Hate Speech at Facebook

This story at the Washington Post begins with a two-paragraph lurid example targeting 'the Squad,' followed by several paragraphs of Facebook 'balking' in a way that is described as 'on the side of racists.' The whole piece is structured to assert that Facebook is actively trying to expose black people especially to harmful views. 

Yet there are two paragraphs in the middle that draw the eye.

The algorithm was aggressively detecting comments denigrating White people more than attacks on every other group, according to several of the documents. One April 2020 document said roughly 90 percent of “hate speech” subject to content takedowns were statements of contempt, inferiority and disgust directed at White people and men, though the time frame is unclear. And it consistently failed to remove the most derogatory, racist content. The Post previously reported on a portion of the project.

Researchers also found in 2019 that the hate speech algorithms were out of step with actual reports of harmful speech on the platform. In that year, the researchers discovered that 55 percent of the content users reported to Facebook as most harmful was directed at just four minority groups: Blacks, Muslims, the LGBTQ community and Jews, according to the documents.

When human users report hate speech, then, 55% of reports are of speech targeting blacks, Muslims, the LGBTQ community and/or Jews; but when the machine blindly identifies and removes denigrating comments, 90% of them target white people and/or men. 

The piece's clear perspective is that this shows that the algorithm is failing, as it is identifying content to remove that is different from what the human users think should be removed. The alternative is that our society has a clear bias towards allowing hate speech towards men and/or white people, and that it is only the other sorts of bias that violate our norms enough for people to report it.