A new ProPublica report featuring internal Facebook documents reveals how the company's censorship policies break down the difference between hate speech and political expression.
According to the report, Facebook "content reviewers" are trained to remove any posts including hate speech attacking "protected categories," but should allow any attacks targetted at "subsets" -- like "black children" or "female drivers."
— Judd Legum (@JuddLegum) June 28, 2017
ProPublica cites two examples of posts targeting "protected categories" and "subsets" in their piece.
In the wake of the London terror attack earlier in June, Republican Rep. Clay Higgins of Louisiana called for the killing of "radicalized" Muslims.
"Every conceivable measure should be engaged to hunt them down," Higgins stated in a Facebook post. "Hunt them, identify them, and kill them. Kill them all. For the sake of all that is good and righteous. Kill them all."
Inversely, ProPubilca cites Black Lives Matter activist DiDi Delgado who at one point posted, "All white people are racist. Start from this reference point, or you've already failed." Delgado's post was removed and her account was disabled for seven days.
According to Facebook's censorship guidelines, Higgins' call to "kill them all" was allowed because it was targeting a "subset" demographic -- that is, "radicalized Muslims." Because Delgado's language included the stance that "all white people are racist," her post was considered an attack on a "protected" group.
For instance, Higgins' incitement to violence passed muster because it targeted a specific sub-group of Muslims — those that are "radicalized" — while Delgado's post was deleted for attacking whites in general.
"The policies do not always lead to perfect outcomes," Facebook head of global policy management Monika Bickert told ProPublica. "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."
According to training slides reproduced by ProPublica, Facebook defines hate speech using the algorithm: "Protected category + Attack = Hate Speech." The company reportedly lists what they do not protect under these guidelines, including social class, political ideology or religion. They do, however, protect sex, religious affiliation, race, sexual orientation, ethnicity, gender identity,
RELATED: Mark Zuckerberg through the years
While Facebook is not required by law to censor content, the company founded in 2004 recently pledged to recommit themselves to monitoring content by doubling their "content reviewers" after a video of a murder was posted to their site.