An odd thing happened to me on twitter today.
UW System had tweeted a statement from system president Ray Cross concerning Trump’s executive order banning refugees and immigrants from 7 countries. Cross’s statement is close to the most empty, useless possible such statement.
So I tweeted back, asking him to revise his statement and actually take a stand against Trump’s executive order.
Note that I mentioned not just @RayWCross and @UWSystem, but also @POTUS.
Twenty minutes later, the tweet had gotten a couple of ❤️’s, and then one RT, but for some reason the tweet itself had disappeared from my twitter feed (in the app I use, Tweetbot). I looked at my twitter account in a web browser, and sure enough the tweet wasn’t there.
Puzzled, and knowing I hadn’t deleted the tweet, I poked around and was able to find, somehow, the URL for the original tweet, and I could see the tweet in a browser. So it hadn’t been deleted, it had just been removed from my feed.
@Kerri_Gilbert then wrote that she could see it in her feed but not if she went to my timeline. So I wasn’t completely crazy.
@CFlensburg later wrote that he could see the tweet, but he’s in Australia.
Now my hypothesis was: someone who looks after @potus had somehow suppressed my tweet. Or was I being paranoid?
A quick google search (for “tweet disappeared from feed”) revealed a Washington Post article from 2015-10-30, “Tweets are disappearing on Twitter. Why?” This seems to explain what happened: my tweet was suppressed by some kind of “abuse filter.”
The article is an interesting read. It describes the experience of Paul Dietrich (@Paulmd199), who wrote an analysis of his situation: “Adventures in Twitter Censorship.” Also an interesting read.
And note that the tweet suppression is location-specific. It’s okay for Australians to read the tweet, but Americans need to be protected from it.
But why this tweet?
This tweet of mine was really a pretty bland tweet. I mean, I was directly criticizing my boss’s boss’s boss, and we just barely have tenure at Wisconsin anymore, so I certainly wasn’t going to be abusive. I’m not sure why I actually mentioned @potus, but I figured, “What the hell.”
But what sort of abuse detection algorithm would decide that this tweet was abusive? If there’s an abuse filter that would suppress this tweet but not all of the actual abuse rampant on twitter…well that’s bullshit.
Maybe it was that I wrote “Sad.” (I thought that was funny. And hardly abusive.)
I favor the theory that it wasn’t an algorithm but rather a person: that someone is assigned to read all mentions of @potus and if they deem something inappropriate, they flip a switch and the tweet gets suppressed.
So maybe I am paranoid.
(I forgot to mention: @NickFleisher had the best response to Ray Cross’s lame statement: “Revise and resubmit.”)