Yesterday I tried to post a link on FB to an "Anonymous video maker" that I saw linked in an article shared here. It's just a web based video maker but when I tried to post it I got a message saying "Facebook thinks this is spam" and FB wouldn't let me share it. Since the page doesn't promote any political view at all we might hesitate to make the leap that they are censoring political views but because the entire pu…
The reason they have is when the spam filter is run against anonymous-video.com it returns positive. You're assigning far too much agency to the decision as if someone at FB reviewed anonymous-video.com specifically and decided to block it or not. Its just a machine learning model that has been trained on a bunch of existing spam finds signals in common with your link.
The policies are explicitly written in code (Haxl) but a lot of them invoke machine learning models to do actual classification of stuff (URLs, images, comments, &c.)—although I’d imagine an increasing amount of the computation is being moved over to the Haxl side because it’s so much less resource-intensive than the FXL system it replaced. Where possible, SI usually errs on the side of “better safe than sorry” when it comes to spam, malware, fake accounts, &c.—hence users’ complaints of “getting zucc’d” when someone reports them, or their activity on the site triggers a filter.