Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×
Twitter Social Networks

Twitter Will Auto-Block Harassers in Bid To Curb Abuse (bloomberg.com) 88

Posted by msmash from the moving-forward dept.
Twitter debuted a new product called Safety Mode that will automatically block users who are being aggressive or hateful toward another person in an effort to help reduce harassment. From a report: The social networking company said it will use automated technology to look at the content of a tweet and "the relationship between the Tweet author and replier" to determine if a block is warranted. Twitter previously introduced the feature during an Analyst Day presentation in February. If the company detects that one user is "using potentially harmful language -- such as insults or hateful remarks -- or sending repetitive and uninvited replies or mentions," it will automatically block the offending user on behalf of the targeted person, Twitter said Wednesday in a blog post. Automatic blocks last for seven days. The goal is to stop "overwhelming and unwelcome interactions that can interrupt" a user's experience, the company said in the post.
This discussion has been archived. No new comments can be posted.

Twitter Will Auto-Block Harassers in Bid To Curb Abuse More

Twitter Will Auto-Block Harassers in Bid To Curb Abuse

Comments Filter:

  • Halfway to hellban. (Score:4, Interesting)

    by Gravis Zero ( 934156 ) on Wednesday September 01, 2021 @01:53PM (#61753109)

    I think this is great but I don't think they are quite there. Instead of blocking a harasser, it should detect the harasser and isolate them and their posts like a shadowban. However, instead of leaving them alone, they should instead employ AI to appear to be the person who was harassed and respond with overt aggression until the harasser finally backs off. After which, it should be revealed which posts were actually generated content, so that the harasser understand they were fighting with a machine and not the person. This completely isolated AI driven aggressive response is what I call a "hellban" because there is no winning in a fight against a machine.

    • Unfortunately, even then I doubt they'd learn the only winning move is not to play.

    • Re:Halfway to hellban. (Score:4, Interesting)

      by t0qer ( 230538 ) on Wednesday September 01, 2021 @02:19PM (#61753241) Homepage Journal

      Generally speaking I think it's stupid. Let me using facebook as an example.

      I get into some pretty heated debates on my page. All someone has to do is use a term facebook deems as harassing, and someone gets hit with a ban. A guy polar of my political beliefs called someone a "Commie" on another page, and had a 2 week ban. Currently I'm serving a 30 day ban for linking to this tweet/video. [twitter.com] Facebook and twitter seem to want to be the nice safe space for boomers with no name calling of any kind, no matter what political ideology you identify with. Here's the thing though...

      The internet is for porn.

      OK maybe that's a bit more overgeneralizing it, but I remember long flame wars that lasted years on newsgroups. By moderating it, you become the bad guy. By not moderating it, sure there's some ugliness but eventually one side or the other tires out. Part of why people gravitate to online discussion is for what I call "The Geraldo Effect" or "The Jerry Springer Effect" (for those too young to remember Geraldo's talk show)

      People want to see online bickering, it's in our nature to slow down and rubberneck at accidents. All twitter and Facebook are going to do is get people to want to move off them.

      • People want to see online bickering, it's in our nature to slow down and rubberneck at accidents. All twitter and Facebook are going to do is get people to want to move off them.

        And nothing of value would be lost.

        • Re: (Score:2)

          by t0qer ( 230538 )

          One of the early things myspace found out was if people start off with friends, they're likely to stick around. Are people going to stick around a place that excludes friends? Maybe not everyone, but there will be some. That's $14@year per user that facebook loses. So there is some value being lost there.

          • Value to Facebook or Twitter is almost directly proportional to the negative impact they have on society. If they disappear, I'll repeat that nothing of value would be lost. And we might actually have a net gain once the dust settled.

      • Allow me to reignite a long standing flame war.
        Emacs sux donkey dix you turd! vi FTW!

        sorry, couldnt resist.

        • Re: (Score:2)

          by t0qer ( 230538 )

          I really miss newsgroups, I also miss BBS's.

          I'm old, not with all this newfangled pronoun stuff. Maybe my preferred pronoun is "root"

          (and VI is awesome)

      • Re: (Score:2, Troll)

        by account_deleted ( 4530225 )
        Comment removed based on user account deletion

        • Re: (Score:2)

          by t0qer ( 230538 )

          Ya you get exactly what I'm getting at here. Once the algo's start discouraging people from participation, people won't participate be it their friends head for greener pastures or they get sick of it themselves.

        • Comment removed based on user account deletion

      • Re: (Score:2)

        by Toad-san ( 64810 )

        I found Geraldo offensive (and often stupid). I always found Jerry Springer offensive, and his guests the scum of the earth.

        I don't slow down to rubberneck at accidents (unless I'm slowing down to stop and help, and there's no one else already there doing that). I dislike real-life bickering, considering it "poor form" and definitely not an example of civilized behavior. So I certainly don't want to see it online.

        I don't consider _this_ bickering, because I'm never coming back to this Slashdot article.

    • However, instead of leaving them alone, they should instead employ AI to appear to be the person who was harassed and respond with overt aggression

      I don't think purposefully making people madder and more unstable is ever a good idea, and could lead to real life deaths (either suicide or murder) in the worst cases...

      • The problem with doing nothing is that it only encourages bad users to continue. The problem with bans is that bad users create new accounts. The only remaining option is to create a virtual environment where they cannot interface with others but are unaware of that fact. Therefore to keep them in the dark, you must engage them with negative feedback (lest positive feedback encourage the behavior). If people are a danger to themselves or others then they should be committed.

        • The problem with doing nothing is that it only encourages bad users to continue.

          Continuing is great because it just means they are typing away, to the detriment of nobody.

          Goading someone into physical action is far worse then them "continuing". It wouldn't goad everyone to action but if even just 1% of people either killed themselves or turned to murdering others, your plan is monstrous.

          Would be far superior if instead of angry hate-forging responses the automated system tried to reason with them using ve

          • Goading someone into physical action is far worse then them "continuing". It wouldn't goad everyone to action but if even just 1% of people either killed themselves or turned to murdering others, your plan is monstrous.

            It's not the plan that is to blame, it is the humans themselves. Humans are monstrous, machines just compute.

            Whatever happened to being responsible for your own actions?

            • Whatever happened to being responsible for your own actions?

              Why not explain that to us all, since your desired action is a plan to manipulate humans into being worse? You seem to be ignoring your own responsibility pretty aptly.

              • The plan isn't to manipulate them into being worse, the plan is to provide negative feedback for harassing others.

        • Making that determination is extremely difficult and similar systems have already been shown to get abused. Meanwhile people can spiral out of control in the course of mere weeks. It just takes a series of events coupled with external hostility related, or even unrelated, to those events. A clever person will know they are talking to a psychological sysop. We havent actually passed a Touring test have we?
      • Yea Its safe to say the majority of people that went postal were triggered by a recent event. What these people really need is a nice long 30 day caribbean cruise with no outside access. Just beaches, pools, recliners, and wine. Lots and lots of wine. Hey! I need one of those! >:-/

    • This completely isolated AI driven aggressive response is what I call a "hellban" because there is no winning in a fight against a machine.

      Sarah Connor would beg to differ.

    • Re: (Score:2)

      by Hentes ( 2461350 )

      Shadowbanning is a tool of psychopaths. Trolls can easily check if they were shadowbanned from a second account, while people banned by accident (and as with any other automated moderation tool, the number of false positives is going to be high) will lose any chance of getting their account back.

      • Trolls can easily check if they were shadowbanned from a second account

        Precisely why you want to keep them engaged. The point is to not let on that they have been banned at all. This occupies their time while protecting your users which is the point of the hellban.

    • Interesting, but it wouldn't work with Twitter. Twitter is public; it takes one incognito browser tab to see that your posts are blackholed. And I would definitely object if the AI makes offensive posts with my name on it. The other party could screenshot them and I would need to defend myself with "I didn't write that" .

  • Politicians are legally required to be excluded from this since they cannot block people
    • If you are saying that politicians are legally exempt from being banned on Twitter, you are sorely mistaken. Politicians still block people. They shouldn't be allowed to block their constituents. If they didn't want people to contact them on Twitter then they shouldn't be on Twitter or they shouldn't be politicians.
  • It’s interesting that technology meant to allow communication now censors it, thanks to monopoly like power websites with large audiences have. Blocking repetitive posts sounds great by itself and should be done anyways separately. However, this safety mode must be at the option of the specific user only and their choice. Harmful language is very different between people, someone may either have thick skin or be a snowflake. This sounds great for accounts that are basically done growing but for any no

    • I'm mostly curious about friendly banter between like-minded people. Every group has their version of the mom jokes. How many, "With your mom" or "yo momma so fat" jokes before the AI thinks it's abusive even if the recipient is laughing their ass off? These automated systems thus far haven't proven smart enough to tell the difference between friendly slagging and downright harassments. And pardon me if I don't believe the company would give enough of a shit about it to fix it when it flags false positi

  • AI screw shit up yet again... [theverge.com]

  • There goes all the blue checkmarks. (Score:1)

    by Anonymous Coward
    Oh wait, these authoritarian idiots are exempt from the Twitter's arbitrary rules. Twitter is its own eco system of stupidity. It should scare everyone here, that the so called elites -- more so Fascists -- that have stolen power, think it reflects the whole of society.

  • And the Twitter slide into irrelevance continues (Score:5, Insightful)

    by memory_register ( 6248354 ) on Wednesday September 01, 2021 @02:45PM (#61753373)
    If you start banning everyone you dislike for vague reasons, pretty soon you get a boring echo chamber.

    • Re: (Score:1)

      by AmiMoJo ( 196126 )

      Nobody else has managed to do better than Twitter so far. Most of the attempts have turned into sick jokes, like Parler and Gab.

      Far from being irrelevant, Twitter remains the premier micro blogging service.

      • This presupposes that a microblogging paradigm is strictly necessary to continue, but I have not seen evidence for this. Outside of the US and Japan, Twitter is basically a non-entity. If Twitter uses the banhammer to destroy the US community / make an echo chamber, it could tank their business.

    • If you start banning everyone you dislike for vague reasons, pretty soon you get a boring echo chamber.

      Twitter's most active period of banning ever has been in 2020, and their user base rose during it and continues to rise now, even with the "boycotts" called by the moronic right during the last moments of the Trump lunacy. They are anything but irrelevant, because it may absolutely shock you but people generally don't like playing in a platform toxic cesspool of shit.

      You're living in some kind of fantasy world if you think curbing abuse will negatively affect them.

      • Nice bait and switch with 'curbing abuse' instead of having the courage to call it censorship.

        And I don't expect there to be short-term consequences, just like I don't expect Amazon to die tomorrow even if they treat their workers like shit. It's just unsustainable long term.
      • For further reading: Twitter was already on a slide before the election due to censorship and mismanagement. There is every reason to expect this trend to return this year and remain for some time. https://backlinko.com/twitter-... [backlinko.com]

        • Twitter was already on a slide

          I think you meant to link to some fake news story rather than statistics which instead show that twitter wasn't on a slide and showed year on year growth from 2018, 2019, and 2020. Maybe read the link you posted.

          • Go down to the chart - slide in DAU and MAU in 2018 & 2019, then an election/covid bump in 2020. I'd be interested to see how '21 treats them.

  • how long before everyone is blocked? (Score:5, Interesting)

    by MNNorske ( 2651341 ) on Wednesday September 01, 2021 @02:48PM (#61753393)
    How long before this process has auto-blocked everyone from everyone else?

  • "using potentially harmful language" (Score:3)

    by Oligonicella ( 659917 ) on Wednesday September 01, 2021 @03:02PM (#61753475)
    Well, *that's* a pretty concise definition.

  • Unintended consequences (Score:3)

    by JeffOwl ( 2858633 ) on Wednesday September 01, 2021 @03:05PM (#61753491)
    It will depend on how aggressive this AI is at suspending people, but I think the likely outcome is is further isolation of people into their respective echo chambers to the point where they think their ideas are closer to being widely accepted than they are. Which will lead to further polarization of those who continue to participate.

  • Why the hell else would I use Twitter?

  • You could already get auto locked for using naughty words while @ing a checkmark or old account over a certain follower threshold.

    Found this out the time I replied to one of my friends with a joking "if you don't shut up I will smack the stupid right out of you, bitch" and was locked within ten minutes.

    The media has really done a shockingly horrible job covering Twitter in general over the years. Features which have been implemented for years getting treated as brand-new revelations, the constant stories ab

  • China's been doing that for a while now. On phones too.

  • "Biden is an idiot". You are blocked for harrassment.
    "Trump is an idiot." Recommended as insightful statement.

  • I see complaint after complaint from people reporting tweets where someone has threatened them, said that they should be raped, said something anti-semetic, etc., and Twitter saying that those tweets don't violate community guidelines. I don't care if you think that death threats aren't harassment because the internet something-something, but some people legitimately have cause to fear for their lives when threats like that are made, so it's easier to just ban those types of tweets. But Twitter reliably do

  • If they're going to block, *how* are they going to block?

    I should be finishing my one week Fecebook jail... someone posted a screen capture of some idiot on twitter, with "I can't believe this idiot". I agreed, and posted, addressing it to the screen capture, *not* the poster, "you're a slut, John".

    Apparently, ancient SNL lines that everyone knows is naughty.

    If the twits want to ban, then I have for them the same question that I have for faceplant: who is this "community" that has standards, who was on the

Slashdot Top Deals

"I'm a mean green mother from outer space" -- Audrey II, The Little Shop of Horrors

Close