There once was a hack called Jack,
His website a punch once packed,
But users got fucked,
‘Cause he’s a cuck,
So they left, and never came back.
On the 21 April 2015, the guardian published this article, about Twitter pledges to crack down on “harassment” and “abuse” (my quotation marks, because I’m a cynic). Three main changes were announced: “a new filter designed to automatically prevent users from seeing threatening messages”, updated, broader guidelines on threats of violence, and “temporary suspensions on users who [Twitter] does not feel justify a full ban from the service”. In a leaked February memo the C.E.O. of the time, Dick Costolo, who resigned in July, made the company’s attitude and goals clear:
We suck at dealing with abuse and trolls on the platform and we’ve sucked at it for years. It’s no secret and the rest of the world talks about it every day. We lose core user after core user by not addressing simple trolling issues that they face every day.
I’m frankly ashamed of how poorly we’ve dealt with this issue during my tenure as CEO. It’s absurd. There’s no excuse for it. I take full responsibility for not being more aggressive on this front. It’s nobody else’s fault but mine, and it’s embarrassing.
We’re going to start kicking these people off right and left and making sure that when they issue their ridiculous attacks, nobody hears them. (my underlining)
Everybody on the leadership team knows this is vital.
Fastforwards to February 2016, and the Trust and Safety Council is announced. It isn’t enormously well-explained what it does other than that it “will help [Twitter] tap into the expertise and input of organizations at the intersection of these issues more efficiently and quickly”. Its purpose, however, is explicit: “to ensure that people feel safe expressing themselves on Twitter”. When they say: “Twitter empowers every voice to shape the world. But you can’t do that unless you feel safe and confident enough to express yourself freely and connect with the world around you”, it’s very clear that suppressing those denizens who say things that make other people uncomfortable is what they intended to do. “Feel” is the key word; I don’t like it when such subjective things are legislated, because they strike me as nothing more than moral judgements.
I’d just like to interject here and say that expressing some thoughts and feelings necessarily stamps on others; this is something we in free societies must all simply get used to. You can cage the tiger or let it roam free, but you cannot and should not try to build some self-contradicting utopia of a “right balance between fighting abuse and speaking truth to power”. No one has ever been abused on Twitter. Don’t be stupid. You’re downplaying what real abuse looks like. I wrote at greater length about my stance on free expression here.
It wasn’t hard to find a list of members, and my goodness . . . I’m just going to drop this here, with this in case you literally tl;drd it. The Yakin Project is still listed despite not actually having an online presence anymore.
At the same time we have the introduction of “shadowbanning”, which is an ominous and confusing, shady but very obvious practice of soft censorship, which was sort-of-possibly confirmed by everyone’s favourite Breitbart writer, who alleged to have an inside source backed up by some higher-up in the publishing industry:
Shadowbanning, sometimes known as “Stealth Banning” or “Hell Banning,” is commonly used by online community managers to block content posted by spammers. Instead of banning a user directly (which would alert the spammer to their status, prompting them to create a new account), their content is merely hidden from public view.
For site owners, the ideal shadowban is when a user never realizes he’s been shadowbanned.
I’m sure you can guess who I mean by “favourite Breitbart writer”, but what he’s talking about is very much observable. I myself have had several accounts’ replies just vanish. Their tweets are visible to their followers, but quoting is the only way to converse. Not even notifications worked and you would not know you had been responded to unless you checked timelines. They are rarely permanent, but if you go to a certain account in a certain period and you’ll find plenty of plenty of people complaining, and just a little evidence, too.
I remember this particular shadowban myself, and I’m annoyed all of my screenshots proving it were on my old account, and on my old desktop, neither of which are around anymore. It lasted a good while from September to November 2016, and if you look at the user’s tweets in September you’ll find out exactly what the controversy was that caused it. Scott Adams has made a similar claim, as have, if you look, Vox Day, Daddy Warpig, Ricky Vaughn, Mark Kern and others. I damn well remember this happening, after a controversial event, to a controversial account, for a clear length of time, and I won’t just let it be dismissed as paranoia over the website’s ordinary rolling and humming (see afterword).
Anyway, moving on, we of course have to tackle the biggie — the Milo ban. It was unjust, even if the platform is a better and more polite place for it. For those who do not know, he — once @Nero, but no more — is a self-described provocateur who deliberately runs around offending people. His star really rose in 2016 but he’s been around since before that. If you haven’t heard of him then just God help you and why are you reading this? He wrote that Breitbart article. Anyway, he was C.E.O. Jack Dorsey’s enemy number one for a good while before he was finally gotten rid of in July 2016. In January, for some strange and obscure reason, despite being a public figure, and a journalist (it’s true whether you like it or not — and I don’t really), he lost his blue verification mark. In the wake of the Orlando Shooting he was suspended for a while (#JeSuisMilo and #FreeMilo trended at number one), but it was after a spat with Leslie Jones prompted by a review of the new Ghostbusters he wrote that he was finally, permanently suspended, apparently for breaking the harassment guidelines. If HeatStreet is to be believed (I was there but cannot for the life of me remember) he had only three interactions with her:
Of course, those links no longer work because the Twitter account no longer exists, and so they don’t appear on the article either; but I was able to glance them from this video. I couldn’t find an archive for the first so I didn’t look for the other two. This does not justify a ban. It was moderate teasing. I’m sorry. This was wrong. He was an arsehole, and should be held as one, but if this is the precedent for getting banned then free speech really is dead on Twitter. Someone stronger than Jones would have laughed it off.
Onwards to August. Remember that “new filter designed to automatically prevent users from seeing threatening messages” from the start? Well, it was finally introduced. Push through that article to the primary source — an official blog post — and we find out what it does:
Last year we began testing a quality filter setting and we’re now rolling out a feature for everyone. When turned on, the filter can improve the quality of Tweets you see by using a variety of signals, such as account origin and behavior. Turning it on filters lower-quality content, like duplicate Tweets or content that appears to be automated, from your notifications and other parts of your Twitter experience.
Fortunately it doesn’t impact accounts you’ve recently interacted with, and you can turn it off; this was an optional feature originally only available to verified accounts that’s just been rolled out to everyone. I don’t even block, so I have no need for my own automated thought police; but I get the sentiment and it’s absolutely fine that this exists. I oppose all attempts to stop the Internet from being a Wild West, but it’s fine for people to want to shield themselves if they can’t stand the adventure. I have one very specific problem with the thing:
Now we’re into the good stuff. In February 2017 (only two days off the anniversary of the Trust and Safety Council) some more changes were made, and they were hardly announced at all. If you wanted to find out about it you would have to check the blog, which left many users dreadfully confused. It opens with some more spineless weaselling:
Making Twitter a safer place is our primary focus. We stand for freedom of expression and people being able to see all sides of any topic. That’s put in jeopardy when abuse and harassment stifle and silence those voices. We won’t tolerate it and we’re launching new efforts to stop it.
And then they announce their changes, which I summarise:
- People who have been suspended once are very naughty and should not be allowed to return to the platform, because chances are they’re just horrible harassers and we’re perfect and never make mistakes (we will revisit this soon).
- We’re going to alter your searches so that you can’t search for things we don’t think you should search for — for your own good, of course, because we know you can’t make such decisions yourself.
- “Low quality” replies that we don’t like will be indiscriminately collapsed so you can’t see them.
The results of these changes have been dramatic and terrible. Perhaps it’s simply the deplorable corner of the Internet I love and hang out in, but I see the third bullet point everyfuckingwhere I go. I went and collected these just now — literally:
The arbitrary suspensions are arguably even worse. As a case study, let’s take a friend of mine, CursedE, who I’ve critiqued in good faith before. She’s been at this since April 2016, and has been suspended from Twitter with hilarious predictability and increasing frequency; the worst has definitely been since I began my own association with her in January. I’ll just let her speak for herself:
- Shadowbanned in November (20/11/16)
- Gone by December (11/12/2016)
- And again (30/12/2016)
- You thought 2017 would be better? (11/1/2017)
- How silly (14/1/2017)
- Ominous (17/1/2017)
- Late January, with my first reply underneath — they will become more common (29/1/2017)
- “Aggressive following”? (3/2/2017)
- Account number Four (8/2/2017)
- And here we are now! (30/3/2017)
I did not even find her first account, but I have followed no less than fucking four in our relatively brief association. In my mind there is no better example of the site’s arbitrary suspension system. As she explains herself, in each occasion the suspension occurred either because she was spam-reported by people who don’t like her, of because of “aggressive following”. What is perhaps most worrying is that most of these happened before the new guidelines came out in February.
We’re all fed up, and this time there’s a fightback. Our tiny little circle has pestered enough larger Twitter accounts and YouTubers to raise some attention, and the industrious young maybe-lady has organised a little bit of a campaign to make the suspension policy more transparent and possible to appeal. I implore you to read these two posts before you proceed any further, and then sign up to her Thunderclap, which as I write is number one on the dot com. It’s time we shouted very loudly to Dorsey that we can take control of our own discussions thank you very much, and you don’t just get away with booting people off arbitrarily, either carelessly or worse still as a political manoeuvre. I know damn well that, if we don’t stand in solidarity, by the time we have a reason we won’t be strong enough to stand at all.
Out, and I don’t think it’s overreacting to tag this in “Military”.
PS. I thought this was funny:
#Bansparency went live at 17:00 G.M.T. on April 13 with 1175 supporters, and a calculated social reach of 1.69 million. Goodness knows how much overlap there was on various different Twitter follow lists, but it’s fair to say the tweet we all sent was seen by a hell of a lot of people, and by extension the blog post that kicked this all off.
CursedE’s Thunderclap had picked up more momentum than we could ever have hoped for, from @Sargon_of_Akkad to @shoe0nhead to @wrongthinknet to @wokieleaks1, @TheSafestSpace, @sjw_nonsense and so many others. We didn’t dare to imagine uptake quite like that. The success was immense, and I’ve been given far more credit for what was accomplished than I’m prepared to accept; I’ll take some thankses, but certainly not the multiple pingbacks my blog received on hers. She has put together something truly commendable, got the attention of many important people, and raised awareness further of a very pressing issue.
Problems continued during the buildup, as problems do. I myself was almost immediately and very unfairly suspended for trolling a bigot. I’ll just let my tweets stand side by side with his:
Twitter seriously need to sort out their suspension system. This is hypocritical. That he could say things like this but not handle my trolling is pathetic; that Twitter suspends me but not him is ridiculous. That is my verdict. Why is it that he was able to get away with the things he said? I think I know: it’s because none of the people he said it to are sensitive authoritarians. I’m sure if I were to go and spam-report him he would be suspended too. This platform rewards the weak.
Another thing that happened was that CursedE was shadowbanned, exactly like how I remember @Sargon_of_Akkad was. This is simultaneously unfortunate and lucky: unfortunate because it was inconvenient, but lucky because it finally gave me an opportunity to take some screenshots:
Unless you’re about to accuse me of doctoring those, they should provide decent enough evidence. Other people saw it too, as a quick browse of her tweets from the time should reveal:
In case you’re wondering about that last interaction — the one I had — her account was also limited for twelve hours:
The account limitation ran in tandem with the shadowban. I’m not entirely sure what the difference between them was — it’s bloody hard to differentiate all of this censorship — but we know the shadowban was also running because it continued afterwards (we also suspect it began before):
The third was when she outright tested whether her shadowban had ended by asking me. I want you to look at the timestamps, on these tweets and on the above one where CursedE mentions her twelve-hour account limitation:
It went on almost twice as long as it should have done. Even if this was only an account limitation, the point still stands. Read the message. “We’ve detected some potentially abusive behaviour from your account”, it says, “so only your followers can see your activity on Twitter for the amount of time shown below”. There’s no option for appeal, no explanation of what happened, and it even admits the behaviour is just “potential”. If it took them close to a full day after the timer ran out to actually end their limitation then that’s just worse.
It seems to me like both ran in tandem, and neither were justified. The evidence I’ve presented here I think is pretty compelling, and I’ve made sure to back it up every step of the way. We made a big noise with #Bansparency — we were heard even if we didn’t elicit any reply — and I am more convinced than ever that what we are doing is important, and right.