Advocacy groups say the quiet policy change and weak enforcement mechanisms leave a lot to be desired for the safety of users on the platform formerly called Twitter.
When I first saw shadowbanning used, it was targeted at spammers. Spammers know their behavior is unwelcome, expect to get banned, and have probably figured out how to make new accounts despite attempts to stop them. Making it harder for them to tell when they’ve been banned slows them down and reduces spam.
Trolls share the same traits, and it’s probably a good fit for them as well.
Any serious attempt at spamming would have mechanisms in place to check for shadow banning. For example, on reddit if I am shadow banned and post a comment, nobody else can see my comment. So I’m a prospective spammer and I have 25 accounts posting spam on a bunch of random threads.
What do I do? I get a few accounts as controls and everytime I post something, I automatically open up the webpage using some sort of webscraper/emulator library like Puppeteer/Selenium and then confirm whether my comment shows up on my control accounts. If my post doesn’t show up in a certain threshold of checks, I scrap that account and start using a fresh one.
Maybe it stops the most basic spammers, but I don’t think it’s worth the collateral damage to individuals. Many people’s main form of communications with the outside world is through the internet. I’ve seen cases of people on reddit being shadowbanned for over a year and posting a comment nearly every day with no response.
I think it’s approaching psychological torture. Even if someone says something that deserves to be banned - I think you should simply tell them what they did wrong and ban them.
Like I said, you would use both browser fingerprinting and IP address to identify people. So hopping IP addresses wouldn’t be enough by itself. If you’re curious about how browser fingerprinting works, there’s this neat website: https://coveryourtracks.eff.org/ where you can test your browser. It’s a fairly accurate way to track people across the internet. A dedicated person could bypass these systems, but a dedicated person is also more likely to be aware of shadowbanning.
Shadowbanning is most likely to affect people who don’t understand a lot about how the technology and systems of a social media site work.
After a certain point, you just have to accept that there will be some trolls that bypass the system but it’s not worth finding 100% of them because you’ll catch a lot of regular people in the process. There’s that famous statement with the criminal justice system “better 10 criminals go free than 1 innocent man go to jail”
no prob. the fingerprinting thing is actually quite common. many popular websites have code on them from big companies like Google & Facebook that will check for browser fingerprints so that they can track you across the internet.
for example some website that is a blog about corgis wants to use google analytics. so they copy paste the google analytics code onto their html. that way they can get all sorts of useful statistics about who is visiting their website, how long they stay, etc. but Google doesn’t do this for free out of altruism - google makes sure that every user that goes onto your website now gets tracked.
so using browser fingerprinting, google knows you visited the corgi blog website even though they don’t own that website and you never used Google to click on it.
basically the whole internet has these types of trackers. so google and facebook know your interests, your porn preferences, your health issues, etc all sorts of interesting things just by things like browser fingerprints.
there are ways around this, although it’s an ever-increasing arms race between privacy activists & companies like Google/Facebook
I think shadow banning is unethical. You trick people into yelling at the void, never knowing why nobody ever responds or interacts with them.
Sure, fuck bigots and all that, but undeserving people inevitably get caught up in automatic filters and by overzealous mods.
I say ban and give a clear reason why you banned and use a combination of browser fingerprinting & IP address to ban new accounts.
When I first saw shadowbanning used, it was targeted at spammers. Spammers know their behavior is unwelcome, expect to get banned, and have probably figured out how to make new accounts despite attempts to stop them. Making it harder for them to tell when they’ve been banned slows them down and reduces spam.
Trolls share the same traits, and it’s probably a good fit for them as well.
Any serious attempt at spamming would have mechanisms in place to check for shadow banning. For example, on reddit if I am shadow banned and post a comment, nobody else can see my comment. So I’m a prospective spammer and I have 25 accounts posting spam on a bunch of random threads.
What do I do? I get a few accounts as controls and everytime I post something, I automatically open up the webpage using some sort of webscraper/emulator library like Puppeteer/Selenium and then confirm whether my comment shows up on my control accounts. If my post doesn’t show up in a certain threshold of checks, I scrap that account and start using a fresh one.
Maybe it stops the most basic spammers, but I don’t think it’s worth the collateral damage to individuals. Many people’s main form of communications with the outside world is through the internet. I’ve seen cases of people on reddit being shadowbanned for over a year and posting a comment nearly every day with no response.
I think it’s approaching psychological torture. Even if someone says something that deserves to be banned - I think you should simply tell them what they did wrong and ban them.
what’s your solution for people who troll and harass and then hop to a different IP and spin up a new account the moment they find out they’re banned?
Like I said, you would use both browser fingerprinting and IP address to identify people. So hopping IP addresses wouldn’t be enough by itself. If you’re curious about how browser fingerprinting works, there’s this neat website: https://coveryourtracks.eff.org/ where you can test your browser. It’s a fairly accurate way to track people across the internet. A dedicated person could bypass these systems, but a dedicated person is also more likely to be aware of shadowbanning.
Shadowbanning is most likely to affect people who don’t understand a lot about how the technology and systems of a social media site work.
After a certain point, you just have to accept that there will be some trolls that bypass the system but it’s not worth finding 100% of them because you’ll catch a lot of regular people in the process. There’s that famous statement with the criminal justice system “better 10 criminals go free than 1 innocent man go to jail”
that’s really interesting, thanks for the link and the writeup!
no prob. the fingerprinting thing is actually quite common. many popular websites have code on them from big companies like Google & Facebook that will check for browser fingerprints so that they can track you across the internet.
for example some website that is a blog about corgis wants to use google analytics. so they copy paste the google analytics code onto their html. that way they can get all sorts of useful statistics about who is visiting their website, how long they stay, etc. but Google doesn’t do this for free out of altruism - google makes sure that every user that goes onto your website now gets tracked.
so using browser fingerprinting, google knows you visited the corgi blog website even though they don’t own that website and you never used Google to click on it.
basically the whole internet has these types of trackers. so google and facebook know your interests, your porn preferences, your health issues, etc all sorts of interesting things just by things like browser fingerprints.
there are ways around this, although it’s an ever-increasing arms race between privacy activists & companies like Google/Facebook
Maybe a middle ground? As eventually they will find out anyway, maybe after x time they should be indicated they got banned?
I think that’s definitely a better policy. It would be a step in the right direction.