I can see the capacity for this technology to do great things. I also see though where many bugs (especially undiscovered ones after deployment) can be had:
If I wear too much rouge one day, will the AI believe I'm inwardly angry and report that to my employer?
Will it give pre-mental health indicators if I keep a poker face the whole time? Will those be reported to my boss?
Will it just all out accuse me of being mad when you can clearly see I have smile on my face?
What if the the AI is programmed to detect anger in someone who expresses political dissent?
I know your post wasn't meant to address these things, so my questions are just to raise awareness. I always hope new technologies are used for good rather than for evil. My luck though? I would get falsely accused by the AI and Human Resources would believe the machine over me.
Just when I thought it might be OK to take off the mask.... :)