Join our growing community to challenge mainstream media bias and fix the news
Facebook claims its AI now detects 94.7% of hate posts

Facebook claims its AI now detects 94.7% of hate posts

Facebook’s CTO Mike Schroepfer announced that the social media’s artificial intelligence now detects 94.7% of the hate speech that gets removed from its platform. Comparing the figures from previous years, Schroepfer said it was 80.5% in 2019 and just 24% in 2017. Social media companies, like Facebook and Twitter, are under fire for failing to adequately curb hate speech on their platforms.

IvoryDove 2 weeks

"Russian Collusion", 100 million posts without restriction. "Hunter's laptop", shut down instantly and never saw the light of day. No wonder the CIA loves social networking.

M. Schlegel
M. Schlegel 2 weeks

Well a private company should be able to "censor" how they want. Freedom of speech is not an obligation towards any company to promote, publish or otherwise distribute your speech. I'm on the left and I don't have facebook and it's censorship rules are part of the reason why - privacy being the main reason. If you don't like what a company publishes and what not, don't be a customer of said company. That said, for companies with (near) monopole-position in the market, it would be nice to a) have a duty to give a (detailed) public reason for any banned content and b) to have some form of independent "judge" who you can call if you think that the rules have been applied unfairly. With such monopolized positions, I think we should at least make sure that they apply the same rules to everyone. If those rules in and of themselves disfavor certain viewpoints - well that's up to them. But they shouldn't allow or disallow the same or comparable "hate speech" depending on who is making said speech as long as their huge market share implicates that you don't really have the chance to "just go to a competitor". Don't know how necessary that is with facebook though, my personal impression is that it's falling out of favor and is already losing it's market share.

Shmee 1 weeks

Strange, these companies said for years that AI was rasist and could not be used to do this. They must have taught it to discriminate like their staff does.

Tevo 2 weeks

How people don't see its a good thing? sure you might not see irony or sarcastic comments, but you also will see tons of less personal attacks and bad judgement. Users yave to start writing more content backed by reasoning.

Stanley 1 weeks

Facebook flags stuff which isn't hate speech, simply because of their illiterate computers which recognize photos but not text. For instance, I shared a meme that had photos of Churchill and Hitler each giving their respective salutes, with the words 'scissors beats paper' across the pictures, and they thought that was Nazi. Likewise an anti-Nazi meme showing Goebbels with a quote from him that in propaganda, one is concerned with persuasion, not with truth. Against, obviously anti-Nazi, but Facebook just saw the photo, not the text. That's illiteracy.

Jaye Muller
Jaye Muller 1 weeks

It also detects posts that are not hatefullbas hate posts. The AI is a bit if an idiot.

Rocky 1 weeks

No such thing as a "hate post". Why do our tech overlords constantly try to criminalize legal and protected behavior. If they aren't threatening violence or advocating violence by others then there is no violation.

ben 2 weeks

Sure, Keywords like "Hunter Biden" "Russiagate" "Covid 19 Methods" have been banned and Mark's wife determines what is and is not hate speech...Hurray!

Doug Star
Doug Star 2 weeks

Is all hate speech violence inciting?

ariel 2 weeks

Noooooo.... stopped. Just cus people cuss, don't mean they are attacking people

America 2 weeks

Delete Fakebook

Steve 2 weeks

I wonder how it handles comments about beheading people.

Dan 2 weeks

Fuk Facecrook and Mark Zuk.

Benjamin 2 weeks

Can they unban me so I can test it out?

Raymond M Hein Jr
Raymond M Hein Jr 2 weeks

Yay censorship

Karl S
Karl S 2 weeks

Yay! More censorship from Facebook! Just what everyone wants... :/

S 1 weeks

Here's a hate post for you. Fü¢k you Facebook! Can you detect that?

bobby_5150 2 weeks

Hate speech = speech we don't like.

Josh 1 weeks

If you don't like how they operate delete them if everyone who complained about fb would just delete it fb would already be a thing of the past

Thrice Hated
Thrice Hated 1 weeks

Can we please get statistics on false positives? The cases when the algorithm flagged a post that in fact wasn't hate speech.

Top in Tech