So… what would YOU say to a whale? When you're done here, follow me to Planet Wild to find out how we used drones to drop camera tags on blue whales! www.planetwild.com/besmart/14_2
One argument to try to communicate with whales is that it would give us insights into how we could potentially communicate with other intelligent life in the universe. Much like in the movie Arrival, communication is about gathering context, gaining understanding, and making connections. Figuring out how to do that with whales is a tough challenge that we would learn from.
Also we might learn a lot of stuff about earth and other animals and many more from those legendary species. I believe they have wisdom that transcends through their generations for all these years!
I think everyone’s missing the point. It’s not about using this info to talk to aliens, it’s just the human condition. We’re social animals and our happiness, survival and evolution is somewhat dependent on how social we are. And I’m not saying that from the perspective of us needing to talk to animals to ensure our survival and happiness, just from the point of socialization being part of the human condition. It’s an indelible part of our existence. Even as an introvert I would love to be able to know what a whale, monkey or tree kangaroo is thinking. How they think, how different it is from the way we think, or whatever. Connecting to others, even as an introvert, is just a supremely satisfying thing to think about and do. Now imagine being about to do that with a freakin’ whale.
I am a retired submarine sonar tech and listening to whale song and other sea creatures was the only thing getting me through those long boring midwatches sometimes
@@diegoandrade3792 Which god, and why that particular one? What’s your best evidence it exists? The bible is the claim, and “look at the trees” or “everything that is created needs a creator” isn’t quite going to cut it, I’m afraid.
Considering that whales have a lifespan of ~100 years, some even above 200 years it is more like "Sorry that we slaughtered your brothers and daughters for candles". Maybe a species that's kicking baby seals 50 m high in the air for fun would understand, though.
Thank you so much, Joe, for hosting our latest Planet Wild mission! Your passion and insights made the exploration of using cutting-edge technology to study blue whales not only educational but truly captivating. We're thrilled to have had the opportunity to collaborate with you on this fascinating project.
I certainly hope one day we find a way to properly communicate with at least one species, and then maybe a few more people will start to understand the world we are part of, rather than thinking we exist separate to nature. If you pull back far enough, we're all one big organism on Earth.
Why should we try to talk with whales: 1. Because it's cool 2. It's a learning experience for communicating with potential extraterrestrial life. 3. Being able to communicate with another species would be a massive boon for conservation efforts. 4. We share this world with them. If they're as intelligent as they seem to be, then we owe it to them to be better roommates, and opening a dialogue would be an excellent first step. 5. (this space reserved for black swan events)
Some of those expert answers were some of the least imaginative answers anyone could come up with to the question. Particularly after driving home the importance of the whale song recordings. Who's to say it wouldn't be good for the whales? It could be the best possible thing for them.
Being German and hearing about "Umwelt" as being a concept introduced to biology in the early 20th century, in turn spawning the field of ecology, is wild. In our everyday language it literally just translates to "surrounding world" - and just means environment and it feels completely natural to use. Hearing English speaking people of scientific background refer to Umwelt as "perceived environment of an animal" feels a little odd, like someone using a word way too specifically, even though they're not wrong. I would've expected that word to be way older, too (it was borrrowed from Danish and used since the 19th century, but only in a geographical sense). Thanks for providing this little trip into my own language's heritage.
"Complex" as in they can identify any of the sixteen thousand annoying objects/creatures that almost step in their holes, right? Not "complex" in an interesting way, right? Lemme know if I'm wrong pls
@@Cajek2 Professor Constantine Slobodchikoff is an animal behaviorist and conservation biologist. He is a professor at Northern Arizona University and he has made a 40 year study of the language of prairie dogs. Their language skills are not only the most complex, but also have traits that are only ever displayed by humans such as displacement. Displacement conversation is talking about something that is not in Immediate sight. So far only humans and prairie dogs are believed to do this. Not even the other primates do it, nor whales or dolphins. His work is fascinating. There have been also cases of where prairie dogs have pre-planned to trap a snake disturbing a prairie dog town by planning dead end holes for these snakes and burying them alive. Prairie dogs have adjectives, verbs and subjects, so they are able to recognise individual species and where they are going. They also come up with new words and have different languages in different areas like humans have Spanish and Chinese.
They also have some kind of social chatter but we haven't cracked the rosetta stone to this due to lack of context and AI would definitely be an interesting use case for this.
@@MisterMakerNL it can be useful. For example if animals understand the idea of a danger call, it could herd wild animals away from an area that could harm them, such as wind turbines.
I'm not at all worried/thinking about what if those whale "songs" we sent into space are actually them being like "Help, the hairless monkeys are after me!" and "Save me from these brutal genocidal maniacs!" etc. Nope, not at all... 😬
Sperm whale's sonar abilities really blow my mind. They are capable of using sonar as a weapon if they aim it and just blast things with extreme sonar. They also dive super deep and hunt squid in pitch black so sonar is crucial. It's fascinating
Here's the thing: the question of "why would you talk to a whale?" is effectively the same as "why would you talk to an extra terrestrial?" All of the objections the interviewees made about talking to whales are just as valid when talking about, say, the inhabitants of Proxima Centauri C.
This is a question that I don't quite understand coming from scientists. My answer to that question would be "Why not?". They dedicated their life to understanding whales so why wouldn't they want to be able to ask the whales their question directly? The same question could be asked about anything: "Why draw pictures?", "Why make music?", "Why study nature?" and the answer is simply "Because we are curious".
@@Shadow05eth Also, whale jokes. I want to know if they laugh. Some primate seem to and many animals show happiness and empathy for their friends/family (barnyard animals do) protecting each other from predators. So ya, let's see what they have to say.
This video made me think about the possible consequences of learning to talk to them, and i'm pretty sur they would suffer from it, as we would quickly put them to work.
Aliens... Because of technology. It's like saying why would you talk to someone else who speaks a different language. Edit: learning how to communicate with another species is the starting point for future communication with aliens
04:26. Also, in 1986, the movie Star Trek IV: The Voyage Home a.k.a. The One With the Whales, made the existence of humpback whales a key for the planet's survival.
@@joejoeington6899 Anatomically modern humans have been around for roughly 300K years. It just took a long time for the technological advances to get going.
Yeah, what was that 150 million about. Sure you had ichthyosaurs during the mesozoic that resembled dolphins but were not really cetaceans. From what I remember it was 55 million years ago that ancestors of cetaceans first started adapting to aquatic life.
Perhaps using AI to listen and understand whale song is best. I think one of thing they might say is "it's too noisy down here". One of the biggest threats to their long term survival is the steady increase in ambient noise due to shipping, fishing, and petroleum extraction, and their related activities. It stands to reason that they depend on their whale song for communication, navigation, finding food, and responding to threats. That is a challenge to deciphering a language if it exists. Whale songs are probably used for other purposes than speech.
6:28 As an AI researcher, it’s worth pointing out that what the host is asking is exactly what these large language models are. They learn to predict any signal. In any number of dimensions
Hmmm, they learn to predict an answer that the user would like. I think the current predictive models would tell a very interesting story about what the whale was saying, and true AGI would say "they are communicating location, desire to mate, food, pain, and little else "
@@TheGrinningViking that’s not how they are trained. At least not the base model. In the base model, the objective is to predict the next word. Using such a pre trained model, you could achieve what you state.
@@LimeyLassenI'm not an ai researcher, but I know some stuff about ai, and the first sort of strategy for tackling this would be 2 fold: take these recordings of whales sounds, and try to use ai to categorize individual sections into different "words" of some sort. Then, we can rewrite the audio in terms of those words, and apply the transformer models to them. The first step would be much harder to get correct, especially with having to deal with the possibility of multiple voices occurring at the same time, but it at least sounds like a feasible path.
@@purewaterruler This is the one AI would help where our ears and eyes and brains can't parse things evenly. good thing there is thousands of upon hours of recordings. Then use that to see how our current understanding of it "fits" those parsed "words", and yeah then applying it what we know is corolated and perhaps see what the image looks like from there.
I love Joe and I love Planet Wild. So great to see them collaborate. I watched both videos and I learned so much about the fascinating world of blue whales. 🐳❤
I was born in 1974, and I remember the Save The Wales campaign. It was everywhere. We even learned about it in school and studied wales. They are magnificent creatures.
Times have changed. Their language is now taught in their schools, it appears on the road signs & they even have their own parliament - all thanks to those campaigns to Save Wales. I'm not sure I'd call them magnificent though - they have an unhealthy interest in sheep.
As someone who has worked with animals for over 25 years I respect and relate to the perspective of each one of the professional speakers in this video. They all echoed the empathy, consideration, respect, and humble understanding we learn to embrace as animal professionals. But given our untrustworthy history as a human species in general, I do not believe as a whole, we are capable of putting aside our arrogance and general propensity towards dominance and therefore should not venture where we are not needed. We don't need the Navy training whales to alert them of the enemy - we need to work on having fewer enemies.
When I was in junior high, I wanted to be a cetacean sociologist...partially inspired by Star Trek IV, and partially by a science research paper I did then. I *loved* listening to whale song and had several tapes and cds, some with classical/new age music with the whale song. Thanks for the happy memories :)
There's a fantastic YA book called "Whale Talk" about a teenager who brings together a group of outcasts to compete on their high school swim team. The book uses whale songs as a metaphor for how we can't understand each other and we just make assumptions
@besmart, what is hilarious is, when you get to the idea that wanting to communicate with whales is human-centred, all of the researchers use human perspective to answer the question, "What would you ask them?" What would *I* ask them? I would ask, "What is important to *you,* and btw, what is the general consensus among you about us humans?" I would want to know what they consider to be beautiful in sound, or whatever other medium they use. I would ask them what they might want to know about me, about us, about our arts and sciences. I would be open and honest. I would ask them what I could do for them, help them with, or share with them, Would they like to hear human music? Would they like to become famous as singers among humans? How can our cultures exchange in a way that would benefit our world?
17:50 Wait, what, 150 MY?! I thought whales appeared quickly after the -65MY extinction event. There's an episode on PBS Eons about it. Someone knows more?
1:57 --- Growing up in the 70s, my parents had that exact same vinyl record of Songs of the Humpback Whale. Fond memories of listening to it. I wonder whatever became of that album.
It's a similar thing with birds. They communicate with what we call songs, Yet we haven't been able to really talk to them in their language. And we've been able to study them closely for a long time.
Maybe birds are just using simple sounds to communicate, not language. Ravens are quite intelligent, and so are octopuses/octopi and chimps, but we're not focusing on them rn. The best candidate for having a language is currently sperm whale.
Another factor worth considering is the fact that whales from different regions might have multiple languages, if I’m not mistaken marine biologists have already found that orca have multiple languages, so we might have to decipher the language of whales from each region.
16:32 Whales aren't the easiest to observe, but we already have other highly social animal species in observation. It wouldn't be too hard to fully cover for example a chimpanzee habitat with sensors and monitor a baby from the moment it's born until it's mature enough to communicate which usually takes much less than 5 years and that's for a very closely related species. Other species with a shorter lifespan would even be faster to comprehend such as our beloved pets: cats and dogs also interact with each other and mature in less than a year. So while whales are the biggest challenge, other species are a lot more accessible.
It. doesn't seem like very many animals have commuication that looks as much. like lannguage as whales, so studying them might not lead to a translator. Birds and bats seem to me like good potentials to study.
I unironically asked an AI this question and it's response - "The idea of communicating with whales is fascinating, and AI could potentially play a role in facilitating such interspecies communication, though there are significant challenges involved. Some ways AI might assist in talking to whales include: - Analyzing whale vocalizations and body language using machine learning models to try to decipher patterns and potential meanings. AI could help identify distinct "words" or signals whales use. - Translating between the acoustics and timing patterns of whale communication and human language formats that are easier for us to comprehend. - Simulating and generating potential whale vocalizations that could be played back to whales to attempt primitive communication once patterns are recognized. - Controlling underwater robots or interfaces that could both receive and broadcast communication attempts with whales more naturally in their environments. However, the barriers are still immense. Whale intelligence, language capacity, and methods of communication are very different from humans. We may be fundamentally unable to communicate rich conceptual information between the species. Additionally, the risk of disturbing whale behavior or causing unintended harm through incorrect assumptions about communication acts as an ethical constraint on such experimentation currently. So while AI analysis and simulation tools could provide avenues for exploration, deciphering and communicating with an alien intelligence like whales may remain extremely challenging even with advanced AI capabilities involved. It's a fascinating goal that may require an extended "listening" period before two-way interaction can be realistically achieved, if ever."
I have no problem with people communicating with them but there is a deeper issue, context. People talk about people things, whales likely talk about whale things. Getting whales to understand human things without experiencing our environment might be exceptionally difficult. What we communicate with them about might just be limited to things that impact them, like ocean conditions and weather, earthquakes, sea life migrations, etc. Some of our concepts might be so alien to them that they may never be able to understand.
If A.I. is capable of decoding whale sounds that’d be an amazing breakthrough and could potentially lead to us understanding Migo’s and other mumble rap artists lyric’s one day. Absolutely terrifying!
I suspect whale languages primarily convey what they need to know every day, including: finding food, avoiding danger, navigating, attracting a mate, locating and identifying members of the group, teaching the young, and expressing emotions.
We should talk to them. We could communicate how we can avoid bothering each other and instead of beaching themselves to keep from predation when they are sick, we could meet them, provide help and protection. We could ask them if they want to remain in a zoo or be free. Unless you are so narcissistic to assume to know what is best for them, it makes sense to want to ask what they think is best for them.
Firstly knowledge for knowledge sake is good and secondly if we figure out what whales are saying to each other it will give us insight into what we might say to them but learning what they’re saying to each other is enough for me to be fascinated
I think the real key to getting this to work is testing different approaches using human speech. Train an AI on human speech without telling it what the words mean, only providing certain context clues. If we can get that to work, then the same approach should work with other species.
They would probably come to call us the "noisy, stinky people" due to the sheer volume of both noise & chemical pollution we saturate their sea with. Currently, the human generated noise from boat engines (louder than you'd think due to cavitation), sonar & even nets grinding up the ocean bottom, has got to be severely cluttering up their pristine ocean umwelten. Because of the extemely powerful low frequency signals it uses, deep sea sonar alone has been linked to hunsreds of whale deaths (Sperm & Minke whales), with it thought that the powerful sounds caused caused their (relatively) delicate ear bones to disintegrate (making entire pods of whales repeatedly beach themselves). It's a damn shame that we don't have more whale song from earlier in their history - before we decimated many whale species (+), killing the singers of old songs before they could pass them on, or the young ones before they could learn them. I would imagine their "lexicon" would have been profoundly richer, with songs going back centuries but also with the addition of new components from each new generation. (+) Not that some groups (like the Japanese) ever stopped the mass killing of whales. Their harvesting of whales for meat - & especially their goddamn horrific Taiji dolphin slaughter - continue to this day. The volume of public outcry, though, has reduced to a whisper.
I think I agree with folks in this video who are more concerned about listening to the whales than talking to them. It's not about what I say to the whale, but what the whale says to me that matters.
I think that it'd be possible to construct a language model that could accurately capture the complexities of whatever language(s) whales are using to speak with each other. But just because it could 'understand' the patterns involved and produce appropriate responses to whales' sounds, doesn't necessarily mean that it can act as a Rosetta stone between our languages and those of whales. As you pointed out, there is very little that the world of whales has in common with our own. I think our best bet would be to go for what little experiences we do share as social, intelligent, long-lived mammals and build an understanding from that; Concepts like family, mother, child, food, breath, pain, temperature, sight, taste, smell, touch.
Thanks for asking the question "should we?". That question should be more widespread in scientific and technological progress. What are the potential benefits but also risks of how we present and use a specific technology.
(I haven't finished the video yet but I just wanted to say this) When you said that it was one thing for ChatGPT to help you write a title for a video but completely different for it to interpret whales, of course ChatGPT wouldn't be able to do it. But AI in general can adapt to nearly anything. Since it adapted to human language, it's reasonable to assume it could adapt to whale language which is probably not much more complex (if at all) than human language.
I hope we don’t take the AI generated language results too seriously. Based on how well ChatGPT is working currently, we’re likely to learn (incorrectly) that whales are plotting to multiply enough to displace ocean water and raise sea levels to drown the humans.
17:30 Ludwig Wittgenstein talked about this in terms of if we could converse with a lion. He reckoned even if we _could_ we would fail to understand each other. We probably see the world in such different ways.
I imagine when they hear us wondering about the fermi patadox, it would go something like this: "there's these guys in a big cylyndrical space ship who like us and they check in every ten thousand years or so, you can say hi to them in a couple hundred years". I imagine while needing more weight, a spaceship for aquatic species would be more resilient to radiation and be better at attacking things in space by using that extra mass to ram through large structures. If something whale like became a spacefairing civilization then they would probably be upset that we drove whales extinct. Sound can travel insane distances underwater, so you dont need radios in that case, and the fermi paradox could be influenced by the proportion of intelligent species who are whale like.
At 18:00, the guest, Michelle Fournet, incorrectly states that "Whales in general have been on Earth for 150 million years. They've been communicating for 150 millions years." The earliest whale ancestors diverged from the other artiodactyl mammals about 54 million years ago. en.wikipedia.org/wiki/Whale
I love the thought of us investing years of AI development into understanding their language only to learn whales exclusively make "Your mom's so fat" jokes.
The question is not what does it do for the whale. The question is will it harm them. If it does not harm them, we can explore our curiosity. If it does, we shouldn't.
To date I think the following is accurate as to the progress made: Notable Progress in AI and Whale Communication Pattern Recognition and Classification: Species-Specific Calls: AI has successfully classified and identified species-specific calls and songs. For example, AI models have been used to distinguish between the songs of humpback whales and the echolocation clicks of sperm whales. Behavioral Correlation: AI has helped in correlating certain whale sounds with specific behaviors. For example, researchers have linked certain types of clicks and calls to social interactions, feeding behaviors, or navigation. Waiting for more results soon. xx
I don't see how we'll be able to discern meaning from whale language without observing those vocalizations in context. Our current iterations of AI are LLMs, which are essentially advanced predictive data generators. So, the best AI could currently do is predict what the next whale vocalization might be, but there's no way it could derive meaning from those vocalizations simply by analyzing the audio. That's why Michelle Fournet said we need to collect data on the context in which those vocalizations were made. That's the only reason we've been able to derive meaning of body language of other animal species, because we've observed those behaviors in context.
From what I know, sperm whales make the loudest sounds. They are by far the loudest, with sounds reaching upto 238 db. Blue whales can make only sounds of about 180 db.
I was on a wähle safari last year. I tried to convince the guide that by 2030 then they don't need to look for the whales anymore, they will just ask them come near the boat and do to something for the tourist in exchange for food or so. They didn't believe that it will ever be possible.
The umwelt thing about a human way of greeting and asking another animal about how it’s doing… might work for an animal that doesn’t have an ultrasonic capability that can literally peer into the physical body and understand the physiology of what it’s encountering/examining. And by understanding how the physiological and biological processes relates to the body of the organism it’s examining in ultrasonic sensory… it may get enough information simply from that observation; that it never “needs” to ask the question. They may solely communicate about observations and needs and future plans of sorts. Their culture and language may be based around the fact that they can’t hide anything in a physical sense from each other… only deeply unique, private thoughts that are rarely sung out loud unless specifically pertinent. It’s fun to speculate.
I think we should. Absolutely. I just want to know who they are and what their world is like. I have a difficult time believing that understanding them would incentivize us to treat them with less compassion.
I think it might help a lot if one could put microphones onto a pair of a whale mother and calf and see how the "language" of the calf evolves over time.
The biggest argument for communicating with whales would be that it would finally start to deconstruct this idea that humans are the only intelligent species. We may even have answers to huge questions we have no way to answer, such as, is consciousness a human condition? There is are many ways in which communicating with animals will help humans become better habitants of earth. We need that. We seem to need a reason to stop placing human needs over those of the planet as a whole.
To answer your question if we should. Yes, unequivocally yes. How would we ever know if we don't learn to ask. We share the planet, that's a good enough reason to learn to coexist better. Communication would drastically help that along.
We kinda know about some animal language already, when a dog is howling or crying we have an idea of why he is doing it. When a cat hisses too. Problem is, we barely know what a whale is doing when it makes an specific noise.
Around 21:15 what if they're social creatures and they crave to be spoken to? Might any social creature enjoy banter at least, and tend to resent someone who just observes and doesn't communicate?
for an AI it really doesn't matter if it is human langue or whale sounds. You have to remember that chatGPT isn't made for humans in any way, it starts of completely alien and but can understand meaning out of the alien langue it have. The biggest problem is not to have a context to whale talking. it having enough whale chat, also translating it.
I do not personally care for what whales say but I do think it would be immensely useful to communicate to an entirely different species, because if we ever meet another advanced alien species it would be great if we had pre-existing tech that worked.
Roger and Katy Payne! Biggest record pressing of all time?! Umwelt. Ai Whale tales, brilliant. 150 million years of Whales communicating with each other. The change in body language of a wild bird, when you mimic their sounds, is priceless! I've had Nuthatches come up to me to see what was making the noise, the dog I was walking with became very angry! Bullfinches replying and popping out of hiding to see me. :)
This could be seriously important technology. If we can translate Whale then we would actually stand a chance translating any actual alien signals we might ever get