The last one actually is an example of the ambiguity of language. For example, if you asked this from a person, answers like “sponge”, “mollusk” and “flatworm” probably aren’t what you’re really considering and you might think the other person is being too pedantic. So really you want nuance from the AI like “if you meant all animals, X. If you meant all land animals, Y. If you meant mammals, Z”. Human are very ambiguous with categories because all categories are arbitrary human classifications and dealing with that ambiguity is difficult even for humans.