In other news, a new rant.. this time to automated AI based "moderating" stuff.
I've seen the rise of these things and are woefully inaccurate that can't handle the small details of the varying ways languages can be written to express different things.
For example. I put a review where a cable I was sold on Amazon was only working partially (aka 1 port works, not the two it's supposed to do ). Apparently the owner reported it and the AI automatically deleted it claiming it was "against" the rules.
There were no insults, no addresses, nothing out of the ordinary for a normal standard review.
And yet, it was flagged claiming it was against the Amazon rules based on "possible contact information, information not related to the product" and other things. I was flabbergasted.
Note that normally you could (before) put a dispute so an human would review this.
But they apparently have removed this button option(on the MX amazon site at least ). So you have to SUCK IT if the moderating thing got it wrong.
Similarly on how many times my comments on facebook has got me suspended for random stuff that makes zero sense while blatant (even racists) comments against me are put aside for "not breaking community standards". I notice the cynicism of Facebook automated alerts that you're suspended and claiming they are offering "unbiased moderating".
The only good changes that have been happening on facebook side is that they now have a third party review board. But facebook is making it harder and harder to make them review the content. (aka the last time I got suspended, I only got the notification on the phone, and when requesting access, it does not fill the form automatically..and if you close the screen on your phone to get the report ID. You can no longer report it to the third party board.
But yes.. more and more companies are inserting AI to everything even in things AI should never work properly.
While at the same time failing monumentally to stop what its supposed to stop .( Facebook groups being completely filled with crypto fake accounts boasting fake agents and how amazing the investments were). Ironically, manual reports to these clear bots also fail with "this comment does not go against the community".
And yes, the "disagree with decision" button does nothing lmao. They do not re-review at all.