ProPublica recently obtained some internal documents related to Facebook's hate speech moderation. Hate speech -- as applied to Facebook -- isn't a statutory term. Much of what Facebook removes is still protected speech. But Facebook is a private company and is able to remove whatever it wants without acting as a censorial arm of the government.
That being said, there's a large number of government officials around the planet who feel Facebook should be doing more to remove hate speech -- all of it based on very subjective views as to what that term should encompass.
It's impossible to make everyone happy. So, Facebook has decided to apply a set of rules to its moderation that appear to lead to completely wrong conclusions about what posts should be removed. A single image included in the ProPublica article went viral. But the explanation behind it did not. The rules Facebook uses for moderation lead directly to increased protections for a historically well-protected group.
[If you can't read/see the image, the slide says "Which of the below subsets do we protect?" with the choices being "female drivers," "black children," and "white men." The answer -- to the great internet consternation of many -- is: "white men."]
Given Facebook's general inability to moderate other forms of "offensiveness" (mainly female breasts) without screwing it all up, the answer to this quiz question seems like more Facebook moderation ineptitude. But there's more to it than this one question. The rest of the quiz is published at ProPublica and it shows the "white men" answer is, at least, internally consistent with Facebook's self-imposed rules.
Facebook must define "hate speech" before it can attempt to moderate it, since there are no statutes (at least in the United States) that strictly apply to this content. Here's how Facebook defines it:
Protected category + attack = hate speech
These are the protected categories:
Sex Race Religious affiliation Ethnicity National origin Sexual orientation Gender identity Serious disability/disease
Here's what's not considered "protected" by Facebook:
Social class Occupation Continental origin Political ideology Appearance Religions Age Countries
"White men" have both race and sex going for them. Any "attack" on white men can be deleted by Facebook. "Black children" only have race. Age is not a protected category. An attack on black men would be deleted but black children are, apparently, fair game. The same goes for white children. In the category "female drivers," only the "female" part is considered protected.
The quiz goes on to explain other facets of hate speech moderation. Calling for acts of physical violence against protected categories is hate speech. If any component of the group targeted is "unprotected," the call for violence will be allowed to stay online. The rules also cover "degrading generalization," "dismissive" speech, cursing, and slurs. If any of these target a protected class (or quasi-protected class, i.e., migrants whose nationality may be in flux), moderators can take down the posts. The QPCs have only slightly more protection than entirely unprotected classes, so they can receive more posted abuse before hate speech protections kick in.
These rules lead to all sorts of things that seem unfair, if not completely wrong:
In the wake of a terrorist attack in London earlier this month, a U.S. congressman wrote a Facebook post in which he called for the slaughter of “radicalized” Muslims. “Hunt them, identify them, and kill them,” declared U.S. Rep. Clay Higgins, a Louisiana Republican. “Kill them all. For the sake of all that is good and righteous. Kill them all.”
Higgins’ plea for violent revenge went untouched by Facebook workers who scour the social network deleting offensive speech.
But a May posting on Facebook by Boston poet and Black Lives Matter activist Didi Delgado drew a different response.
“All white people are racist. Start from this reference point, or you’ve already failed,” Delgado wrote. The post was removed and her Facebook account was disabled for seven days.
Religions are unprotected. Races are. That's why this happens. At best, it would seem like both should be taken down, or the less violent of the two remain intact. But that's not the way the rules work. People who criticize Facebook's moderation efforts are asking for something worse than is already in place. To right the perceived wrongs of everything listed above, the rules would have to be replaced by subjectivity -- setting up every moderator, all over the world, with their own micro-fiefdom to run as they see fit. If people don't like it now, just wait until thousands of additional biases are injected into the mix.
That's the other issue: Facebook is a worldwide social platform. Protecting white men may seem pointless here in the US, but the United States isn't the only country with access to Facebook.
“The policies do not always lead to perfect outcomes,” said Monika Bickert, head of global policy management at Facebook. “That is the reality of having policies that apply to a global community where people around the world are going to have very different ideas about what is OK to share.”
This is the unfortunate byproduct of a job that's impossible to do to everyone's satisfaction. Blanket rules may seem dumb on a case-by-case basis, but the alternative would be even worse. If a company is going to proactively protect sexes and races, it's inevitably going to have to stand up for white men, even if the general feeling is white men are in no need of extra protection.