Pixar's Oscar-winning 1988 short "Tin Toy" by John Lasseter and, in particular, the baby character (Billy) was the first time I had encountered the monstrousness of the uncanny valley.
Uncanny Valley has only been proven on 2d to my knowledge.
It has no relevancy to real life (Outside of watching TV)
They took a robot famous for being in the Uncanny Valley to an old age home and the residents loved it.
It could be possible we don't have the tech yet to produce a 3D object in the Uncanny Valley, but even when you see artistic work by artists like Ron Mueck you'll notice in real life there is no Uncanny Valley
> Uncanny Valley has only been proven on 2d to my knowledge.
I think that it's probably just far more likely to be encountered with CG than something most of us see in real life. Plenty of researchers are working on the issue because it does make some people uncomfortable, not just so that their perfectly fine robots will look better on instagram. I occasionally run into real life examples around Halloween in some of the masks but the effect isn't ever consistent. It's just certain moments when movements (or a lack of them) seem to defy what the brain expects. I've heard the Uncanny Valley used to help explain the reason people can find disabled people, taxidermy, the dead at open coffin funerals or clowns "scary" in real life.
It's not even something everyone experiences, or maybe it's better to say that what triggers it in one person might not bother another, so it's hard to pin down. I'm guessing familiarity and time helps a lot too so even if a 20 second video of Repliee R1 creeps you out you'd eventually not be bothered by it if you had one in your house all day. I wouldn't be surprised if a sculpture, properly contextualized in an exhibition environment, was less jarring than a random photo online for the same reason.
The movies "The Polar Express" and "Beowulf" both by Robert Zemeckis were early attempts at using facial animation captured and transferred to digital characters. Both had problems with the uncanny valley.
The HN thread on This person does not exist [1] 9 months ago about generated human faces is where I first met this concept. Search for "uncanny" and "disturbing" (several occurrences).
The article offers several explanations. Let me offer you another: Upon perceiving a face, the brain reads its visual cues to determine its bearer's intentions. The uncanny valley occurs when we fail to determine a perceived individual's intent due to conflicting visual cues.
I had a real-life experience with this phenomenon recently, when I got a robocall made by an interactive voice responder. The dialog at first sounded almost human, but there was something that was just a bit off with the responses not quite matching my questions.
Typically I like to play around with telemarketers, keeping them on the line for a while, asking dumb questions etc. instead of just hanging up on them -- especially if it sounds like a scam to begin with. But in this case it was either a well done chat bot that I was talking to, or now that I think of it (more likely) someone had a bunch of pre-recorded phrases that they would choose as responses (I can see this being done by an offshore telemarketing pit).
I think somehow our body is pre-wired for some smells.
Healthy food is nutritious, so it smells good.
In the same way our body knows "dead person" is a very terrible smell, even if we have never encountered it in our life.
Maybe the uncanny valley is the same thing for vision -- we are wired to be wary of sick or dead people, who look "off" because of an expression, their posture, skin color, etc.
That sounds like the first theory in the linked Wikipedia article.
> Mate selection. Automatic, stimulus-driven appraisals of uncanny stimuli elicit aversion by activating an evolved cognitive mechanism for the avoidance of selecting mates with low fertility, poor hormonal health, or ineffective immune systems based on visible features of the face and body that are predictive of those traits.
It’s very surprising that in-group/out-group factors didn’t really make the theories’ section.
“Looks human, but not quite what I’m used to.”
Seems like a decent way to quickly spot nonmembers of a social group. I wonder if an experiment could classify the revulsion and compare/contrast it to something like, when a person meets someone of an unfamiliar subculture.
> I think somehow our body is pre-wired for some smells.
This is a result of millions of years of biological evolution. Bodies that were "less" wired for certain smells were likely at some sort of disadvantage. After a lot of generational churning, the average body appears to be "prewired" for the most important smells.
> In the same way our body knows "dead person" is a very terrible smell, even if we have never encountered it in our life.
I only recently discovered the band Steam Powered Giraffe, a group of people pretending to be 19th century robots, and particularly in some of their older videos, they sometimes appear to enter the Uncanny Valley from the other side, by moving in a robotic way that's simultaneously fluid, but also appears slightly non-human.
I guess this might also be the original attraction of robotic dancing, but for the most part we're used to it. But SPG seems to demonstrate that you can still get that uncanny eerie feeling if you do it in just the right way.
16 comments
[ 4.6 ms ] story [ 40.8 ms ] threadIt has no relevancy to real life (Outside of watching TV)
They took a robot famous for being in the Uncanny Valley to an old age home and the residents loved it.
It could be possible we don't have the tech yet to produce a 3D object in the Uncanny Valley, but even when you see artistic work by artists like Ron Mueck you'll notice in real life there is no Uncanny Valley
I think that it's probably just far more likely to be encountered with CG than something most of us see in real life. Plenty of researchers are working on the issue because it does make some people uncomfortable, not just so that their perfectly fine robots will look better on instagram. I occasionally run into real life examples around Halloween in some of the masks but the effect isn't ever consistent. It's just certain moments when movements (or a lack of them) seem to defy what the brain expects. I've heard the Uncanny Valley used to help explain the reason people can find disabled people, taxidermy, the dead at open coffin funerals or clowns "scary" in real life.
It's not even something everyone experiences, or maybe it's better to say that what triggers it in one person might not bother another, so it's hard to pin down. I'm guessing familiarity and time helps a lot too so even if a 20 second video of Repliee R1 creeps you out you'd eventually not be bothered by it if you had one in your house all day. I wouldn't be surprised if a sculpture, properly contextualized in an exhibition environment, was less jarring than a random photo online for the same reason.
[1] https://news.ycombinator.com/item?id=19144280
Typically I like to play around with telemarketers, keeping them on the line for a while, asking dumb questions etc. instead of just hanging up on them -- especially if it sounds like a scam to begin with. But in this case it was either a well done chat bot that I was talking to, or now that I think of it (more likely) someone had a bunch of pre-recorded phrases that they would choose as responses (I can see this being done by an offshore telemarketing pit).
I think somehow our body is pre-wired for some smells.
Healthy food is nutritious, so it smells good.
In the same way our body knows "dead person" is a very terrible smell, even if we have never encountered it in our life.
Maybe the uncanny valley is the same thing for vision -- we are wired to be wary of sick or dead people, who look "off" because of an expression, their posture, skin color, etc.
Just a theory.
> Mate selection. Automatic, stimulus-driven appraisals of uncanny stimuli elicit aversion by activating an evolved cognitive mechanism for the avoidance of selecting mates with low fertility, poor hormonal health, or ineffective immune systems based on visible features of the face and body that are predictive of those traits.
“Looks human, but not quite what I’m used to.”
Seems like a decent way to quickly spot nonmembers of a social group. I wonder if an experiment could classify the revulsion and compare/contrast it to something like, when a person meets someone of an unfamiliar subculture.
This is a result of millions of years of biological evolution. Bodies that were "less" wired for certain smells were likely at some sort of disadvantage. After a lot of generational churning, the average body appears to be "prewired" for the most important smells.
> In the same way our body knows "dead person" is a very terrible smell, even if we have never encountered it in our life.
Again, Darwinian evolution can explain this.
I guess this might also be the original attraction of robotic dancing, but for the most part we're used to it. But SPG seems to demonstrate that you can still get that uncanny eerie feeling if you do it in just the right way.