Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Facebook Censorship Communications Social Networks The Internet

Facebook's Uneven Enforcement of Hate Speech Rules Allows Vile Posts To Stay Up (propublica.org) 171

ProPublica has found inconsistent rulings on hate speech after analyzing more than 900 Facebook posts submitted to them as part of a crowd-sourced investigation into how the world's largest social network implements its hate-speech rules. "Based on this small fraction of Facebook posts, its content reviewers often make different calls on items with similar content, and don't always abide by the company's complex guidelines," reports ProPublica. "Even when they do follow the rules, racist or sexist language may survive scrutiny because it is not sufficiently derogatory or violent to meet Facebook's definition of hate speech." From the report: We asked Facebook to explain its decisions on a sample of 49 items, sent in by people who maintained that content reviewers had erred, mostly by leaving hate speech up, or in a few instances by deleting legitimate expression. In 22 cases, Facebook said its reviewers had made a mistake. In 19, it defended the rulings. In six cases, Facebook said the content did violate its rules but its reviewers had not actually judged it one way or the other because users had not flagged it correctly, or the author had deleted it. In the other two cases, it said it didn't have enough information to respond.

"We're sorry for the mistakes we have made -- they do not reflect the community we want to help build," Facebook Vice President Justin Osofsky said in a statement. "We must do better." He said Facebook will double the size of its safety and security team, which includes content reviewers and other employees, to 20,000 people in 2018, in an effort to enforce its rules better. He added that Facebook deletes about 66,000 posts reported as hate speech each week, but that not everything offensive qualifies as hate speech. "Our policies allow content that may be controversial and at times even distasteful, but it does not cross the line into hate speech," he said. "This may include criticism of public figures, religions, professions, and political ideologies."

This discussion has been archived. No new comments can be posted.

Facebook's Uneven Enforcement of Hate Speech Rules Allows Vile Posts To Stay Up

Comments Filter:
  • censor all the things
  • by AHuxley ( 892839 ) on Friday December 29, 2017 @08:02PM (#55831805) Journal
    The US university system graduates so many SJW every year. They would enjoy the feeling of power to censor the internet for a social media company.

    Just list all the terms, music, art, culture, music/movie reviews, cartoons, blasphemy to be de ranked, banned and removed.

    SJW can also report users, accounts, art work, history to EU and US law enforcement too.
    Just hire a lot more SJW and let them censor social media.
    Social media's got what governments crave. They crave censorship. It's got social justice.
    • Re: (Score:1, Insightful)

      by Anonymous Coward

      To be fair SJWs worming their way into worthwhile positions are few (compared with how many get their "degrees".) Most end up serving coffee to people of actual worth, or writing fake news for free as interns while living in the places their hard working moms and dads' bought or rented for them.

  • by Anonymous Coward on Friday December 29, 2017 @08:03PM (#55831817)

    Anything else means that you're getting in the way of somebody's Freedom of Speech.

    In fact, it might be nice to know that Fred Bloggs can't go three posts without using the N-word.
    It will inform me when I'm making decisions about who to invite to a party, recommend for a job opening, etc.

    • by Anonymous Coward

      Agree 100%. It's really unfair to SJWs how this thing works. Their speech gets left alone and so employers can only filter them out and not the nazies that get censored.

    • "Remove illegal"

      I'm sorry, what?

    • And yet, posts on Facebook don't qualify for any Constitutional "freedom of speech" protections. Facebook wasn't formed by Congress, and that is the entity referred to in the Bill of Rights. They don't even receive and government funding, so technically they can censor anyone they want. That would probably be a really bad business decision, but totally legal.
      • by jon3k ( 691256 )
        At what point do we admit that Facebook (which also owns Instagram and WhatsApp) has a monopoly on social media and require government intervention? My (probably naive, incorrect) understanding of the laws in US vs EU is that in the US they must abuse their market position to harm other companies but in the EU merely having a monopoly could invite government intervention.

        Personally I couldn't care less, I'm 35 and I haven't logged into Facebook in years and everyone younger than me is using it even less
      • by RedK ( 112790 )

        And yet, posts on Facebook don't qualify for any Constitutional "freedom of speech" protections.

        He said "Freedom of Speech", not "1st Amendment rights". You silly people are so quick to say "Private entities don't have to support Free Speech!" based only on the Constitution. It's almost like you hate actual free speech.

        Freedom of Speech is an old concept, dates back to Ancient History. The question is, do we as consumers value it enough to force our service providers to adhere to it through our wallets, and should service providers recognize it and support it on their platforms as a selling point ?

        • Well, seeing as Facebook is a "free" product to the general population, that makes it pretty much impossible to use the wallet to force them to do anything. They "sell" their user's data to advertisers...in that regard, this "hate speech" is just as valuable as any other as it helps potential advertisers narrow down their target market promotionals. As someone who has taken several college-level marketing and promotion classes, I think that this is the real reason why FB isn't hard-core in their "policing"
    • Maybe, just maybe, the concept of "hate speech" is so vague that no two people will ever agree on what it constitutes, so nobody should be removing anything for violating it? Nah, that's just crazy talk, take down everything that somebody disagrees with.
    • "N-word" is such a cop-out. Just say n*gger (thanks, Slashdot, for forcing censorship!)--you're already implying it to everyone that reads your comment.

      Also a big thanks to /. for their "lameness" filter automatically blocking my submission. Those scary words sure are better off blocked! Someone might be offended, after all, which is literally the worst thing that can happen to a white, upper/middle-class woman.
  • by HiThere ( 15173 ) <charleshixsn.earthlink@net> on Friday December 29, 2017 @08:09PM (#55831837)

    I'm not saying that Facebook is making a good faith effort to solve the problem. I've never looked at it, so I have no idea. I'm saying that the problem as stated is a hard problem. It's easy to, say, ban certain particular words, but that doesn't accomplish very much.

    I'm not sure that the problem as stated could be addressed by anything much short of a human equivalent AI, and even that would only allow some particular set of standards to be applied uniformly. It sure couldn't guarantee that the standards were fair.

    As an example consider the text "You with a donkey's member!" This is apparently a violently abusive comment, but that depends on context that isn't present. I'm sure I could come up with a context where that would be encouragement. And every single word in that sentence is perfectly harmless. Or what about "Pepe the frog"? That was intended to be a humorous children's cartoon character....but it didn't stay that way, much to the annoyance of the creator.

    That said, the evidence seems to support the assertion that Facebook encourages hateful posts, and is more reluctant to censor nazi-ish posts than those with an opposing message. Again, I have no direct evidence for this as I never visit that site, and am relying on material published by others.

    • As an example consider the text "You with a donkey's member!" This is apparently a violently abusive comment,

      Proof positive that the snowflake generation needs to lighten the hell up. Violently abusive? Really?

    • by AmiMoJo ( 196126 )

      Facebook doesn't seem to have any particular political bias. TFA says they removed a comment stating "men are trash", for example.

      • TFA says they removed a comment stating "men are trash", for example.

        Ha! I recently got blocked by a woman in a polya group shortly after telling her that it wasn't just men running out on families any more (even though mine did) and immediately after she told me "men are trash it's not my fault". I hope that's the comment they are talking about :)

    • > "You with a donkey's member!"

      It's nice that I give that impression, thank you,. but it's really not that big...

    • The problem with AI is someone has to program it, program the patterns for it to recognize. It's already been shown that many algorithms are developing racial biases. So most likely an AI would only go much faster, but not be much more accurate...at least at first.
      • by HiThere ( 15173 )

        If it's an actual AI, the bias is not caused by the programming, but rather by the training data set. As to whether it would be more accurate, that's a separate problem from either consistent or fair. My assertion was that it could be more consistent, but that this wouldn't imply it would be fairer. Accurate is a yet more difficult problem when dealing with people, and it's one that human level intelligences have yet to solve.

        To be fair, "fairness" is only a solved problem in the opinion of the person do

  • by Anonymous Coward

    Arbitrary, subjective rules can never be enforced fairly. Create more objective criteria, or stop moderating altogether and let people do their own blocking like adults.

  • It means that they can filter whatever they want and blame it on the "automated" filter....

  • I miss Usenet (Score:5, Insightful)

    by mrsam ( 12205 ) on Friday December 29, 2017 @08:15PM (#55831877) Homepage

    There was nothing comparable to that -- no bureaucracy that needed to employ a small army to act as a thought police, enforcing vaguely-defined thoughtcrime. There were just a few, content-neutral rules one had to follow, to post on Usenet. You were free to write anything you wanted, no matter how vulgar or obscene. Complete and unrestricted freedom of speech. You could not be silenced. When some snowflake or a SJW got triggered, too bad, so sad. They could do nothing about it. In its heyday, I had a blast of a time trolling the snowflakes and giving them daily aneurisms. I miss those days.

    Of course, Usenet's still around, if one knows where to find it. And, come to think of it, I think I will. The riff-raff, the millenial snowflakes can have Faceboot, Twatter, and the rest of that junk. They should stay off Usenet. They wouldn't be able to handle it.

    • by Anonymous Coward

      They are all dying.

      I have been poking around into a few communities I used to visit or knew people in, and visiting a few communities I didn't visit back when they were fresh and new. Almost all of them have waning activity. The largest active IRC now seems to be freenode, and mostly in the developer channels, combined with #hardware. 2600's irc is dead. IRC2P on I2P has about 20-30 regulars, spread across the i2p developer channels, #salt, and the russian/chinese channels and is otherwise dead. MU* communi

    • Re:I miss Usenet (Score:4, Interesting)

      by Pfhorrest ( 545131 ) on Friday December 29, 2017 @10:39PM (#55832377) Homepage Journal

      You could be "silenced" on UseNet, with a killfile. But the thing was, killfiles were user-specific. Each user could decide who they wanted to "ban" from the discussion, and to that user, whoever they wanted gone, was gone. Any nobody else had any idea.

      I don't see why modern discussion systems can't implement something like that. Allow people to share "kill-lists" too, so you can get a list of all the known [the-other-side] trolls you don't want to deal with from other people who've had to deal with them. Everyone gets their own personal shadowban button, but it only shadowbans people from that user's perspective. Eliminate anyone you want from your view of the discussion and build your own personal filter bubble if you want. Nobody else will be any the wiser. Nobody gets censored, and anyone who wants to see the shitstorm of unfiltered content can, at their own expense and nobody else's.

      • by AmiMoJo ( 196126 )

        Block lists have been tried and don't work. They are too easy to circumvent and don't deal with viral posts well. Every user has to build their own from scratch, or if someone shares their list it gets instantly condemned. Then you end up with shitty "internet security" products promising to automate the block list for you, and failing.

        From Facebook's point of view they want people to have a good experience. Telling them that they have to manually block the bullshit isn't exactly a great marketing strategy.

        • Facebook (et al) can themselves offer prebuilt block lists of known-bad actors (like they already build), and allow people to choose between them, or choose none of them, and of course to make their own modifications to (their copies of) the prebuilt lists once they're picked. It's just like what they're doing now, except you can opt out of it if you want, and also opt into alternatives, and custom configure your own alternatives if you want. The default block list would just be the same list of people curr

      • Killfiles are not a panacea. The person in question can still reply to posts and everyone else sees it. This leaves open a situation where people who don't know either party read the posts and think that the person in the killfiles made a point to which you have no counter-argument, etc.
        • In this case facebook can do better (while they don't). If facebook users can block someone, with a tick of saying, "I want my blocking of this guy publicly known", then for trolls there will be a tag which people can see oh that guy has been publicly blocked by hundreds or thousands of people. But no, facebook choose to let users to only block/flag/report someone silently, and hold the power to stamp someone "hate speech" solely on the hand of facebook employees.
          • That is also a good idea, but the improvement on usenet killfiles I thought you were going to suggest it: modern systems like Facebook are better able to track what is a reply to whom than usenet was, so if you block a user, you can also (optionally) block all replies to that user as well, so you don't see one side of an argument against the trolls, you just don't see the argument at all.

    • Usenet posts could be canceled [wikipedia.org]. That was a message distributed the same way as Usenet posts directing servers to delete another Usenet post (usually after discussion in one of the net-abuse groups). Only admins were authorized to send a cancel notice, and the admin of an individual server didn't have to set it to honor cancel requests (most didn't). But it's not exactly true that Usenet was complete and unrestricted freedom of speech.
    • by ncc74656 ( 45571 ) *

      Maybe I shouldn't post this here where the snowflakes can see it, but you can get free accounts here [eternal-september.org]. Point your newsreader to their servers and you're in business. No binaries, but if you want those, there are lots of paid services offering them.

      I even have an ebuild for trn in my Portage overlay [gitlab.com], if you're using Gentoo. Builds and runs like a champ on x86 and AMD64, at a minimum.

  • by RightwingNutjob ( 1302813 ) on Friday December 29, 2017 @08:29PM (#55831919)
    One left-wing troll story after another. Are you retards trying to hit your quote before the year is up or are is the supervision on their Christmas vacation this week?
    • Re: (Score:1, Troll)

      by AmiMoJo ( 196126 )

      So you would prefer that people can just post what they like and others deal with it... Except "left wing troll" stories that upset you.

    • by ncc74656 ( 45571 ) *

      Are you retards trying to hit your quote before the year is up or are is the supervision on their Christmas vacation this week?

      s/Christmas/winter/

      That's how they'd want it, after all. :-P

  • by Patent Lover ( 779809 ) on Friday December 29, 2017 @08:40PM (#55831961)
    How scientific.
  • by SlaveToTheGrind ( 546262 ) on Friday December 29, 2017 @09:11PM (#55832075)

    have been directed toward conservatives or others who don't mindlessly toe the party line. Strangely, those all seem to stay up. If you want to talk about uneven enforcement, how about starting there?

    • I want to add that if anyone finds their opinion exactly matching either party platform, then you are either demented or have some kind of Stockholm syndrome. The party platforms of both parties were made by a series of compromises, designed to appeal to a broad spectrum of ideologies, change often, and thus can be contradictory from time to time.

      Anyone who strictly follows either party platform is a tool.
    • Conservatives, by definition, "toe the party line". To oppose it would be iconoclastic, and conservatives are anything but that.

    • It's a sad truth that reporting accurately and objectively on the activities of conservatives is often perceived by them to be "vile" and unfair. They tend to get upset when the happy face they try to paint on what they are doing is allowed to stand unchallenged.

      Examples of this unfortunate situation are legion. I can go all the way back to when Ronald Reagan's administration tried to avoid meeting even the minimal nutrition requirements of a school lunch program by re-defining ketchup as a vegetable. Th

  • I had to look that up. Seems to exist, so don't hate me for it...

  • Or has the comment section here been more cesspool-y in the last few weeks?

    • It's you, Hentai007.

      But seriously, don't want to see people fight back SJWs, don't post SJW stuffs. Have you noticed that just 1 or 2 slashdot users are responsible for 100% of the SJW posts?

      • by RedK ( 112790 )

        Have you noticed that just 1 or 2 slashdot users are responsible for 100% of the SJW posts?

        They're easy to spot too. They actually use the Slashdot Friend/Foe system and so their posts have a little Yellow dot next to their names for me.

        I dunno what was more surprising, seeing that thing being used, or the fact that people I had no idea even existed had found me important enough to flag.

      • Eh there is a difference between âoefighting sjwsâ and calling for the extermination of people... it was like a few threads in a row and I wondered if the 8chan crowd was hitting /. For some reason...

        • What do you mean "calling for the extermination of people"?
          Like those
          "Let's band together to kill all men"
          or
          "All I want for Christmas is white genocide"?
          that proliferate mainstream academia/media?

          If so, then golly, you are right and I must have missed them since I browse at +1.

          How I miss those days when not everyone is a Nazi... where there are hot grits everywhere, frosty piss in all your base, and the Soviet Russia, caveman Ogg were all still alive.

    • It is you. It has been more "cesspool-y" for much longer than that. You are evidently just noticing.
  • This kind of thing is inevitable when you try to police free speech.

    Don't like it? Tough.

  • by Mr307 ( 49185 ) on Friday December 29, 2017 @10:02PM (#55832261)

    According to these idiots insults are hate speech:
    http://www.dictionary.com/brow... [dictionary.com]
    noun
    speech that attacks, threatens, or insults a person or group on the basis of national origin, ethnicity, color, religion, gender, gender identity, sexual orientation, or disability.

    Websters seems to have it simplified down to a literal state which could be fine:
    https://www.merriam-webster.co... [merriam-webster.com]
    Definition of Hate speech
    : speech expressing hatred of a particular group of people

    Wikipedia is all over the map but at least seems to only report on various countries:
    https://en.wikipedia.org/wiki/... [wikipedia.org]

    These people are subjectively confused thinking "any form of expression regarded as offensive":
    https://definitions.uslegal.co... [uslegal.com]
    Hate Speech Law and Legal Definition
    Hate speech is a communication that carries no meaning other than the expression of hatred for some group, especially in circumstances in which the communication is likely to provoke violence. It is an incitement to hatred primarily against a group of persons defined in terms of race, ethnicity, national origin, gender, religion, sexual orientation, and the like. Hate speech can be any form of expression regarded as offensive to racial, ethnic and religious groups and other discrete minorities or to women.

    These people get it:
    https://www.urbandictionary.co... [urbandictionary.com]
      Hate speech
    A highfalutin' way of saying "I disagree with your meticulously-researched, irrefutable facts, so I am going to organize a social media campaign to demonize you and ruin your life. But don't forget to donate to my Patreon."
    Sane, rational human being: "I sure do loves me some grapes!"
    Filthy SJW bacterium: "OMFG GRAPE HAS 'RAPE' IN IT THAT'S HATE SPEECH! RAAAAAAAPE CULTUUUUUURE!"

    Disparaging a social group is hate speech to these people:
    https://www.thefreedictionary.... [thefreedictionary.com]
    hate speech
    n.
    Bigoted speech attacking or disparaging a social group or a member of such a group.

    • by LynnwoodRooster ( 966895 ) on Friday December 29, 2017 @10:35PM (#55832363) Journal
      Who knew that an entire movie is hate speech [imdb.com] based upon its insults laced throughout the movie...
    • by Subm ( 79417 )

      > Websters seems to have it simplified down to a literal state which could be fine:
      > https://www.merriam-webster.co... [merriam-webster.com]
      > Definition of Hate speech
      > : speech expressing hatred of a particular group of people

      This definition says that when Indiana Jones says, "Nazis, I hate these guys," the movie is hate speech.

      Is that what we want?

      • by Mr307 ( 49185 )

        Well as a simple and seemingly correct definition I think its fair.

        But in no way shape or form would I ever advocate for it to have a law associated with it, its all too subjective, which was the point of the original post.

        Short of that very literal one, I dont believe there is a reasonable definition of 'hate speech', it would appear to be a tool mostly used to silence discussion, probably for political or dogmatic reasons.

  • Twitter's fallacy here -- and the one which powerful forces are trying to foist upon other big names -- is that you can have a global communication platform with one set of rules. There is no one set of rules that refers to all speech. There is no way to measure "hate" fairly. Twitter has painted themselves into a corner: they will forever be too strict, or too lenient, often being loudly accused of both at the same time.

    • by AHuxley ( 892839 )
      They should have kept their US branding globally.
      Freedom of speech, freedom after speech sells around the world.
      People who need to interact with their own governments can use boring national sites.
      Keep social media fun, free and full of different news.
      Thats what attracts people and profits. Users can all get 100% censorship in their own nations for every day.
      US social media does not need to report users to they own nations police for enjoying US freedom of speech.
  • One of the problems with filters is something I see on other sites, especially on political issues. Things escalate. When a post is finally flagged, often it's in response to a post that's similar or worse. Sites have trolls using moderation to eliminate opinions they don't like. If you just mildly insult someone who holds an opinion you don't like, then if they respond in kind, report them.

    Deep Learning should be used to identify those who make the most complaints, or have a disproportionate number o
  • by MSTCrow5429 ( 642744 ) on Saturday December 30, 2017 @01:17AM (#55832821)

    I'm old enough (i.e., not middle-aged yet) to remember when "hate speech" was on the anti-free-speech fringes. Then it started moving in on us. And the closer and closer it got, the greater and greater amount of largely anodyne words and thoughts became verboten. Now some claim that certain words or thoughts are equivalent to physical violence. It was better when people were just seen as obnoxious dicks, and not some form of Neo-Grammar(ish) Nazi guilty of breaking windows of the mind.

    • by mjwx ( 966435 )

      I'm old enough (i.e., not middle-aged yet) to remember when "hate speech" was on the anti-free-speech fringes.

      I'm old enough to remember when the old racist brigade was nought but a small group of sad men sitting in the corner of the local pub. They sat their in their corner muttering about "them" ("them", the target, changed over time, first it was Asians, then Arabs, then Muslims, however the hate always remained the same). They'd never venture out of their corner because no-one put up with their bullshit, you'd note that none of the local racist brigade had all of their teeth because back in those days if you we

  • Comment removed based on user account deletion
  • Truth (Capital "T" Truth) and groupthink don't mix. Humans WILL lie to themselves about anything and everything. And we get very upset when reality refuses to follow our whims. And why, oh why, are you socializing with people you don't like on a platform that doesn't like you?
  • ...Setting the standards for double-standards!

Do you suffer painful hallucination? -- Don Juan, cited by Carlos Casteneda

Working...