Facebook employs 7,500 to parse free speech from hate speech. Does it know the difference?
This morning, The New York Times launched a clever crowdsourcing campaign to help determine whether or not six “deeply insulting” statements should qualify as hate speech on Facebook.
In June, ProPublica posted the hate speech rules Facebook uses to train its reviewers. It later came under fire for prioritizing white men over Black children in the screening process, prompting it to change its policy to cover age as a protected category. Some question the maneuver’s sincerity.
Muddying the water further is the platform’s policy on modifiers. For example, “women need to be hit in the head” does qualify as hate speech, because it advocates violence based on gender (a protected category). But “female sports reporters need to be hit in the head” does not qualify. Because, as the Times observes, including occupation in the attack “negates the protection based on gender.” Weird.
One ProPublica commenter says, “The idea of censorship of social media just feels like a slippery slope. When some humans are setting a rubric for other humans, however thoughtful and logical it may seem, it makes free speech meaningless.”
Another says, “They protect based on gender identity? Tell that to their ‘real name’ policy enforcers.” ProPublica is asking users to report hate speech through its Facebook page.
Did any of the answers surprise me? Sure. Facebook considers “white men are assholes” hate speech, but saying “poor black people should still sit at the back of the bus” is okay. If the social network has taught me anything, it’s how to judge a book by its cover. Am I crazy to think so?