News for hackers. Stuff that matters.
Re:Facebook moderators in America by aboutruby on Monday February 25, 2019 @12:09PM (#19247179)
For those used to shock sites from 10-20 years ago, this is very different, I'm not going to link to anything but the disturbing content is on a few whole new levels.

I can't even imagine doing this job. This should be 100% automated with only 0.01% error rate for the first-pass filter (e.g. before it reaches humans).

They are willing to spare their users the horrors but not their own employees/contractors...

Re:Facebook moderators in America by ddalex on Monday February 25, 2019 @02:40PM (#19248676)
I developed the first version of the screening system for a corporation about 10 years ago, when UGC was still in its infancy. We had the help of a Microsoft tool that would match the visual fingerprint of a photo against a known-images database that prevented human operators from re-seeing an already tagged image in a different cloud, hoping to reduce the psychological scarring of the human workers that screened the content. We also evaluated automatic classification of images, but it's very hard to detect soft core porn from shock images even with today's algos, never mind 10 years ago.

One advice from a legal counsel that I received at that time still stands in my memory; she said: "Never look at the images, nevermind how curious you are. It's not just illegal and can land you in prison, but it's permanent. You cannot un-see something you've seen."

Re:Facebook moderators in America by intended on Monday February 25, 2019 @12:30PM (#19247404)
Facebook tried the automated system by using employees to train the machine. It failed.

They returned to non Facebook employed humans doing the moderation

Re:Facebook moderators in America by fencepost on Monday February 25, 2019 @01:49PM (#19248223)
One thing I hope they're doing is advanced detection of duplicate inappropriate content - e.g. by splitting a confirmed properly blocked video into frames and identifying matching videos based on frames matching (or even key parts of frames).

You'd still have to investigate/review partial matches, but something like that could cut down a lot on duplicate effort and could auto-identify a lot of things that would need manual review.

Re:Facebook moderators in America by rock_hard on Monday February 25, 2019 @01:21PM (#19247962)
I read somewhere that it’s actually a hybrid approach!

For some things liken nudity ML based solution work fine.

But other things such as hate speech are more nuanced and require human oversight.

That said ML is used everywhere possible to flag content to moderators.

For some categories the work of human moderators is used to train the ML models.

Facebook and other companies like YouTube are well insentivized to do the right thing here and automate as much as possible for all the reasons outlined in OP

Re:Facebook moderators in America by agentdrtran on Monday February 25, 2019 @02:25PM (#19248550)
Machines cannot do this job yet.