a person. He has been generated by a 0:27:48.160 --> 0:27:53.159 computer that's in Palo Alto, California. It's a supercomputer. I mean, 0:27:53.160 --> 0:27:55.680 it's a really good computer. But yeah, Well, I often 0:27:55.720 --> 0:27:58.720 feel the same way about Michael Fastbender, And granted it's 0:27:58.720 --> 0:28:01.280 complicated by the fact that he has he has a 0:28:01.359 --> 0:28:04.919 thing for playing androids recently, but at times you're just like, no, 0:28:05.160 --> 0:28:07.359 he's just a little too handsome. There's something in human 0:28:07.400 --> 0:28:10.879 about this, this man's handsomeness, in his charm. So I 0:28:10.920 --> 0:28:13.639 want to talk about one thing that is that has 0:28:13.680 --> 0:28:16.640 given me pause on this subject, and it's going back 0:28:16.680 --> 0:28:18.000 to what we talked about at the beginning of the 0:28:18.080 --> 0:28:21.840 last episode Rogue one. Yes, So back to the c 0:28:22.000 --> 0:28:24.640 G I grandmof Tarken when I saw when I first 0:28:24.680 --> 0:28:26.600 saw Rogue one, I liked a lot of things about 0:28:26.600 --> 0:28:28.720 the movie. But I did not really like the the 0:28:28.800 --> 0:28:32.719 c g I Tarken, the the almost Peter Cushing was 0:28:33.320 --> 0:28:36.040 very very good, and I really mean that, I mean 0:28:36.720 --> 0:28:40.280 shockingly good, but still not quite real to me, still 0:28:40.360 --> 0:28:44.160 kind of distracting because of how slightly off it was. 0:28:44.320 --> 0:28:47.080 I would not have mistaken it for a real person. 0:28:47.680 --> 0:28:50.240 But the other day I was talking in the office 0:28:50.280 --> 0:28:53.200 to Holly fry Who. It's Holly a Star Wars fan. 0:28:54.760 --> 0:28:56.760 She's so, she's one of the hosts of Stuff You 0:28:56.800 --> 0:28:58.880 Missed in History Class, one of our one of our 0:28:59.440 --> 0:29:02.600 podcast in the podcast family here. Yeah, she has hands 0:29:02.640 --> 0:29:05.320 down the most Star Wars knowledgeable person in the office 0:29:05.360 --> 0:29:08.360 in an office full of nerds. Yeah, I should point 0:29:08.400 --> 0:29:12.280 out that Holly has slash had like a golden ticket 0:29:12.320 --> 0:29:14.479 to go see Rogue one anytime she wanted to at 0:29:14.520 --> 0:29:18.360 the theater. What are you serious? And serious? This is? Yeah? Yeah, 0:29:18.400 --> 0:29:20.960 leget where'd that come from? I don't know, I don't know. 0:29:21.000 --> 0:29:23.280 I'm not I'm not at that level of a fandom 0:29:23.320 --> 0:29:25.840 where I'm even offered such things. I know. So if 0:29:26.160 --> 0:29:28.160 you want to have a funny experience, go up to 0:29:28.200 --> 0:29:32.280 Holly and just like ask some really obscure random question 0:29:32.360 --> 0:29:35.240 about Star Wars where you think it could not be 0:29:35.360 --> 0:29:38.960 possible that there's an actual answer to this, like, oh, 0:29:39.040 --> 0:29:42.040 that stormtrooper on the left, where did he? What planet 0:29:42.080 --> 0:29:44.400 was he born on? Holly will have an answer. She'll 0:29:44.440 --> 0:29:46.840 be like, oh, that was actually addressed in dialogue in 0:29:46.880 --> 0:29:50.360 the Greek dub of this episode of Clone Wars. So 0:29:50.480 --> 0:29:53.160 Holly has amazing Star Wars knowledge. She is super fun 0:29:53.200 --> 0:29:56.160 to talk to about the Star Wars universe. But anyway, 0:29:56.600 --> 0:29:58.840 Holly pointed out that while a lot of people like 0:29:58.960 --> 0:30:01.160 me were saying that the c G I. Tarkan was 0:30:01.200 --> 0:30:04.760 a few pixels short of escaping the valley, then again, 0:30:04.800 --> 0:30:07.280 there were plenty of people, including some older people that 0:30:07.400 --> 0:30:10.560 she knew, who couldn't tell that it wasn't a real person. 0:30:11.080 --> 0:30:14.480 They literally they couldn't tell. Well, I'm worried about my 0:30:14.560 --> 0:30:17.280 viewing upcoming viewing of the film because I have I've 0:30:17.320 --> 0:30:21.840 been preconditioned to have a certain response to the to 0:30:21.960 --> 0:30:25.440 the Tarkan bot here. Oh, I'm sorry that we've had 0:30:25.440 --> 0:30:27.520 this discussion before you were able to see the movie 0:30:27.560 --> 0:30:30.840 for yourself. It'll be interesting being preconditioned. Will I go 0:30:30.920 --> 0:30:34.880 into it, you know, expecting an abomination, and like that 0:30:35.000 --> 0:30:37.440 the level of detail will overwhelm me and it won't matter. 0:30:37.920 --> 0:30:40.040 Or am I going to go into there and nothing 0:30:40.120 --> 0:30:41.960 is gonna fool me because I'm gonna be looking for 0:30:41.960 --> 0:30:46.440 the cracks, You know, I can. I can put myself 0:30:46.440 --> 0:30:51.480 in a mindset where I think it's possible. I might 0:30:51.520 --> 0:30:54.160 not have known that it was a c g I 0:30:54.200 --> 0:30:56.760 effect if I wasn't familiar with what to look for, 0:30:56.840 --> 0:30:58.680 Like if I didn't watch a lot of movies that 0:30:58.800 --> 0:31:00.960 had c g I effects in them, and if I 0:31:01.000 --> 0:31:03.680 didn't complain about c g I a lot. I'm sorry, 0:31:03.680 --> 0:31:07.400 guilty is I'm guilty of that? Uh? If I wasn't 0:31:07.480 --> 0:31:11.280 aware that Peter Cushing was dead, Um, if it wasn't 0:31:11.320 --> 0:31:13.400 if I wasn't sort of prepared to see a lot 0:31:13.440 --> 0:31:15.760 of high tech c g I by virtue of the 0:31:15.800 --> 0:31:18.480 fact that I'm sitting in a theater for a Lucasfilm movie, 0:31:18.960 --> 0:31:21.240 all those things. If you took away all the context 0:31:21.320 --> 0:31:25.080 and my pre knowledge, I'm very might possibly have fallen 0:31:25.120 --> 0:31:27.440 for it. I think I might have just it might 0:31:27.480 --> 0:31:29.520 have just gone past me. If I was absorbed in 0:31:29.520 --> 0:31:31.840 the story, I might have thought, Yeah, it's kind of strange. 0:31:31.880 --> 0:31:35.440 Looking dude. But it's just a dude, and I think 0:31:35.480 --> 0:31:39.320 it gets better or worse depending on your perspective. So 0:31:39.400 --> 0:31:44.880 I have in October newspiece from BBC Asia here about 0:31:44.880 --> 0:31:49.800 a character called Saia, a computer animated character created by 0:31:49.920 --> 0:31:53.320 the Japanese husband and wife graphic design team Terry Yuki 0:31:53.400 --> 0:31:56.800 and Yuka Ishikawa. And I mentioned this one in particular 0:31:56.840 --> 0:31:58.719 because before we did this episode, I went and I 0:31:58.760 --> 0:32:00.840 looked up what can that ared a lot of the 0:32:00.880 --> 0:32:05.560 most realistic c g I character creations, the most impressive animations, 0:32:06.040 --> 0:32:08.120 this one came up, and I think this was the 0:32:08.120 --> 0:32:12.000 most impressive to me. It's probably the most photo real 0:32:12.600 --> 0:32:16.320 computer generated human I've come across so far. So Siah 0:32:16.480 --> 0:32:19.400 is supposed to be a seventeen year old Japanese student 0:32:20.080 --> 0:32:22.760 and the creators have been working on her design for 0:32:22.800 --> 0:32:25.840 a couple of years now, and as versions of Siah 0:32:25.920 --> 0:32:28.720 have been posted on the Internet, people have widely reacted 0:32:28.720 --> 0:32:31.760 with comments like I can't believe that's not a real person, 0:32:32.920 --> 0:32:35.520 and I kind of have to agree. I'm looking at 0:32:35.560 --> 0:32:37.760 these these pictures of her. There are a couple of 0:32:37.760 --> 0:32:41.720 different generations of her design up, and the most recent 0:32:41.720 --> 0:32:45.880 one just looks like a photograph of a person. Yeah, 0:32:45.920 --> 0:32:48.960 I I can't tell that that is not a person. 0:32:49.160 --> 0:32:53.680 I I have no recourse to critical faculties in my 0:32:53.720 --> 0:32:56.080 mind that would say no, here's where you can tell 0:32:56.160 --> 0:32:59.200 that that's not a real person. Now, at the same time, 0:32:59.240 --> 0:33:00.720 I do have to come to to come in a 0:33:00.800 --> 0:33:03.520 made earlier that that this is also a it's a 0:33:03.640 --> 0:33:06.840 it's a it's a pretty face, it's a very standard 0:33:06.880 --> 0:33:10.480 face like this is this is leading lady material. Whereas 0:33:10.840 --> 0:33:12.880 I think it gets more problematic when you look at 0:33:12.920 --> 0:33:16.760 the character actor type figures such as Peter Cushing exactly. 0:33:16.840 --> 0:33:19.520 They have such a distinctive face, yes, and Peter Cushing. 0:33:19.840 --> 0:33:22.040 So this is one thing that helps, I think is 0:33:22.080 --> 0:33:26.800 that it's a young character who has very smooth features. Uh, 0:33:26.840 --> 0:33:29.480 Peter Cushing has a lot of cracks and crags, a 0:33:29.560 --> 0:33:31.479 lot of wrinkles, And I think that that may actually 0:33:31.520 --> 0:33:35.400 be simply having more texture on your face could make 0:33:35.440 --> 0:33:38.479 it much more difficult to make a photo real copy 0:33:38.560 --> 0:33:42.840 of you. That's entirely possible. But to get back to Saia. 0:33:43.000 --> 0:33:46.760 So in October of last year, the Artist's debuted the 0:33:46.800 --> 0:33:50.800 first animated clip of Siah, which they created using motion 0:33:50.840 --> 0:33:54.760 capture technology, and they debuted at a Japanese consumer electronics show. 0:33:55.520 --> 0:33:59.280 I watched this footage, and I think Maury's distinction about 0:33:59.320 --> 0:34:03.800 having different standards for motion and still images does apply, 0:34:03.960 --> 0:34:07.200 because while with the still image, I can't tell that's 0:34:07.240 --> 0:34:10.600 not a real person, with the short animated clip, I can. 0:34:10.760 --> 0:34:12.560 I can tell it's not a real person, but it's 0:34:12.560 --> 0:34:18.520 still very very impressive, not as absolutely photo real as 0:34:18.520 --> 0:34:21.880 the still images. But I don't know. I wonder to 0:34:21.960 --> 0:34:26.239 what extent this gap is just um that motion animation 0:34:26.280 --> 0:34:29.400 is a bigger technical project, it takes more investment and 0:34:29.440 --> 0:34:32.400 money and all that um, And to what extent the 0:34:32.400 --> 0:34:36.000 gap is within the viewers mind? Essentially, to what extent 0:34:36.080 --> 0:34:38.480 it's caused by the fact that the climb out of 0:34:38.520 --> 0:34:42.560 the uncanny valley is steeper if you're moving. Now. I 0:34:42.600 --> 0:34:44.719 know we're not talking about robots here, but this, of course, 0:34:44.760 --> 0:34:47.120 this brings up the thought that as where as we 0:34:47.160 --> 0:34:51.000 attempt to conquer this in the realm of humanoid robotics, 0:34:51.680 --> 0:34:54.319 you're gonna inevitably have situations where oh, it looks just 0:34:54.360 --> 0:34:56.600 like a person that it's walking down the street, but 0:34:56.680 --> 0:35:00.000 if it climbs stairs, there you go it's not necessarily 0:35:00.200 --> 0:35:03.160 edge to oh nine, there it lays in the air, 0:35:03.480 --> 0:35:05.640 but maybe there would be something telling like, oh, it 0:35:05.680 --> 0:35:07.640 looks like a human most of the time. But they're 0:35:07.640 --> 0:35:12.880 gonna be certain movements, certain environmental reactions that are just 0:35:12.920 --> 0:35:15.920 not going to hold up to scrutiny, right, So, I 0:35:15.920 --> 0:35:18.520 don't know. Looking at these things, looking at Tarkan in 0:35:18.840 --> 0:35:22.000 Rogue one, looking at Siah, I think it's clear that 0:35:22.040 --> 0:35:25.880 we're getting closer and closer to really bridging the gap 0:35:26.080 --> 0:35:31.520 on indistinguishable photo real humanity and computer animation. Whether you'd 0:35:31.520 --> 0:35:34.080 call that an uncanny value or not. Obviously this is 0:35:34.120 --> 0:35:38.000 a related but maybe different issue. What we're definitely drawing 0:35:38.040 --> 0:35:41.759 near is that peak of reality where there are synthetically 0:35:41.760 --> 0:35:45.440 generated images of humans that you can't tell from the 0:35:45.520 --> 0:35:49.600 real thing. And since things in the uncanny value are creepy, 0:35:49.640 --> 0:35:52.120 I think we usually just assume that overcoming it is 0:35:52.120 --> 0:35:55.359 a good thing, right, like designs getting better? Uh, and't 0:35:55.360 --> 0:35:57.400 it kind of cool that we can generate these photo 0:35:57.440 --> 0:36:02.000 real images without without actually having to photograph someone. But 0:36:02.160 --> 0:36:04.919 I'm not so sure that's a good thing. I think 0:36:04.960 --> 0:36:07.359 we should maybe think about the implications of this, like 0:36:08.000 --> 0:36:11.880 what would happen in the world where human simulations, especially 0:36:11.920 --> 0:36:16.719 computer animation, can reliably climb up that second peak. So 0:36:16.840 --> 0:36:18.600 we are going to take a quick break and when 0:36:18.600 --> 0:36:25.040 we come back, we will go beyond the Uncanny Valley. 0:36:25.920 --> 0:36:28.000 All right, we're back. You know, I'm glad you brought 0:36:28.040 --> 0:36:29.879 up this idea of you know, is it a bad thing? 0:36:29.920 --> 0:36:32.000 Is it a good thing? It does remind me of 0:36:32.040 --> 0:36:33.920 a fabulous book that came out a few years back. 0:36:33.960 --> 0:36:35.880 I think I just referenced it in our Sex Spots 0:36:35.920 --> 0:36:40.480 episode toddled The wind Up Girl by Paolo bag Um. 0:36:40.520 --> 0:36:44.560 It's a near future science fiction tale, just really a 0:36:44.560 --> 0:36:48.279 wonderful novel, very fun. But the wind up Girl in 0:36:48.400 --> 0:36:53.040 question is a essentially a sex Spot character. It goes 0:36:53.280 --> 0:36:55.200 you know that that ends up rebelling, and you have 0:36:55.280 --> 0:36:58.239 sort of a typical narrative with her. But they call 0:36:58.320 --> 0:37:02.319 her a wind up girl because she's she's she's very 0:37:02.320 --> 0:37:05.520 convincing as a humanoid, except that her skin pours are 0:37:05.520 --> 0:37:10.200 too small and she has an intentionally herky jerky movement 0:37:10.280 --> 0:37:13.880 to uh to as she walks around. That they did 0:37:14.080 --> 0:37:17.000 so that she could not be mistaken as a person. 0:37:17.800 --> 0:37:20.200 So that so that apparently, like all the I think 0:37:20.200 --> 0:37:22.359 they were called the new people and some in some 0:37:22.400 --> 0:37:25.640 cases so that the new people could not be mistaken 0:37:25.920 --> 0:37:28.920 for the old people. Whoa. It reminds me of how 0:37:28.960 --> 0:37:32.799 they had to add artificial sounds to electric cars for 0:37:32.840 --> 0:37:36.160 safety purposes, because the cars are too quiet, they can 0:37:36.160 --> 0:37:38.239 really sneak up on you from behind, so they had 0:37:38.280 --> 0:37:43.239 to make them rumble a little bit. Yeah, I think 0:37:43.239 --> 0:37:46.360 that's an apt comparison. So here's a question for you, Robert, 0:37:47.320 --> 0:37:52.000 what is the gold standard of evidence that somebody did something? 0:37:53.160 --> 0:37:56.319 Imagine you're on a jury. I'm the defendant. I've been 0:37:56.360 --> 0:37:59.440 accused of offering a cash bribe to a police officer 0:37:59.560 --> 0:38:02.919 if she'll me borrow her gun for five minutes. She says, 0:38:02.960 --> 0:38:05.640 I did it. I plead not guilty. What what's the 0:38:05.680 --> 0:38:10.160 best evidence to convince you that I really did that. Well, 0:38:10.239 --> 0:38:13.279 it's not human memory, because if we've touched on many 0:38:13.280 --> 0:38:16.080 times before, human memory is fallible and uh, and it's 0:38:16.080 --> 0:38:20.000 a legitimate problem when it comes to to eyewitness testimony. 0:38:20.520 --> 0:38:24.080 But when the eyewitness is a video camera digital camera 0:38:24.600 --> 0:38:27.760 that has a long photographic evidence as well to certain 0:38:27.760 --> 0:38:30.240 degrees like this has been held up as the gold standard. 0:38:30.719 --> 0:38:33.400 I mean assuming the footage is clear enough that the 0:38:33.400 --> 0:38:37.720 individual's face is visible. All of that. Uh Like even 0:38:37.719 --> 0:38:40.600 in our science fiction right, we have so many examples 0:38:40.600 --> 0:38:43.200 of like on Star Trek. Again, there would be scenes 0:38:43.480 --> 0:38:46.480 where the card would command it we zoom in and enhance, 0:38:47.080 --> 0:38:49.520 and it was never questioned that there were any problems 0:38:49.520 --> 0:38:51.839 with the enhancing of the image. It was just something 0:38:51.840 --> 0:38:53.880 that was done. It's like, oh, yeah, the image is enhanced, 0:38:53.880 --> 0:38:56.400 and now we see exactly who the killer is. So 0:38:56.520 --> 0:38:59.360 we should look at this this booming new research field. 0:38:59.440 --> 0:39:01.920 I shouldn't booming. I just mean there are some papers 0:39:01.960 --> 0:39:07.120 on it, okay, called facial reenactment. So this employs some 0:39:07.160 --> 0:39:09.680 of the same techniques that you would see used in 0:39:10.160 --> 0:39:13.520 UH in studios. If if people are doing motion capture 0:39:13.600 --> 0:39:16.280 for c g I characters in movies and video games, 0:39:16.800 --> 0:39:19.919 you have an actor performer who puts on special gear 0:39:20.000 --> 0:39:23.120 and a special environment surrounded by lights and cameras, and 0:39:23.160 --> 0:39:26.640 the performer acts out motions. These motions are captured from 0:39:26.719 --> 0:39:30.839 multiple angles and different lighting conditions, and then they're translated 0:39:30.840 --> 0:39:33.040 by a computer into the motions of a c g 0:39:33.239 --> 0:39:35.400 I character. You could make a c g I me 0:39:35.760 --> 0:39:37.840 that was doing all the same things I did with 0:39:37.880 --> 0:39:41.800 my body. But what if instead of a c g 0:39:41.960 --> 0:39:47.400 I character, you used captured motion to manipulate existing video 0:39:47.560 --> 0:39:50.480 or images of a real person, not a c g 0:39:50.640 --> 0:39:54.759 I character. This technology is already in development today, and 0:39:54.960 --> 0:39:57.440 one example is the research being done under the heading 0:39:57.440 --> 0:39:59.799 as I said, of facial reenactment. There are a couple 0:39:59.840 --> 0:40:03.839 of papers along with accompanying video demonstrations by a group 0:40:03.840 --> 0:40:06.600 of researchers based out of Stanford, out of the Max 0:40:06.640 --> 0:40:11.279 Plank Institute for Informatics and the University of Erlanga, Nuremberg, 0:40:11.880 --> 0:40:14.960 and in their own words quote, we present a method 0:40:15.080 --> 0:40:19.200 for the real time transfer of facial expressions from an 0:40:19.200 --> 0:40:22.359 actor in a source video to an actor in a 0:40:22.440 --> 0:40:26.279 target video, thus enabling the ad hoc control of the 0:40:26.320 --> 0:40:30.600 facial expressions of the target actor. And so if you 0:40:30.600 --> 0:40:32.400 haven't seen video of this, you should look it up. 0:40:32.440 --> 0:40:37.120 Try facial reenactment video. If you have sufficient sample video 0:40:37.160 --> 0:40:40.719 of your target, you can use a regular camera to 0:40:40.800 --> 0:40:45.840 project new facial expressions, including mouth movements which form the 0:40:45.880 --> 0:40:49.560 shapes of words, onto your your target in your video. 0:40:49.880 --> 0:40:52.560 So I could take video of Robert talking if I 0:40:52.600 --> 0:40:55.960 had all this technology, If I could take video of 0:40:56.040 --> 0:41:00.000 Robert talking, and then I could film myself saying how 0:41:00.000 --> 0:41:03.239 Halloween five is the best entry in the Halloween franchise, 0:41:04.120 --> 0:41:07.319 and then map that onto Robert's face to make his 0:41:07.440 --> 0:41:11.080 lips say those words, to make his face move along 0:41:11.160 --> 0:41:13.920 with my face as it's being recorded. And in the 0:41:13.960 --> 0:41:17.120 demonstrations of this it looks nearly photo real. They do 0:41:17.200 --> 0:41:21.240 it with with public figures, making them move their faces around, 0:41:21.400 --> 0:41:24.400 move their lips to say things. In some cases, I 0:41:24.480 --> 0:41:27.520 think the average observer already would not be able to 0:41:27.600 --> 0:41:31.239 tell the difference in this video, and in fact that 0:41:31.360 --> 0:41:33.960 a similar thing appears to be happening with voice, this 0:41:34.040 --> 0:41:36.239 thing you might have read about last year, this thing 0:41:36.280 --> 0:41:39.000 Adobe Voco, where they came out with this announcement that 0:41:39.040 --> 0:41:41.840 Adobe is working on software where you can take a 0:41:41.880 --> 0:41:44.640 twenty minute sample of your audio to learn from. If 0:41:44.640 --> 0:41:47.080 you've got a recording of somebody talking for twenty minutes, 0:41:47.520 --> 0:41:50.080 I can take it make a recording of you saying 0:41:50.120 --> 0:41:53.280 things you never said in your own voice. I wanna. 0:41:53.480 --> 0:41:56.960 I seem to remember that before his death, Roger Ebert 0:41:57.600 --> 0:42:00.239 was involved with the project with some of this technology, 0:42:00.400 --> 0:42:02.040 the idea of being that, of course he had lost 0:42:02.120 --> 0:42:06.440 his his availability to talk due to illness, But there 0:42:06.520 --> 0:42:11.359 was so much Roger Ebert audio out there from all 0:42:11.400 --> 0:42:13.040 of his years as a as a as a film 0:42:13.080 --> 0:42:16.360 critic and a TV personality that they had this they had, 0:42:16.400 --> 0:42:19.800 they had everything they needed to enable him to say 0:42:19.920 --> 0:42:23.360 anything new. Yeah, and that that explores the totally non 0:42:23.360 --> 0:42:26.040 nefarious aspect of this. I mean, I I don't think 0:42:26.080 --> 0:42:28.719 people who are pursuing these lines of research are just 0:42:28.800 --> 0:42:31.160 trying to create a world where we can fake video 0:42:31.239 --> 0:42:34.279 evidence of things. But wouldn't that be a wonderful thing 0:42:34.280 --> 0:42:37.239 for somebody who lost their capacity for speech they had 0:42:37.280 --> 0:42:39.920 recordings of their voice to be able to create a 0:42:40.000 --> 0:42:42.719 text to speech voice box that could speak with their 0:42:42.760 --> 0:42:48.840 own voice. That's amazing, That's kind of beautiful. But there 0:42:48.880 --> 0:42:51.839 are these other ways of looking at this. And the 0:42:51.840 --> 0:42:54.400 authors also that they point out that, you know, they 0:42:54.560 --> 0:42:56.920 they in their own defense, they're like, look, we're not 0:42:56.960 --> 0:42:59.240 trying to create a world where people can fake video. 0:42:59.560 --> 0:43:02.560 We also try to show how you can detect altered video. 0:43:03.160 --> 0:43:05.880 So that's another thing they're trying to explore and make public. 0:43:05.920 --> 0:43:09.520 Because you know, they're not the only people pursuing this research. Obviously, 0:43:09.560 --> 0:43:13.000 people all over the place are doing stuff like this, 0:43:13.080 --> 0:43:15.279 and stuff like this has been in the use, has 0:43:15.280 --> 0:43:18.000 been in use in the movie industry for years. Yeah, 0:43:18.160 --> 0:43:19.719 I mean it makes me think that what you would 0:43:19.760 --> 0:43:22.520 need to go for is the equivalent of a water mark. 0:43:22.840 --> 0:43:25.480 I don't know exactly what that watermark would be and 0:43:25.520 --> 0:43:28.960 what form it would take, but it does make me think, well, 0:43:29.040 --> 0:43:30.840 we're going to reach a point where any kind of 0:43:30.840 --> 0:43:35.400 footage has to have the watermark of authenticity otherwise doubt 0:43:35.440 --> 0:43:40.040 will be cast upon it. Yeah. I'm concerned about the 0:43:40.080 --> 0:43:43.080 idea of living in a world where you can make 0:43:43.280 --> 0:43:48.560 very convincing looking fake video evidence of things, and not 0:43:48.680 --> 0:43:51.960 just because of the specific example of somebody can make 0:43:52.000 --> 0:43:54.359 a video of me or somebody I like, you know, 0:43:54.440 --> 0:43:57.200 saying or doing something that they didn't do. It's not 0:43:57.239 --> 0:44:00.759 just the specifics. It's the general degray eating of our 0:44:00.880 --> 0:44:04.440 trust in the ability to look at things and know 0:44:04.560 --> 0:44:06.800 that they're true. Yeah. I mean, we look at the 0:44:07.640 --> 0:44:09.680 current news cycle and there's been a lot of discussion 0:44:09.680 --> 0:44:15.040 about the reliability of information, of so called fake news. Uh, 0:44:15.120 --> 0:44:17.840 the idea that you read, do you reach this point 0:44:17.840 --> 0:44:20.520 where nobody knows what to trust anymore? End of just 0:44:20.520 --> 0:44:23.399 trusting nothing and you already know this because you don't 0:44:23.400 --> 0:44:26.839 trust any weird looking picture of somebody you see, right, 0:44:26.840 --> 0:44:29.120 because you know what can be done with photoshop. We're 0:44:29.160 --> 0:44:33.239 already there with still images. Photoshop is is basically in 0:44:33.239 --> 0:44:38.640 many contexts become a verb for for the the distortion 0:44:38.680 --> 0:44:42.520 of truth. But what what if we had photoshop to 0:44:42.719 --> 0:44:46.400 undermine moving video evidence to the same extent that photoshop 0:44:46.480 --> 0:44:49.840 has undermined still images. I mean, this is what a 0:44:49.840 --> 0:44:52.360 lot of these viral fake news stories are based on, 0:44:52.480 --> 0:44:55.400 is a photoshopped image. If you go to Snopes or 0:44:55.440 --> 0:44:57.160 something and you look at what a lot a lot 0:44:57.200 --> 0:44:59.719 of what they're debunking is, it's just an image that 0:45:00.000 --> 0:45:03.279 aims to be real of somebody doing something, and they 0:45:03.280 --> 0:45:06.440 have to track down where it came from. Yeah, it's 0:45:06.480 --> 0:45:10.200 difficult to imagine what that what would slash will be 0:45:10.320 --> 0:45:13.439 like when we reach that point to where there's where 0:45:13.480 --> 0:45:16.160 where video, even digital footage is no longer the gold 0:45:16.200 --> 0:45:18.640 standard it was. I mean, I think it's actually a 0:45:18.760 --> 0:45:24.400 very important project to maintain a version of the Uncanny Valley, 0:45:24.920 --> 0:45:29.520 to to help people find a way to all to 0:45:29.600 --> 0:45:33.240 separate real video evidence from fake video evidence of things, 0:45:33.320 --> 0:45:35.520 to understand that there are things you can look for 0:45:35.640 --> 0:45:40.680 that separate real moving imagery from falsified or synthetic moving imagery. Now, 0:45:40.719 --> 0:45:43.960 one way you can approach this is to do what 0:45:44.120 --> 0:45:46.279 the authors of this research I was talking about do 0:45:46.440 --> 0:45:49.399 is they say, look, here are things that are are 0:45:49.600 --> 0:45:53.920 signs that video has been manipulated. And that's one thing. 0:45:53.960 --> 0:45:55.759 And maybe there will be a lot of experts on 0:45:55.760 --> 0:45:57.600 this in the future, like it could be a whole 0:45:57.640 --> 0:46:01.000 field of people who are just there to have expertise 0:46:01.040 --> 0:46:07.120 in authenticating purportedly real video of you doing things or not. Uh. 0:46:07.280 --> 0:46:11.080 Then on the other hand, we could hope that there 0:46:11.239 --> 0:46:16.040 is in fact an adaptive response in our discernment in general. 0:46:16.160 --> 0:46:18.160 And this is where I want to go to the 0:46:18.200 --> 0:46:22.920 concept of an of the uncanny wall. So a paper 0:46:22.960 --> 0:46:26.239 in the International Journal of Arts and Technology, authored by 0:46:26.280 --> 0:46:31.080 Tinwell Grimshawn Williams offers this interesting counter hypothesis to the 0:46:31.200 --> 0:46:34.520 uncanny Valley. And I want to emphasize again we've sort 0:46:34.560 --> 0:46:37.440 of shifted back and forth between the uncanny valley itself 0:46:37.680 --> 0:46:41.319 in terms of what causes negative affinity and then over 0:46:41.400 --> 0:46:44.839 slightly just to the issue of nearing photo realism or not. 0:46:45.000 --> 0:46:48.120 So keep in mind the difference in those subjects, but 0:46:49.800 --> 0:46:53.560 they proposed this idea of the uncanny wall. To put 0:46:53.560 --> 0:46:59.120 it succinctly, they proposed that quote, increasing technological sophistication in 0:46:59.200 --> 0:47:03.760 the creation of realism for humanlike virtual characters is matched 0:47:04.360 --> 0:47:08.920 by increasing technological discernment on the part of the viewer. 0:47:09.680 --> 0:47:14.200 In other words, as humanoid characters become more real, our 0:47:14.280 --> 0:47:18.680 standards for what looks realistic go up. And I do 0:47:18.760 --> 0:47:22.680 think they're just anecdotally personally that I think there's some 0:47:22.760 --> 0:47:24.799 support for this, and I kind of hope this is 0:47:24.840 --> 0:47:28.400 true so we can avoid this world where all video 0:47:28.440 --> 0:47:31.839 evidence is in question. Because I immediately think, right now, 0:47:31.880 --> 0:47:35.920 I've got I've got antenna for photoshopped images, unlike I 0:47:36.000 --> 0:47:39.040 had ten years ago. I think stuff that would look 0:47:39.160 --> 0:47:42.799 obviously photoshop to me today would have fooled me ten 0:47:42.880 --> 0:47:46.919 years ago. I think I've simply adapted. And another thing 0:47:47.120 --> 0:47:49.120 is it makes me flashback to the early days of 0:47:49.160 --> 0:47:52.000 c G I in movies and like the thirty two 0:47:52.080 --> 0:47:56.439 bit video game era, or think about like PlayStation one 0:47:56.640 --> 0:48:00.160 games now. Back then, I remember looking at games for 0:48:00.160 --> 0:48:04.480 the original PlayStation and thinking, Wow, that looks so real, 0:48:05.560 --> 0:48:07.680 and you try and play him. Now and it's painful. Yeah, 0:48:07.719 --> 0:48:12.680 you can't blocky polygons. People's faces have all these sharp corners. 0:48:12.840 --> 0:48:16.800 It's it's hilarious. Uh, there was some kind of geometrical 0:48:17.000 --> 0:48:20.239 nightmare world everything was taking place in where there's just 0:48:20.960 --> 0:48:24.120 lots of sharp angles. But at the time it looked 0:48:24.239 --> 0:48:26.719 so real to me. And another fun trick is go 0:48:26.800 --> 0:48:29.719 back and read movie reviews for movies with bad c 0:48:29.840 --> 0:48:34.239 g I from the nineties. Professional movie reviewers at the time, 0:48:34.280 --> 0:48:38.040 we're often praising the effects. One example is like the 0:48:38.120 --> 0:48:41.840 Mortal Kombat movie, the original Mortal Kombat movie. Remember this. 0:48:42.160 --> 0:48:44.680 You can find reviews at the time where people are like, well, 0:48:44.719 --> 0:48:49.160 the stories then an immature but dazzling special effects. Now 0:48:49.239 --> 0:48:53.680 even mentioning those special effects conjures a kind of delirious hilarity. 0:48:53.760 --> 0:48:56.040 You just start laughing when you think about the c 0:48:56.200 --> 0:48:59.360 g I in Mortal Kombat. But that being said, the 0:48:59.360 --> 0:49:02.239 gore pup it was above approach just looks really good. 0:49:03.040 --> 0:49:05.920 It did have kind of nasty, beady eyes, kind of 0:49:05.960 --> 0:49:09.040 crypt keep keeper asked, Yeah, oh yeah, it was like 0:49:09.080 --> 0:49:12.080 a very it was like a bloated, buffed up crypt 0:49:12.160 --> 0:49:16.600 kicker at the buff But yeah, it's so ugly. It 0:49:16.680 --> 0:49:20.120 provokes uncontrollable laughter, But at the time people were like, 0:49:20.200 --> 0:49:24.000 dazzling looks amazing. Uh. So it makes me think that 0:49:24.120 --> 0:49:27.880 I hope that there is something to this, this idea 0:49:27.960 --> 0:49:31.800 that these authors have that as things continue to chase 0:49:31.880 --> 0:49:35.520 photo realism, as synthetic imagery of humans gets closer and 0:49:35.560 --> 0:49:38.239 closer to the real thing, we just get more and 0:49:38.320 --> 0:49:43.279 more attuned to the minute problems with him with them 0:49:43.320 --> 0:49:47.480 and never really get fully fooled. Well. I have two 0:49:47.600 --> 0:49:50.800 thoughts here, One on the whole watermark thing. Maybe it 0:49:50.800 --> 0:49:54.040 would be some sort of a Bitcoin type of authentication 0:49:54.160 --> 0:49:58.439 system that would be in place. The other is maybe 0:49:58.440 --> 0:50:00.600 you have to go beyond the real may be safe 0:50:00.600 --> 0:50:02.600 for a head of state to appear in a video 0:50:02.719 --> 0:50:06.080 and it'd be authentic. They have to appear as an 0:50:06.080 --> 0:50:10.040 as a computer generated avatar so advanced that it is 0:50:10.200 --> 0:50:13.600 that it is beyond the ability of any like non 0:50:13.680 --> 0:50:18.279 state production or company to create like something I'm I'm 0:50:18.920 --> 0:50:21.920 something that at this point in what would be their past, 0:50:22.040 --> 0:50:24.200 I cannot even conceive of like a I don't know, 0:50:24.280 --> 0:50:29.480 like a five dimensional unfolding c G I god being. Uh, 0:50:29.800 --> 0:50:31.960 because how he's gonna fake that? You can fake a person, 0:50:32.000 --> 0:50:35.840 but good luck faking the fifth dimension hole of Avatar 0:50:35.960 --> 0:50:40.680 of Vishnu. Okay, like like the algorithms for faking a 0:50:40.760 --> 0:50:45.240 person with a lot of photo and an audio visual 0:50:45.360 --> 0:50:47.640 cues to sample from. If you've got a lot of 0:50:47.680 --> 0:50:50.480 footage out there, you could simulate that person. But you 0:50:50.520 --> 0:50:54.759 couldn't simulate this brand new creation that is it requires 0:50:54.760 --> 0:50:57.279 you know, supercomputers to generate. And yeah, and it was 0:50:57.280 --> 0:51:00.080 probably built from the bottom up with completely alien is 0:51:00.160 --> 0:51:02.960 theology and movements. Just a thought. That's kind of a 0:51:02.960 --> 0:51:05.120 crazy idea, but I like it. Well that's what I'm 0:51:05.160 --> 0:51:07.880 here for with the crazy ideas. I don't know. Have 0:51:07.960 --> 0:51:10.319 you got anything else, Robert, any anything else you can 0:51:10.320 --> 0:51:13.240 think of to save us from the future of uh 0:51:13.360 --> 0:51:16.239 synthetic human imagery? Oh you know, I could sit around 0:51:16.239 --> 0:51:17.759 here all day and talk about the c g I 0:51:18.000 --> 0:51:21.600 monsters and UH and Uncanny Valley and films and video 0:51:21.640 --> 0:51:23.719 games and whatnot, But you know, we'll have to what 0:51:23.920 --> 0:51:25.719 to save that for another time, maybe save some of 0:51:25.719 --> 0:51:28.920 it for trailer talk, which will If you're listening to 0:51:28.960 --> 0:51:31.279 this on a Thursday, hopefully you can tune in tomorrow 0:51:31.560 --> 0:51:35.359 around eleven am on our Facebook page, like our Facebook page. 0:51:35.360 --> 0:51:37.520 While you're at it, follow us there, but tune into 0:51:37.600 --> 0:51:40.759 a little discussion of trailers that are associated with the 0:51:40.840 --> 0:51:44.319 Uncanny Valley. Oh yeah, and in the meantime, heading over 0:51:44.320 --> 0:51:46.320 to stuff to Blow your Mind dot com. That's the mothership. 0:51:46.360 --> 0:51:50.080 That's we will find all of our blog posts, our podcast, 0:51:50.200 --> 0:51:52.759 our videos, and links out to our various social media 0:51:52.760 --> 0:51:55.360 accounts such that Facebook page or the Twitter page, the 0:51:55.400 --> 0:51:58.239 Instagram account, you name it. And if you want to 0:51:58.239 --> 0:52:00.440 get in touch with us directly as all ways, you 0:52:00.480 --> 0:52:03.600 can email us at blow the Mind at how stuff 0:52:03.600 --> 0:52:15.600 works dot com for more on this and thousands of 0:52:15.640 --> 0:52:40.680 other topics. Is it how stuff works dot com.