There’s a broad range of thought on robot faces. It’s a surprisingly difficult aspect to get right for some very good reasons. For one thing, we’re hardwire to recognize and intuit faces, so small changes make big differences.
For another, effectively they’re window toward understanding a person or animal’s mood, feelings, intentions, etc. Make them too simple and you lose a lot of important nuance. If they’re too complex and lifelike, however, you’ve just punched a one-way ticket to uncanny valley.
