WEBVTT - Selects: Is The Uncanny Valley Real? 0:00:00.880 --> 0:00:04.120 Hey, everybody. In this week's Select, it's Josh. By the way. 0:00:04.320 --> 0:00:07.480 We take a look at the Uncanny Valley, a theoretical 0:00:07.520 --> 0:00:10.560 phenomenon where robots give us the creeps as they get 0:00:10.560 --> 0:00:13.680 closer to looking human but still don't seem quite right, 0:00:14.240 --> 0:00:18.000 and we explore academic research into the creeps, which is 0:00:18.079 --> 0:00:26.279 pretty cool. So enjoy this episode. Welcome to Stuff You 0:00:26.280 --> 0:00:35.760 Should Know, a production of iHeartRadio. Hey, and welcome to 0:00:35.800 --> 0:00:39.000 the podcast. I'm Josh Clark, and there's Charles w Chuck Bryant, 0:00:39.000 --> 0:00:42.880 and there's Jerry over there doing the robot, which means 0:00:42.920 --> 0:00:48.560 this is stuff you should Know robot style. I knew 0:00:48.560 --> 0:00:50.680 i'd get a laugh out you sooner or later. Did 0:00:50.720 --> 0:00:53.040 you do the robot? Can you do the robot? I 0:00:53.040 --> 0:00:55.960 think I've seen you do pretty bad robot. Sure, I 0:00:56.040 --> 0:00:58.920 don't know about pretty bad robot. I can do a 0:00:58.960 --> 0:01:02.560 pretty great robot, that's what you've seen. I can't do 0:01:02.560 --> 0:01:06.440 any of that stuff. Yeah, I can't really either. And then, really, 0:01:07.600 --> 0:01:10.480 if your claim to fame as like a really great 0:01:10.600 --> 0:01:14.200 robot dance, I don't know, maybe take up some other 0:01:14.280 --> 0:01:17.720 hobbies as well. Kind of round that out. You don't 0:01:17.720 --> 0:01:20.920 want that to be the only thing you're good at, right, 0:01:21.000 --> 0:01:24.080 because if you list that on a dating site, you 0:01:24.160 --> 0:01:29.720 might turn ladies off. Yeah, according to e Harmony. Yeah 0:01:29.720 --> 0:01:33.440 that's foreshadowing. I love that one, don't you. Yeah, it's 0:01:33.440 --> 0:01:36.800 some issues with that whole Oh yeah, bit, Yeah, we'll 0:01:36.800 --> 0:01:39.360 get to that. This is tad all right. Well, let's 0:01:39.360 --> 0:01:41.880 start at the at the beginning, almost the beginning, Chuck, 0:01:41.959 --> 0:01:44.959 Let's go back to nineteen seventy, which was the beginning 0:01:45.000 --> 0:01:48.040 of the greatest decade in the history of humanity. Yeah, 0:01:48.200 --> 0:01:50.400 neither one of us are born yet. I can finally 0:01:50.440 --> 0:01:53.560 even say that I'm still not even born. It must 0:01:53.600 --> 0:01:59.120 feel good. Yeah, okay, well, welcome to the club. I 0:01:59.200 --> 0:02:02.640 guess thanks. In nineteen seventy, we're not just going just 0:02:02.720 --> 0:02:05.720 anywhere in nineteen seven, We're going to Japan in nineteen seventy. 0:02:05.960 --> 0:02:08.920 I bet Japan was pretty cool in the seventies. Yeah. 0:02:08.960 --> 0:02:12.959 A lot of bell bottoms, a lot of ninja running around. Still. Yeah, 0:02:14.280 --> 0:02:19.680 there were calculators being wheelded all over the place. Probably 0:02:19.840 --> 0:02:22.600 it was a good time, good time for Japan, right, 0:02:23.919 --> 0:02:26.640 And one of the one of the things that was 0:02:26.720 --> 0:02:29.600 going on in nineteen seventy I could not for the 0:02:29.639 --> 0:02:33.799 life of me, find what issue of this journal it 0:02:33.919 --> 0:02:36.080 came out in what month? But at some point in 0:02:36.120 --> 0:02:41.520 nineteen seventy there was an obscure journal, a Japanese academic 0:02:41.600 --> 0:02:45.320 journal called Energy, and at some point during that year 0:02:45.400 --> 0:02:50.760 it published a article by a Japanese roboticist. And his 0:02:50.919 --> 0:02:55.360 name is Massa Hiro Morri. Nice, thank you, I have 0:02:55.400 --> 0:02:59.840 a lot of practice. And Massa Hiro Amori published this 0:03:00.080 --> 0:03:03.280 article and he named it Bukimi no tani gen show 0:03:03.960 --> 0:03:06.320 is actually the full name of the whole thing. And 0:03:06.400 --> 0:03:11.000 as we'll see, it's kind of difficult to translate into English, right, 0:03:11.960 --> 0:03:15.480 and it took many, many years after he wrote this 0:03:15.639 --> 0:03:18.840 article for it to be translated into English for anybody 0:03:18.840 --> 0:03:22.400 even tried to attempt it. So basically more was this 0:03:23.240 --> 0:03:26.840 roboticist and he wrote this essay, and at the time 0:03:26.840 --> 0:03:29.240 he just put it out there and went back to work, 0:03:29.680 --> 0:03:33.080 started teaching more and more roboticists. The whole new generation 0:03:33.120 --> 0:03:36.320 of roboticists learned under him, and his work just kind 0:03:36.360 --> 0:03:41.400 of sat there unobserved that article, i should say. And 0:03:41.440 --> 0:03:45.120 then in two thousand and five a rough translation of 0:03:45.160 --> 0:03:48.920 it was leaked out wasn't intended for publication, and the 0:03:48.960 --> 0:03:54.600 world entirely changed, right because massa hiro Omori had in 0:03:54.760 --> 0:03:57.680 his article put his finger on something that no one 0:03:57.760 --> 0:04:01.200 had before. In his capacities are roboticists and a human, 0:04:01.680 --> 0:04:07.800 and that was what we call today the uncanny Valley. Yeah, 0:04:07.840 --> 0:04:12.840 so that's um the idea that you're making a robot, 0:04:13.040 --> 0:04:15.240 and we'll see this apply to more than just robots, 0:04:15.720 --> 0:04:17.560 but in his case, you're making a robot and you 0:04:17.600 --> 0:04:21.720 want to make it look like a person, which I 0:04:21.760 --> 0:04:24.479 guess not all roboticists do. Some of them like the 0:04:24.520 --> 0:04:28.239 clunky jets and style robots like Rosy, But I guess 0:04:28.240 --> 0:04:31.159 if you're morey, you're you're on the path to designing 0:04:31.560 --> 0:04:36.479 lifelike robots. And the closer you get to that lifelike look, 0:04:36.720 --> 0:04:39.440 everything's going great, everything's going great, people are like, this 0:04:39.480 --> 0:04:41.640 is so cool, this is so cool. Yeah. Then all 0:04:41.640 --> 0:04:45.400 of a sudden, people go oh, like, right as it 0:04:45.400 --> 0:04:49.120 approaches it's most or basically, when it reaches it's most 0:04:49.160 --> 0:04:54.440 lifelike capacity, that this whoever's making it can conjure people 0:04:54.440 --> 0:04:59.440 are repulsed by it. Yeah, which is something that most 0:04:59.440 --> 0:05:02.560 people who whoever hear of the Uncanny Valley or like, yeah, 0:05:02.600 --> 0:05:05.039 you know, that's I've noticed that that's happened to me 0:05:05.120 --> 0:05:08.120 before too. But the thing is, Chuck, it doesn't it 0:05:08.160 --> 0:05:10.760 doesn't actually make sense, right, Like we know a robot 0:05:10.880 --> 0:05:14.800 is a robot. Yeah, so you know, maybe you could 0:05:14.800 --> 0:05:16.760 be afraid that's gonna like pick you up and break 0:05:16.760 --> 0:05:19.600 you in two or something like a cartoon. But that's 0:05:19.640 --> 0:05:21.960 different than being creeped out by Like why would we 0:05:22.040 --> 0:05:23.880 be creeped out by a robot? And this is what 0:05:23.960 --> 0:05:27.920 Morey put his finger on, was there's something to this 0:05:28.160 --> 0:05:32.360 and it doesn't make sense. And he didn't. It wasn't 0:05:32.400 --> 0:05:35.960 even just um this article that he wrote. He created 0:05:36.000 --> 0:05:39.920 a graph as well that's become quite famous that um 0:05:40.880 --> 0:05:43.840 really kind of gets the point across more than anything else. Yeah, 0:05:43.839 --> 0:05:46.240 And he wasn't even the first person to go over 0:05:46.320 --> 0:05:48.640 this and to put a put some thought to it. 0:05:49.440 --> 0:05:52.440 Freud of course, because he'd like to think about everything. 0:05:53.320 --> 0:05:56.080 He thought about a little bit. And before Freud, there 0:05:56.160 --> 0:06:01.599 was a Geman name anst Yinched. Oh nice, I did 0:06:01.600 --> 0:06:04.080 not realize that's how his last name should be pronounced. Yinch. 0:06:04.600 --> 0:06:06.040 That's good stuff. I think I put a T on 0:06:06.080 --> 0:06:09.320 the end, but the teas in the middle. Yinch. Yeah, 0:06:09.400 --> 0:06:12.960 I think that's right. I've been saying Ginch, Well, we'll 0:06:12.960 --> 0:06:14.680 have to look that up. Then I think you're no. 0:06:14.839 --> 0:06:16.240 I think of the two of us, you you get 0:06:16.240 --> 0:06:20.480 the German down. And he had a little term called 0:06:20.560 --> 0:06:24.599 Umheimlich that he called it so like you know, different 0:06:24.640 --> 0:06:28.400 languages had different names for it, and you'd go back 0:06:28.440 --> 0:06:30.919 in time, all the way back to like the seventeenth century, 0:06:31.360 --> 0:06:34.280 and people were and I guess you know, robots didn't 0:06:34.279 --> 0:06:36.839 look super lifelike back then, but whatever their version of 0:06:36.920 --> 0:06:41.280 lifelike was in the sixteen hundreds, people were like, uh, 0:06:41.720 --> 0:06:44.160 I don't like that. Why is it looking at me? Yeah, 0:06:44.200 --> 0:06:47.000 it's got a quill and it's writing things. But like 0:06:47.040 --> 0:06:51.440 you said, Moriy made this graph because he was a roboticist, 0:06:51.480 --> 0:06:53.200 and he thought, you know, let's look at this on 0:06:53.240 --> 0:06:57.039 a plotted out so we can stare at it. And 0:06:57.279 --> 0:07:01.360 on the x axis he had likeness, then on the 0:07:01.480 --> 0:07:04.160 y axis he had affinity, like whether or not you 0:07:04.240 --> 0:07:07.200 like the way this thing looks? Right? And just as 0:07:07.400 --> 0:07:09.960 we're talking about um the graph went up and up 0:07:10.360 --> 0:07:13.480 as things got more lifelike and people like the way 0:07:13.480 --> 0:07:15.840 it looked. And then at a certain point there is 0:07:15.840 --> 0:07:20.480 that valley, there's a big dip that really just kind 0:07:20.480 --> 0:07:24.800 of says it all right, And again, this all makes 0:07:24.800 --> 0:07:28.120 sense intuitively, but as we'll see, that's it's been very 0:07:28.160 --> 0:07:31.080 difficult to prove. And one of the reasons why it's 0:07:31.240 --> 0:07:36.600 confounded research thus far is because we're not even one 0:07:36.680 --> 0:07:40.120 hundred percent sure what Morey meant by some of the 0:07:40.200 --> 0:07:42.680 words he chose, at least as far as translating them 0:07:42.760 --> 0:07:48.600 to English. Right. For example, bukimi Right, yeah, it was 0:07:48.640 --> 0:07:53.280 translated in two thousand and five as uncanny, but um 0:07:53.400 --> 0:07:56.960 again that that original translation was not intended for publication, 0:07:57.360 --> 0:08:01.360 but it leaked out, and so uncanny Anny Valley became, 0:08:01.480 --> 0:08:04.680 you know, the way we all think of it here 0:08:04.680 --> 0:08:09.920 in the West. But boukimi more closely resembles something like eerie. 0:08:10.360 --> 0:08:14.320 Like I've seen it explain that a word like boukimi 0:08:15.000 --> 0:08:19.480 means more than Uncanny's just weird or remarkable or noteworthy. 0:08:19.520 --> 0:08:22.520 It's not necessarily something that gives you the creeps. Boukimi 0:08:22.600 --> 0:08:28.240 is something that gives you the creeps, like Steve Bukimi's exactly. 0:08:29.520 --> 0:08:34.160 So boukimi probably more should be should have been translated 0:08:34.200 --> 0:08:37.960 the Erie Valley, but by the time an actual official 0:08:38.040 --> 0:08:42.719 translation that m maury signed off on came out in 0:08:42.760 --> 0:08:45.640 twenty twelve. The cat was out of the bag. Everybody 0:08:45.720 --> 0:08:47.880 knew of it as the Uncanny Valley, and there's no 0:08:47.920 --> 0:08:50.120 way anybody who's going to come back and be like, no, no, no, 0:08:50.160 --> 0:08:54.200 everybody stop calling it that. It's now the Eerie Valley, okay, 0:08:55.520 --> 0:08:57.839 right right, And it may be one of those things 0:08:57.880 --> 0:09:00.200 where we're so used to Uncanny Valley now that it's 0:09:00.200 --> 0:09:02.480 hard to imagine Eerie Valley. But right, I think that 0:09:02.559 --> 0:09:05.520 was the issue. Yeah, Like, nobody's gonna go along with that. 0:09:06.640 --> 0:09:09.080 So this graph, like I said, it starts off on 0:09:09.120 --> 0:09:11.400 that left hand side, and this is where you have 0:09:11.480 --> 0:09:16.280 things that are super robotic, like you know, a packaging 0:09:16.520 --> 0:09:21.880 robot in a factory, right that you know, apparently most 0:09:21.920 --> 0:09:24.280 people don't have fondness for I do because I love 0:09:24.559 --> 0:09:29.320 mechanical processes, right right, Okay, So there's there's part of 0:09:29.320 --> 0:09:33.160 the problem. It's like that's not necessarily the kind of 0:09:33.240 --> 0:09:38.520 feeling that Massa Hiro Morty was talking about. He was like, yeah, yeah, 0:09:38.559 --> 0:09:42.520 you're interested in robotics and robotic arms and the industrial processes, 0:09:42.559 --> 0:09:45.559 and you love watching how it's made. Right, what he 0:09:45.640 --> 0:09:49.240 was talking about was more like how it resembles a 0:09:49.360 --> 0:09:52.360 human and then how it makes you feel in relation 0:09:52.440 --> 0:09:55.840 to its resemblance of a human. Right. Well, in that case, 0:09:55.880 --> 0:09:57.720 it makes me feel nothing because it doesn't look at 0:09:57.720 --> 0:09:59.560 all like a human, right, Okay, so that would be 0:09:59.600 --> 0:10:02.160 at about the origin of the graph. It has no 0:10:02.240 --> 0:10:05.800 resemblance to a human really, and it's not eliciting any 0:10:05.880 --> 0:10:09.640 real affinity in you at all as far as it 0:10:09.720 --> 0:10:12.760 looking like a human, right, But lots of affinity as 0:10:12.760 --> 0:10:16.640 a thing that's just that's called props. So you go 0:10:16.679 --> 0:10:18.720 a little bit further on the graph and then you 0:10:18.800 --> 0:10:23.160 have things like um little stuffed animals and I No. 0:10:23.320 --> 0:10:26.640 C three Po is a common one that's mentioned because 0:10:27.120 --> 0:10:30.480 C three Pou, you know, is built to look like 0:10:30.520 --> 0:10:33.000 a human. He does a great robot kind of it, 0:10:33.120 --> 0:10:36.240 talks like a human and acts like a human. But 0:10:36.320 --> 0:10:38.080 when it comes to that face, and as we'll see, 0:10:38.120 --> 0:10:39.480 the face is kind of the key to all this. 0:10:40.679 --> 0:10:43.120 For the most part, C three PO looks nothing like 0:10:43.160 --> 0:10:45.840 a human in the face. So everything's still good and 0:10:45.920 --> 0:10:48.800 people love C three PO. Right, So if you're looking 0:10:48.840 --> 0:10:51.640 at the graph, C three PO is going up in 0:10:51.760 --> 0:10:54.200 human likeness because he kind of you know, he's got 0:10:54.240 --> 0:10:59.560 some some commonality there, and we're feeling affinity for him 0:11:00.200 --> 0:11:04.080 on that human likeness. So it's he's going up. Yes, Okay, 0:11:04.840 --> 0:11:07.840 we're going everything's going pretty well so far, right, Chuck, 0:11:08.400 --> 0:11:10.959 that's right. Okay, So then we're gonna start hitting some 0:11:11.120 --> 0:11:14.400 areas where things start looking a little more human, a 0:11:14.440 --> 0:11:16.680 lot more human, I would say than c three po 0:11:16.800 --> 0:11:21.800 like say, the characters in Moanna or Frozen picks our 0:11:21.920 --> 0:11:25.560 characters that kind of thing where they look like they're 0:11:25.600 --> 0:11:28.480 supposed to be human, like they're based on human, but 0:11:28.520 --> 0:11:31.880 they have very exaggerated features that you would never confuse 0:11:31.960 --> 0:11:35.360 at first glance for an actual human. Right, So they 0:11:35.360 --> 0:11:39.480 have like big guys, small noses, things that make them cute, right, 0:11:40.200 --> 0:11:43.280 And so our affinity for them is going up as 0:11:43.320 --> 0:11:46.080 the human likeness is going up. Again, things are going 0:11:46.120 --> 0:11:49.240 really well so far, that's right, because in Mohanna and 0:11:49.280 --> 0:11:51.640 Frozen they look a little bit more like people, and 0:11:51.679 --> 0:11:53.760 we like them a lot more for that reason. Right. 0:11:54.000 --> 0:11:57.960 And then, like you said earlier, out of nowhere, the 0:11:58.080 --> 0:12:01.040 whole thing, this line that's just been going up very pleasantly, 0:12:01.840 --> 0:12:06.360 and a nice little slope just drops downward. Right, and 0:12:06.360 --> 0:12:10.000 it doesn't drop just downward, it goes actually below the 0:12:10.280 --> 0:12:14.600 X axis into negative territory. And now this is the 0:12:14.720 --> 0:12:18.600 uncanny valley, that's right, And that's why it has that name, 0:12:18.679 --> 0:12:21.520 because it's a valley, right, and this is where those 0:12:21.559 --> 0:12:28.600 things like really really lifelike androids live, or corpses live, 0:12:29.160 --> 0:12:34.800 or zombies live. Because Maury he had the idea that 0:12:35.080 --> 0:12:39.880 if something's moving, is even creepier than something similar to 0:12:39.920 --> 0:12:42.720 it that's not moving. So he actually created two lines 0:12:42.760 --> 0:12:45.840 on this graph, one for things that are animate and 0:12:45.840 --> 0:12:48.440 one for things that are inanimate. So if you look 0:12:48.480 --> 0:12:52.200 at this Uncanny Valley, on the inanimate line, the non 0:12:52.280 --> 0:12:55.480 moving line, you've got corpses are at the bottom of it. 0:12:56.040 --> 0:12:58.960 But if you look at the animate line, it's even 0:12:59.280 --> 0:13:02.120 it dips even further below than the inanimate line, and 0:13:02.240 --> 0:13:04.840 at the bottom of those are zombies. So dead people 0:13:05.240 --> 0:13:09.199 up and moving around and saying brains is as creepy 0:13:09.200 --> 0:13:12.080 as it gets as far as this graph is concerned. Yeah, 0:13:12.120 --> 0:13:15.360 and Maria wasn't the only one that earns yinch that 0:13:15.440 --> 0:13:19.600 we talked about the German psychiatrists. He also talked about 0:13:19.640 --> 0:13:22.120 the fact that if you are looking at something that 0:13:22.120 --> 0:13:25.040 should not be moving, and it moves. I mean, I 0:13:25.040 --> 0:13:27.920 think we can all agree that a baby doll that 0:13:28.040 --> 0:13:31.080 suddenly turns its head and looks at you, right, probably 0:13:31.080 --> 0:13:34.120 one of the creepier things you can witness, right, you know, Yeah, 0:13:34.160 --> 0:13:36.600 it's about as creepy as it gets. Or have you 0:13:36.640 --> 0:13:41.920 ever been to an open casket funeral? A few? I'm 0:13:42.320 --> 0:13:46.800 not a fan at all. No, it is it makes sense. 0:13:47.000 --> 0:13:49.800 You know, we've really kind of closed or put a 0:13:49.840 --> 0:13:52.400 lot of space in between us and death, way more 0:13:52.400 --> 0:13:54.679 than we used to have in like the nineteenth century. Oh, 0:13:54.800 --> 0:13:57.160 they either would sit up with the dead, sure, right, 0:13:57.200 --> 0:13:59.160 so this seems to be like kind of a holdover 0:13:59.280 --> 0:14:02.760 from that. But if you've ever been to an open 0:14:02.800 --> 0:14:06.680 casket funeral and have just stared at the corpse long enough, 0:14:06.720 --> 0:14:09.720 like maybe it's arm or it's fingers or something, your 0:14:09.760 --> 0:14:12.720 brain is so anticipating that they it's about to start 0:14:12.760 --> 0:14:15.680 moving that sometimes you can creep yourself out and make 0:14:16.040 --> 0:14:18.720 yourself think you did actually see it move. You'll also 0:14:18.760 --> 0:14:21.000 be asked to leave the funeral. Well you shouldn't be 0:14:21.040 --> 0:14:24.040 like giving a commentary about this out loud, but you can, 0:14:24.240 --> 0:14:26.240 you know, you can do it to pass the time 0:14:27.200 --> 0:14:30.840 in the funeral if you're looking to kill some time. 0:14:31.840 --> 0:14:34.640 So here's the thing with all this. We know this 0:14:34.680 --> 0:14:38.080 happens because everyone kind of has this feeling, but no 0:14:38.120 --> 0:14:40.120 one and all this research has been done, and no 0:14:40.160 --> 0:14:44.640 one is exactly sure why this happens. So, uh, Maury's 0:14:44.760 --> 0:14:47.920 essay and especially once it was translated, a lot of 0:14:47.960 --> 0:14:53.320 research started happening in this area. And it's problematic though, 0:14:53.360 --> 0:14:57.800 because there are a few different problems. One is, well, 0:14:57.840 --> 0:15:01.560 it's it's subjective. This dependent variable, whether you have an 0:15:01.560 --> 0:15:05.000 affinity for something is very subjective, so it's hard to 0:15:05.040 --> 0:15:08.840 kind of nail that down scientifically, right, all right, So 0:15:08.880 --> 0:15:12.720 then number two is human likeness, right, this is the 0:15:12.840 --> 0:15:19.680 independent variable. Yeah, and if you have human likeness, like, 0:15:19.920 --> 0:15:22.280 what does that mean? Like what looks human? What doesn't 0:15:22.320 --> 0:15:25.720 look human? The like, we haven't pinned that down. So like, 0:15:25.760 --> 0:15:28.600 if you can't pin the dependent variable down and the 0:15:28.640 --> 0:15:32.000 independent variable down, it makes it really tough to study correct. 0:15:32.680 --> 0:15:35.160 And then there's a third one too. I love this one. Yeah. 0:15:35.160 --> 0:15:38.320 The third one is you know, the original hypothesis. It 0:15:38.360 --> 0:15:43.560 doesn't have a mathematical model that really specifies the shape 0:15:43.560 --> 0:15:48.680 of this curve. Right, So it's still hypothetical, I guess, right, 0:15:49.080 --> 0:15:52.400 which means that so if you look at Mori's graph, 0:15:52.480 --> 0:15:54.720 it was he just basically made a line, right. It 0:15:54.760 --> 0:15:57.480 wasn't based on any studies he'd done. The whole thing 0:15:57.560 --> 0:16:02.720 was really an essay more than anything else. So researchers 0:16:02.720 --> 0:16:06.640 who are trying to seriously study the scientifically have nothing 0:16:07.120 --> 0:16:11.320 that they're actually trying to place their findings against, which 0:16:11.440 --> 0:16:14.600 leads to put it puts them at risk for what's 0:16:14.600 --> 0:16:18.680 called the Texas sharpshooter fallacy. Yeah, greatest name fallacy around, 0:16:19.120 --> 0:16:22.720 and it's based on the idea that if you take 0:16:22.960 --> 0:16:25.120 a sharpshooter out in Texas and have them shoot at 0:16:25.120 --> 0:16:27.520 the side of a barn a bunch of times, some 0:16:27.560 --> 0:16:30.520 of them are inevitably going to hit the barn, and 0:16:30.560 --> 0:16:33.080 then the Texas sharpshooter walks up and then draws the 0:16:33.080 --> 0:16:36.680 bull's eye around the bullets that he already sunk into 0:16:36.680 --> 0:16:39.240 the side of the barn. That's the Texas sharpshooter fallacy. 0:16:39.280 --> 0:16:41.920 It's ignoring data like the ones where he missed the 0:16:41.960 --> 0:16:45.040 barn in favor of ones that fall into what you're 0:16:45.080 --> 0:16:48.280 looking for, the bullet holes in the barn. You could 0:16:48.280 --> 0:16:52.320 do the same thing with the data that you get 0:16:52.400 --> 0:16:55.280 from testing the Uncanny Valley if you have no model 0:16:55.320 --> 0:16:58.400 to fit it into already. Yeah, I think they would 0:16:58.400 --> 0:17:00.400 have done better if they would have just instead of 0:17:00.440 --> 0:17:05.680 trying to prove something, to maybe just research and call 0:17:05.720 --> 0:17:09.120 it a thought experiment, right, you know, right, But people 0:17:09.160 --> 0:17:11.520 are taking it seriously, and we'll we'll talk about some 0:17:11.560 --> 0:17:40.359 of this research right after this, Chuck, all right, So 0:17:40.400 --> 0:17:45.399 we're back. And despite the fact that this is really 0:17:45.440 --> 0:17:48.000 tough to study, it's not even established that it's a 0:17:48.040 --> 0:17:52.160 real thing in everyone's mind. By the way, there are 0:17:52.320 --> 0:17:56.600 people out there who are really studying the Uncanny Valley 0:17:56.600 --> 0:17:59.040 and trying to pin it down. Yeah. One of these 0:17:59.080 --> 0:18:02.480 people is at Dartmouth College psychologist. And I didn't look 0:18:02.560 --> 0:18:11.240 up their mascot, the pub Darts, the fighting pub darts 0:18:11.240 --> 0:18:13.959 at Dartmouth College. We're gonna hear from Dartmouth. But her 0:18:14.000 --> 0:18:17.280 name is Talia Wheatley, and she's done some research and 0:18:17.320 --> 0:18:21.360 has found that it's not just like some uniquely Western 0:18:21.440 --> 0:18:24.320 thing or American thing. It's kind of all over the world. 0:18:24.440 --> 0:18:28.400 She studied tribes in Cambodia and they have these same 0:18:28.440 --> 0:18:32.240 sensitivities to these things that look human but aren't human. 0:18:33.040 --> 0:18:35.560 And they've even found that and I think it kind 0:18:35.560 --> 0:18:37.960 of all comes down to the eyes. But they found 0:18:37.960 --> 0:18:42.680 just looking at the eye can be enough. Yeah, somebody 0:18:42.720 --> 0:18:45.280 can tell whether it's a human or not just looking 0:18:45.320 --> 0:18:47.480 at a picture of the eye, right. Yeah. And that's 0:18:47.480 --> 0:18:52.560 where I think people lose credibility in And we'll talk 0:18:52.600 --> 0:18:55.439 about movies and sculpture and all that stuff, but they 0:18:55.520 --> 0:18:57.959 just never get you can't get the eyes right, Like 0:18:58.000 --> 0:19:02.840 you can't put life in lifeless eyes. Yeah, hard, they try. 0:19:03.200 --> 0:19:07.800 Only God can u. And there was this other experiment 0:19:07.840 --> 0:19:11.000 where they, you know, like where you can morph af 0:19:11.000 --> 0:19:14.520 face digitally or whatever like that Michael Jackson Black or 0:19:14.560 --> 0:19:17.240 White video. Yeah. I think some people were creeped out 0:19:17.240 --> 0:19:20.240 by that even sure, but they would show this doll 0:19:20.400 --> 0:19:23.280 image and it would morph into a human face, and 0:19:23.359 --> 0:19:26.720 basically they would have people mark where they where they 0:19:26.720 --> 0:19:29.960 thought that it would look more human than doll and 0:19:30.040 --> 0:19:32.760 it you know, it landed about the sixty five percent 0:19:33.080 --> 0:19:36.240 mark as far as morphing into human, which I mean 0:19:36.280 --> 0:19:40.280 you can't really apply that necessarily, but just it's interesting offhand. 0:19:41.240 --> 0:19:45.520 Sixty five point is about where the Uncanny Valley happened 0:19:45.760 --> 0:19:48.280 in Maury's mind. Yeah, I would think it would be 0:19:48.480 --> 0:19:52.360 higher than that. Yeah, but um yeah, it's still super interesting. 0:19:53.040 --> 0:19:56.480 Um and you were saying that the eyes, and that's 0:19:56.520 --> 0:19:58.240 what you're betting on, is that it's going to turn 0:19:58.240 --> 0:20:00.879 out to be the eyes, right. Yeah. So trying to 0:20:00.960 --> 0:20:06.360 investigate what constitutes human likeness, there's a researcher named Angela 0:20:06.400 --> 0:20:10.160 Tinwell and she basically says, like, yes, it's all about 0:20:10.160 --> 0:20:15.320 the upper facial features and that we detect those, we 0:20:16.200 --> 0:20:19.400 read those. And so if there's any anything that's even 0:20:19.440 --> 0:20:22.679 just slightly off in like you know, the eyes or 0:20:22.720 --> 0:20:27.840 the brows or the wrinkles that form, that will lead 0:20:27.880 --> 0:20:31.639 to the Uncanny Valley. That's the creeping part or the 0:20:31.720 --> 0:20:34.320 smile too. She also says, well, yeah, and all these 0:20:34.320 --> 0:20:36.560 things kind of come down to evolution. And her point 0:20:36.680 --> 0:20:39.879 is like, you can't battle millions of years of evolution 0:20:40.480 --> 0:20:44.000 that has honed our our dumb, little human brain to 0:20:44.119 --> 0:20:48.639 detect something that's off about a face, right, it's just 0:20:48.680 --> 0:20:53.480 too much to overcome, basically, Right. This other researcher named 0:20:54.320 --> 0:20:59.439 Carl F. McDorman, who's from the University of Indiana, who actually, 0:21:00.200 --> 0:21:04.240 he's basically like dedicated his career to this. Now. He 0:21:04.400 --> 0:21:09.560 found that certain kinds of people, if you do like 0:21:09.600 --> 0:21:14.560 a personality inventory before testing for Uncanny Valley sensitivity, yeah, 0:21:15.080 --> 0:21:19.119 some types of people are predictably more sensitive to the 0:21:19.200 --> 0:21:23.720 Uncanny Valley than others. Specifically, he found that very religious 0:21:23.760 --> 0:21:29.840 people that makes total sense. Yeah, Neurotic people, yeah, and 0:21:30.520 --> 0:21:36.400 people with high sensitivity to animal reminders. It's basically anything 0:21:36.800 --> 0:21:40.760 that reminds you that, hey, you're super civilized and you 0:21:40.880 --> 0:21:43.959 drive a car and you know how to play poker, um, 0:21:44.160 --> 0:21:46.520 but you're still an animal, just as much as that 0:21:46.600 --> 0:21:50.879 ape over there is an animal reminder. People who are 0:21:50.880 --> 0:21:53.680 sensitive to that kind of thing tend to go off 0:21:53.720 --> 0:21:56.040 on the Uncanny Valley as well. And then people who 0:21:56.160 --> 0:22:00.320 are anxious are more likely to experience the Uncanny Valley 0:22:00.320 --> 0:22:03.280 as far as McDorman is concerned. Yeah, that anteresting makes 0:22:03.280 --> 0:22:06.399 sense too, because they're probably just more prone to be 0:22:08.320 --> 0:22:10.040 I don't know, just have a reaction to a lot 0:22:10.080 --> 0:22:12.640 of things like this. Right. But but we should say 0:22:13.400 --> 0:22:16.399 the science and all this, the fact that the independent 0:22:16.400 --> 0:22:20.520 and the dependent variable are still not defined. The science 0:22:20.640 --> 0:22:23.960 is this, This is like the scientific equivalent of that 0:22:24.080 --> 0:22:28.320 backward over the head half court basketball shot. Yeah, that's 0:22:28.359 --> 0:22:30.600 the level of science that these people are carrying out 0:22:30.640 --> 0:22:33.880 at this point, because there they're a lot of them 0:22:33.920 --> 0:22:38.480 sadly are conducting experiments based on something that again doesn't 0:22:38.480 --> 0:22:42.320 have a set, dependent or independent variable. So how can 0:22:42.359 --> 0:22:44.800 you do that? As my question, Well, yeah, I mean 0:22:44.800 --> 0:22:50.520 because in each experiment, they're going to be using different stimuli, 0:22:51.600 --> 0:22:55.879 different faces, whether it's a doll or a wax figure 0:22:56.119 --> 0:22:59.440 or a CGI character, and then they're gonna be doing 0:22:59.480 --> 0:23:02.440 different things and have different expressions, and each person has 0:23:02.480 --> 0:23:04.960 their own subjective take, so it is a very tough 0:23:05.000 --> 0:23:07.160 thing to kind of nail down. Yeah, and I think 0:23:07.200 --> 0:23:09.760 some of them are actually trying to form the basis 0:23:09.880 --> 0:23:14.480 of this field of study right now, they're doing the groundwork. 0:23:14.880 --> 0:23:16.879 But I think some of them also are like just 0:23:16.960 --> 0:23:19.800 chasing headlines, Like there's no better way to get get 0:23:19.840 --> 0:23:23.439 into the media cycle with your study than releasing some 0:23:23.520 --> 0:23:27.399 findings on the uncanny Valley. People just tied up. Sure 0:23:28.680 --> 0:23:31.320 they love it. One thing I thought was interesting was 0:23:31.840 --> 0:23:34.880 at Princeton they tried this with monkeys and they found 0:23:34.920 --> 0:23:38.160 the same thing happen when they had these realistic looking 0:23:38.200 --> 0:23:42.399 but fake monkey faces. The monkeys were like ah and 0:23:42.480 --> 0:23:46.680 turned away. It did make me think, though, like all 0:23:46.720 --> 0:23:50.720 the you've seen these situations where like an orphaned animal 0:23:50.760 --> 0:23:54.639 has a creepy puppet mother. Yeah, I know exactly what 0:23:54.640 --> 0:23:57.760 you're talking about, and they seem to like that. But however, 0:23:57.880 --> 0:24:01.000 and this is a bit of a spoiler, but towards 0:24:01.000 --> 0:24:03.840 the end of this article, it points out that human 0:24:03.880 --> 0:24:06.399 babies don't have this reaction at first either, and that 0:24:06.720 --> 0:24:10.400 it's kind of learned. So maybe that explains it. Maybe 0:24:10.400 --> 0:24:14.280 with the animals. I know you're talking about that that cage, 0:24:14.320 --> 0:24:18.480 like why or monkey mother, it's super creepy. It's a 0:24:18.480 --> 0:24:20.560 black and white photo. Well, no, I mean they do it. 0:24:20.600 --> 0:24:22.639 There's all kinds of animals. Well, they'll they have like 0:24:22.680 --> 0:24:26.600 a fake tiger or a fake duck or whatever, just 0:24:26.640 --> 0:24:28.760 so the animal will feed or I mean it's usually 0:24:28.760 --> 0:24:30.840 an animal that that milks from the mother. I guess 0:24:31.080 --> 0:24:33.959