Facebook suspends you for quoting Dune, promotes you for spreading COVID and vaccine lies

Originally published at: Facebook suspends you for quoting Dune, promotes you for spreading COVID and vaccine lies | Boing Boing


His mistake was being on Facebook in the first place.



FACEFORK’s bot will suspend anyone who types the phrase “I will kill _____” in any context, for any reason. I’ve fallen for this twice now, when exchanging teases with friends. The suspension will happen an hour or later and you’ll first learn about it when you try to comment or post something and “unknown errors” will occur. After a few rounds of this, no doubt designed to frustrate you, THEN it will announce that you are a bad little girl or boy, and you’ve done something you know you shouldn’t have, and you’re being punished. It asks if you’d like to protest. No? you’re suspended. Yes? “In these times of COVID we’re running short staffed so whatevah” and you’re still suspended. Great little system there, zuckerberg. Would be a shame if you had to live in it yourself.


I’m glad @orenwolf is not an algorithm.


Are you really that sure?


I saw someone who was suspended for quoting the Declaration of Independence on July 4th. They’ve got some dumb algorithms (enforced by people as well, no doubt) that are triggered by certain kinds of words and phrases, but not by anti-vaxxer stuff, obviously. That’s guaranteed to lead to bad outcomes.




You told the secret of secrets, now face the consequences for your folly. /s


Well his username is @orenwolf


You can’t expect an algorithm to understand the difference between a quote and an actual threat. Putting an actual threat in quotes doesn’t suddenly not make it a threat, nor is a quote (with or with out marks) a threat. Humans, usually, can tell the difference. AI isn’t there yet.

So, you know, expect some posts to land people in water in the interests of curbing actual threats.

As for the second issue, they need some sort of manual over ride that either reviews and deletes these misinformation groups, or make them islands that never get suggested.


I got my Twitter account locked down until I deleted a post… I was wishing that the energies and group abilities of these anti-vax protesters could be put towards a better use, like cleaning up garbage in a community, or otherwise do something with a positive effect… a tangible result, instead of getting lathered up for a pointless cause that only leaves harm in it’s wake.

And Twitter didn’t like that. But they were fine with the two jackholes that replied with threats.

I’ve also posted some stuff here that wasn’t so kosher… upon reflection I was wrong and took my lumps.

Back in the day if you said something nasty to someone your mom would give you a tweak and tell you to straighten up… that is missing from social media platforms… now it’s more like your brother does the deed and you get in shit for it because that is just how it is, don’t like it? Get a job and move out, but as long as you’re under my roof, my rules.


I’m fairly certain that the people who look at any of the reported comments contents are typically not native speakers of the language and miss all kinds of context. I don’t know how many time I reported horribly racists memes like “you’ve seen elf on the shelf, get ready for piglet on a …” (insert picture of piglet and a small child of a particular ethnicity) and then was told “Sorry this upset you but it doesn’t go against our community standards”. Yet something like “effing white people” gets mean put in FB jail.

1 Like

Emphasis mine. This is almost always a sign that AI is going to have an uphill battle.

Plus, I can imagine the exact same post being an actual death threat, depending on who is posting and their history.

Also, people learning that you could get away with anything if it is in quotes.


Is FB using an AI algorithm to mod things? I was under the impression the only stuff that’s gets looked at is reported content and then it’s looked at by low paid humans in India who understandably don’t understand a lot of the subtext/context/nuance.

I don’t know, but if I were them I would be using sentiment analysis/AI as triage before sending it to humans. It doesn’t really matter though.

At this point? Probably. I mean, they’ve blocked award-winning photographs from the Vietnam war and bog only knows what else, but allow actual disinformation to perpetuate (and even amplify) all because it makes them money.

“But now the holy dollar rules everybody’s lives
Gotta make a million, doesn’t matter who dies”
Queensryche, Revolution Calling

1 Like

For the second issue, I think they are making money from it, or at least it’s not costing them money and there’s a lot of ‘engagement’ on those pages. Because I don’t think it’s that hard to fix.


Probably… the more posts that are made,the more your feed updates, the more ads you see, the more revenue they get.

There should be a base level of integrity for what they allow, and I am sure there is to a degree, but it is way lower than what it should be.