Most of social media is bots. News organizations didn't start putting (and that's a good thing!) or (and that's a bad thing!) in their headlines because they're afraid you might form the wrong opinion, although they are. It's to let their bots know how to pretend to feel, because they don't actually feel anything. They're still just glorified chatbots responding to keywords. They cannot handle implications, insinuations, or subtext. An unstated point will drive them insane very quickly.
That first bot got tripped up because it cannot read the text on the image, and even if it could (they are being trained for that) it could not connect "she" to "daughter" because the connection wasn't stated.