not sure why - just scrape all their message data and social media feeds to build up a model of each child and then jumble it up with the current work.
cameras trained on children to measure engagement could vaguely work (it'd almost certainly be crap of course).
Oh I'm with you on that bit, sure. I think the leap is then turning that into understanding why a child is engaging with the work or not, etc. People are more than just models!
absolutely - that end of it is nowhere near working and probably never will because I don't think humans would ever accept empathy from a machine. so even if it was technically possible, it still wouldn't work.
(though I do think a lot of people WOULD accept it from a machine - in fact a lot of people are willing to anthropomorphise existing models which is part of the problem. "Oh ChatGPT really gets me" etc)
Comments
cameras trained on children to measure engagement could vaguely work (it'd almost certainly be crap of course).
AGAIN: THESE ARE TERRRIBLE IDEAS