[HN Gopher] AI can interpret animal emotions better than humans
___________________________________________________________________
AI can interpret animal emotions better than humans
Author : marojejian
Score : 85 points
Date : 2025-02-13 21:16 UTC (3 days ago)
(HTM) web link (www.science.org)
(TXT) w3m dump (www.science.org)
| jaharios wrote:
| While Humans can be seen as the most intelligent, we are hyper
| focused on "human way of thinking" in a way that we lose our
| "basic" instincts and abilities that other animal have.
|
| My dog can understand my voice tone and emotions way better than
| I can understand hers, also animals can understand the difference
| sounds we make (words) that affect them, way better than our
| understanding of animals sounds.
|
| Don't get me wrong we can make tools and we can experiment and be
| able to suppress all other animals. But a solo, "naked" human is
| like an office worker in world of manual labor.
| unification_fan wrote:
| > Don't get me wrong we can make tools and we can experiment
| and be able to suppress all other animals. But a solo, "naked"
| human is like an office worker in world of manual labor.
|
| Nah most of it is nurture. Raise a human in the wild and he'll
| be more in tune with nature. We have become alienated from the
| environment we evolved in and that's why you feel like a "naked
| office worker" on your own planet despite being the result of
| billions of years of adaptation.
|
| Most humans simply ignore animals when they communicate. Both
| because they're ignorant and because they won't bother to
| listen. You can't expect an animal to talk with human words,
| but they talk all the time. Pets actively have conversations
| with us.
|
| Plus there's this hardwired notion in our culture that humans
| are inherently superior to all animals but that's a very self-
| centered and short-sighted understanding of the world. We are
| more intelligent, yeah, but that's about it.
| neom wrote:
| I'd argue we've gone far beyond alienated. We're actively
| rejecting and dismantling, en-masse, the very systems we
| foundationally operate within.
| jaharios wrote:
| I agree, I am not saying that it is not nurture, in fact the
| opposite. Hyper focused on "human way of thinking" is not
| something you are born with, you adapt to it. In fact if you
| don't do it early you will never be able to 100%, in a way we
| rewire our brain to cope with how we want it to operate to be
| able fit in.
|
| Our language for example, requires to be "forced on us" from
| early stages or you will never be able to "get it" [1]
|
| > Most humans simply ignore animals when they communicate.
| Both because they're ignorant and because they won't bother
| to listen. You can't expect an animal to talk with human
| words, but they talk all the time. Pets actively have
| conversations with us.
|
| With my dog I can understand angry/playful/sad/afraid/(give
| me food) barking/sounds and especially body language. But
| hearing "dog words" in random barking? Impossible.
|
| [1]https://en.wikipedia.org/wiki/Language_deprivation_experim
| en...
| unification_fan wrote:
| > With my dog I can understand
| angry/playful/sad/afraid/(give me food) barking/sounds and
| especially body language. But hearing "dog words" in random
| barking? Impossible.
|
| No what I meant is that animals use body language, smells
| and all kinds of non-auditory cues in order to communicate.
| And they rely a lot on behavioral reinforcement in order to
| communicate efficiently with us.
|
| They don't really use verbal language like we do and
| therefore they also lack the tools that are required for
| abstract conversation.
|
| They don't talk about complex topics like astrophysics.
| They usually just talk about their immediate needs but they
| can also convey more complex emotions like trust and guilt
| -- displaying a rudimentary theory of mind.
|
| Regardless, there are many interactions you can have with
| your pets that entail a string of questions and answers.
|
| Example:
|
| 1. My cat comes up to me and sits there staring. She means:
| "I need something of you, but I can wait"
|
| 1a. If I don't get up in a while, she will come closer and
| bump my leg. She means: "come on, please"
|
| 2. I ask her what's up and _get up from my chair_ to signal
| that I am ready
|
| 3. She recognizes this signal, having seen it many times
| before, and heads for the bowl/the door/the balcony/the cat
| tree depending on what she needs
|
| 4. I understand her need and give her what she wants
|
| 5. She trills or purrs to tell me that I'm on the right
| track/my assessment is correct/to thank me
|
| That is clearly a conversation, albeit a simple one.
|
| Lately she's become addicted to bird and mice videos on
| Youtube so she comes up to me and stares intently at my
| laptop and/or desktop until I put those on for her.
| idiotsecant wrote:
| >We have become alienated from the environment we evolved in.
|
| Before we start painting with all the colors of the wind too
| much in this thread, this is not entirely a bad thing. We are
| removed from stressors such as 'being eaten by large
| predators' and 'dying of infections from wounds'. There is a
| lot of 'nature' that out ancestors would be quite happy to be
| 'alienated' from.
| taurknaut wrote:
| Sure, but this completely neglects the aspects of humanity
| we left behind moving into sedentary communities. We will
| be forever blind to what we lost, and the morons among us
| will claim we lost nothing.
| theshackleford wrote:
| > morons
|
| Ah yes, that old chestnut. "Anyone who disagrees with me
| is a dum dum"
|
| How persuasive your arguments are.
| talldayo wrote:
| Which parts are you referring to? The part where you kill
| a guy with a rock if he looks at your wife funny?
|
| We always lose something when evolving, that's okay. You
| can keep living in whatever way you want to, as long as
| it doesn't disrupt the liberty of another person. If
| you're mad that the world embraced secularity over
| spiritualism, or that men aren't fist-fighting for
| resources, blame yourself for not modernizing. Without
| any serious examples, your comment basically just reads
| like a trad dogwhistle.
| darkerside wrote:
| > We are more intelligent, yeah, but that's about it.
|
| You say that like it's not the defining characteristic of our
| power over the natural world
| Kostchei wrote:
| Depends on your life experiences and working environment. If
| you have worked in prisons and places with a lot of physical
| violence you can (some don't) acquire a distinct and accurate
| sense for emotion and threat, based on sound and body language.
| The actual words don't matter so much, but the interaction of
| tone, distance, stance etc, they tell you a huge amount. People
| can be saying "no" and be just asking a question or pleading
| their case, and they can be saying "yes" and mean "i want to
| kill you". I used to follow the tone, and when it was going to
| end badly, make sure I was standing behind the person who was
| about to start violence (being responsible for physical
| security in that environment), just as it was about to kick
| off... Pretty grim work. But yes, you can use your intelligence
| to learn that stuff. Don't need to be a puppy.
| darkerside wrote:
| Yeah, I think we spend so much of our childhoods, if they are
| healthy, learning to disregard those signals. Authority
| figure yelling but will not hurt us. Trust.
|
| We rational humans overthink our first instinct and even
| learn to ignore it. And it helps us function in traditional
| society.
| knallfrosch wrote:
| If you had a 5 meter tall dog master family responsible for
| giving you food, shelter and keeping you safe from dog-driven
| cars, you'd be quite good at reading dog expressions too.
| block_dagger wrote:
| Misleading HN title. The article says AI can predict stress
| better than humans and only poses the question of general
| emotions.
| trebligdivad wrote:
| I bet someone will turn that to Horse race betting then based on
| the look of the horses before/at the start of the race.
| terrut wrote:
| This is a thing. Going to the races and watching horses warming
| up in a paddock is very informative to those in the know. The
| first time i went, my girlfriend's father pointed out a horse
| that none if the others would run past, and that horse won it's
| heat.
| shakna wrote:
| After a year-long experiment, my professor's experiment at
| training AI on horse races found that it could fairly
| accurately predict podium winners - it hyperfixated on red
| shirts. Happened that a rather famous jockey usually wore a red
| shirt. Him being in a race was a good indicator that he'd win
| it.
| neom wrote:
| I guess in theory you could maybe apply that to many things.
| Horse racing, poker, the courts system, commerical pilots.
| taurknaut wrote:
| Hard to imagine this hasn't already been done for decades. What
| role would AI play?
| anon-3988 wrote:
| Humans are clearly bad at analyzing other people emotions based
| on how much misunderstanding there are out there. Just look at
| how bad people are in relationships. Someone shared their
| experience in an event talking to a girl. I just listened,
| thinking, "Do they not realize that you were clearly in the wrong
| here?".
| wruza wrote:
| There's a huge difference between analyzing emotions and being
| well-trained on a spectrum of behaviors. You may spend hours at
| a therapist to understand your own emotions, but then you go
| out and are expected to read people in seconds. That's
| nonsense. I will even say bullshit. You may just know the
| social protocols better, but there's nothing to analyze
| usually. It's all common bugs in the heads of those you
| communicate with. The true analysis could be possible if people
| expressed their emotions properly, but most social games are
| almost designed to be as misleading as possible.
| anon-3988 wrote:
| > It's all common bugs in the heads of those you communicate
| with.
|
| The problem is that people are not equipped to fix those
| bugs, its very hard to fix a bug with a buggy software after
| all. Which is why, a second opinion, in this case, an AI (or
| a therapist, a friend, etc) will help significantly.
|
| I _should_ have commented on it, but I kept quite. If he
| really wanted to know what went wrong, I should have told him
| what happened, but I don't know that. If he had some ML
| chatbot analyzing the images and such, he would have had a
| second opinion.
| hombre_fatal wrote:
| Reading a facial expression is one thing.
|
| But I think what you're talking about is someone's ability to
| simulate the internal state of someone else. Or its little
| brother: simulating it correctly and then getting mad at the
| result due to ego.
| cardanome wrote:
| Misunderstanding will always be a fact of life. The diversity
| of humans makes it impossible to reliably interpret other
| people's emotions, especially when there is not a strong
| cultural context. We just need to learn to communicate openly
| and explicitly.
|
| What creeps me out is that so many people have zero self-
| awareness that there is a difference between what has been
| communicated and their interpenetration of it.
|
| They will be in a bad mood and conclude the text they just
| received must have been written in a very rude "tone".
|
| They see your face and conclude you must be angry at them.
|
| They take their subjective interpretation as the same as the
| objective truth and absolute hate to be challenged. The believe
| themselves to be "empaths" and "good communicators".
|
| If you think you are very accurate at understanding other
| people's emotions, you are not. That is not possible. The
| inside state of people can not be measured by looking at their
| outside expression, you can only make predictions. You have to
| ask people how they feel.
| dyauspitr wrote:
| One of the problems here is that "wrong" is not universal in
| places that don't have a single, universal culture.
| spicy-punk-fog wrote:
| AI hype is right at its equivalent of Radium Chocolate stage of
| nuclear physics today
| egberts1 wrote:
| But AI is currently failing at American Sign Language, notably at
| facial expressions, miserably.
|
| Yes, we are working on that but I see writing on the wall and it
| is not soon enough.
| taurknaut wrote:
| ASL seems like a really hard problem. I've learned a fair bit
| myself being friends with a few interpreters and deaf people
| and two different people can sign the same thing in a way that
| would look very different to an AI. Sometimes it feels like you
| have to put a fair bit of effort into understanding how a
| specific person signs (from my very inexperienced perspective).
| I'm curious if this could be overcome with sufficient data, but
| where is this massive archive of videos of sign? There are some
| databases certainly, but nothing close to the level of which we
| have written and spoken english via the internet. Plus then you
| get into regional and cultural dialects... i think banks will
| be obligated to hire interpreters for the forseeable future.
| blogabegonija wrote:
| Then AI is ready and can into U.S presidency.
| farleykr wrote:
| Maybe if we spent more time learning how to interpret animal
| emotions than we do building AI to do it for us the title would
| read vice versa.
| RobertDeNiro wrote:
| Doesn't take a phd to know that those pigs are unhappy.
| chillingeffect wrote:
| I wonder if they trained the systems on the times when theyre
| lowered?
| lcnPylGDnU4H9OF wrote:
| I'm not particularly happy to read about researchers
| intentionally causing distress to animals in order to study the
| response but the documentation of their research is a few
| grades above "obviously it's distressed". The point is to be
| able to determine whether they are feeling stress in situations
| that don't have such obvious stressors.
| hombre_fatal wrote:
| Takes a certain kind of person to care, though.
| taytus wrote:
| This type of headline is like: 'computers are better than humans
| at chess!' Like... isn't this something obvious by now?
| jon9544hn wrote:
| Mention of an app, but no references. Does anyone know of it?
| SapporoChris wrote:
| These hungry piggies are the unsuspecting beta testers of
| Intellipig https://pure.sruc.ac.uk/en/projects/a-face-based-
| automated-o...
| gizajob wrote:
| Correction - AI can interpret _videos_ of animal emotions
| (recorded under controlled conditions) better than humans
| (allegedly).
| taurknaut wrote:
| ...than untrained humans, presumably, because otherwise how would
| you produce the model?
| d--b wrote:
| Says who? The cows?
| Dotnaught wrote:
| We can understand animal emotions (mammals at least) reasonably
| well if we choose to try, and even in the absence of
| understanding, we can assume, for example, that caged animals
| would rather be elsewhere. But self-interest (or economic
| interest) often makes it more convenient to ignore what's
| obvious.
| harimau777 wrote:
| I wonder if it's possible that what is actually happening is that
| human ability to understand animals is not optimized towards
| "understanding objectively what the animal is feeling" so much as
| it is optimized towards "understanding how to get them to do what
| we want."
|
| Assuming that a significant portion of our ability to understand
| animal behavior comes from evolutionary instinct or ancestral
| folklore, then it seems reasonable that the result might be
| highly pragmatic. For example, our ancestors may have only cared
| about identifying whether a dog was communicating sufficient
| submissiveness to indicate that it would follow orders. Whether
| that submissiveness came from love or fear of punishment may not
| have been important.
| simianparrot wrote:
| If humans don't understand animals how do we evaluate whether the
| language model does..?
___________________________________________________________________
(page generated 2025-02-16 23:01 UTC)