It sounds pretty darn creepy to me. And it's sure to lead to social disasters - how can you boil down all the nuance and uniquenss of an individual human's facial expression to a smiley? It's bound to make the wrong face at least sometimes.
Is this different from using an
emoji to approximate an emotion?
If, on the other hand, my smile looks like 8-D but someone else's looks like :=) and that other guy's looks like %} - the consistency vanishes. And that's before you start factoring in people with, say, Bell's Palsy.
If it's not useful, people won't use it. If there's a miscommunication, we have ways to resolve those. I can't imagine this is something people will be forced to use, just as people aren't forced to use emoji, or markup images.
I'm guessing you can cancel and try again before sending it in that case.
When my little sister saw the demos that were on 9to5mac that was sort of her reaction. She thought somewhere cute but at the same time she thought others looked kind of creepy. She had a similar reaction to some of the giant emojis The watch could send.
At the same time I could see having a ton of fun with this. Make some sort of funny voice, choose an emoji to go with it, and send it to someone as a joke.
My guess (in this is slightly based on some of the leaks from last month) is that it doesn't do any actual representation of your face. It figures out specific faces/gestures (smiling, winking, raised eyebrow, etc.) and shows those along with maybe the direction your head is looking. I'm assuming it's not real motion capture style.
It would be interesting to see if there's a new permission they need to get access to this kind of stuff, but I haven't seen anything about that in the leaks.
I would assume trying to do this all the time would use up a ton of battery.