Taste is a tricky ephemeral thing though. Even if an AI's algo can quantify taste, which it probably can (to some extent), will the taste always be backward looking?

Moreover, will AI ever grow a backbone and tell the user that the user's taste is bad? And stand firm and say, "No, we're going with my vision, it's better."

And... even if it does, what if the user is right and the AI is wrong?

So many questions to think about