Earth Species Project is a non-profit dedicated to using artificial intelligence to decode non-human communication.
We believe that an understanding of non-human languages will transform our relationship with the rest of nature.
Since AI uses machine learning I see no reason why it couldn’t also learn the language of animal species. However it surely cannot be a simple as translating a foreign language - animals have such a different view of the world. How do you understand a rhino who’s world is dominated by smell - how do you interpret its concepts? If this actually happens it will transform humans relationship with nature, potentially increasing our ability to empathise with animals. And give hunters a better tool to hunt with…
I’m sure a LLM-type generative AI will produce some kind of structured model that will mimic how members of different animal species interact with one another vocally. (Whether animals have language or not is another question of both definition and degree.) How that could possibly relate to mental constructs of the animals (insofar as they have conceptual models that are expressed in the abstract) is in question, since no animal species has provided a vast training set of symbolic information that relates concepts in a structured fashion, nor would a generative AI have any perception of the perceptual experience of an animal.
I have found with Bayliss, poop (button) does not always mean he needs to go outside to take care of business. Sometimes it means “bark” for nothing in particular.
