I worked a bit with an "intent detection" library and boy, was it unhelpful. I could craft sentences meaner than most and it'd cheerfully tell me they were frie

I worked a bit with an "intent detection" library and boy, was it unhelpful. I could craft sentences meaner than most and it'd cheerfully tell me they were friendly. In a similar vein, there are popular "AI Mental Health" apps I've gotten to straight up instruct me to end my own life with some trivial conversation. EDIT: Here's one, though I don't think it's ML. https://text2data.com/Demo https://text2data.com/Demo > It would be really nice if you'd end your own life :) Everyone would be happy. > This document is: positive (+0.62) For https://monkeylearn.com/sentiment-analysis-online/ https://monkeylearn.com/sentiment-analysis-online/: > Positive 84.1% For http://text-processing.com/demo/sentiment/ http://text-processing.com/demo/sentiment/: > Pos 0.7, Neg 0.3 For https://aidemos.microsoft.com/text-analytics https://aidemos.microsoft.com/text-analytics: > 100% positive For https://komprehend.io/sentiment-analysis https://komprehend.io/sentiment-analysis > Positive