I don't know exactly what training material went into Suno's models, but if it includes random collections of text from the internet, it could very well have included "man, AI is fucking stupid and I treat it as such" in it's training datasets.

Now I won't claim that that suddenly makes the models "aware" of it, because surely we'll understand "aware" different and this will turn into a different conversations, but I don't see it as impossible that some models could have training data that includes text with how some humans feel about AI.