Unsupervised learning exists.
It could identify patterns in bird speech. It can’t identify *what* they serve, until you get some clues about what those patterns correlate to (human analysis)
Discussion
An autoencoder is the simplest example. You have two models: an encoder and decoder.
You give the encoder an input, and it produces a vector (the embedding). You then give the decoder the embedding, and expect to get the input.
You then train this model, as if the encoder and decoder were joined, and the embedding another internal layer. Goal: make the output as close as to the input.
At the end, you are left with a model that can turn data into embeddings and embeddings into a lossy version of what it thinks that embedding would correspond to.
The best part is, after it’s done, you do not need the AI. You can then reverse engineer the model into rules by analyzing inputs and outputs, and its state.
We know for sure they are communicating among each other .. many humans understand their language ! It is only a matter of putting right sort of data and people together .. I am pretty sure people are working on it .. that is why I say AI will exponentially increase the jobs ..
Imagine humans communicating with every species on planet .. I would definitely want to monitor what my dog is talking to my neighbours dog :-)