News for hackers. Stuff that matters.
Re:Facebook Apologizes After A.I. Puts “Primates” Label on V... by hodgesrm on Saturday September 04, 2021 @02:24PM (#28417699)
Glad you posted. A lot of AI code simply cannot be trusted. In particularly, it's pretty obvious that the teams using it either don't do a lot of QA or don't understand how to test their systems in the first place. I find it mind-boggling that this technology is so widely used with such limited understanding.

The interesting question is how the hell do we fix this?

Re:Facebook Apologizes After A.I. Puts “Primates” Label on V... by jpambrun on Saturday September 04, 2021 @03:55PM (#28418502)
One example doesn't support your claim that QA was not sufficient. You will always get misclassifications, most won't be offensive to anyone. This is not a medical device or a self driving car, this is Facebook trying to squeeze more engagement..
Re:Facebook Apologizes After A.I. Puts “Primates” Label on V... by hodgesrm on Saturday September 04, 2021 @08:19PM (#28420654)
This type of misclassification is offensive and for good reason. Similar errors have happened regularly for a long time and in many different systems. [1]

The comments about enlarging training sets illustrate the blinders that many technical people have about this technology. There does not seem to be a notion of independent quality assurance to detect what should by now be obvious problems.

[1] https://www.bostonmagazine.com/news/2018/02/23/artificial-in...

Edit: oops, left out source

Re:Facebook Apologizes After A.I. Puts “Primates” Label on V... by jrm4 on Saturday September 04, 2021 @10:25PM (#28421196)
I have never seen a more sadly oblivious statement than this: the suggestion that Facebook squeezing engagement is significantly less dangerous than malfunctioning medical devices or self driving cars.

Facebook is more dangerous; precisely because the harm in getting it wrong is so much easier to hide.