Which group is protected from hate speech? The correct answer: white men
Pro Publica provides a large bale of information on why Facebook does the strange things it does.
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.
A trove of internal documents reviewed by ProPublica sheds new light on the secret guidelines that Facebook’s censors use to distinguish between hate speech and legitimate political expression. The documents reveal the rationale behind seemingly inconsistent decisions. 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.
Over the past decade, the company has developed hundreds of rules, drawing elaborate distinctions between what should and shouldn’t be allowed, in an effort to make the site a safe place for its nearly 2 billion users.
…
While Facebook was credited during the 2010-2011 “Arab Spring” with facilitating uprisings against authoritarian regimes, the documents suggest that, at least in some instances, the company’s hate-speech rules tend to favor elites and governments over grassroots activists and racial minorities. In so doing, they serve the business interests of the global company, which relies on national governments not to block its service to their citizens.
Who ya gonna suck up to? Established power, or rebels?
The question answers itself, doesn’t it.
One document trains content reviewers on how to apply the company’s global hate speech algorithm. The slide identifies three groups: female drivers, black children and white men. It asks: Which group is protected from hate speech? The correct answer: white men.
The reason is that Facebook deletes curses, slurs, calls for violence and several other types of attacks only when they are directed at “protected categories”—based on race, sex, gender identity, religious affiliation, national origin, ethnicity, sexual orientation and serious disability/disease. It gives users broader latitude when they write about “subsets” of protected categories. White men are considered a group because both traits are protected, while female drivers and black children, like radicalized Muslims, are subsets, because one of their characteristics is not protected. (The exact rules are in the slide show below.)
Of course this further rests on the absurdity that all members of the “protected categories” get the protection – so whites as well as non-whites, men as well as women, locals as well as immigrants, gentiles as well as Jews, straights as well as gays.
Unlike American law, which permits preferences such as affirmative action for racial minorities and women for the sake of diversity or redressing discrimination, Facebook’s algorithm is designed to defend all races and genders equally.
“Sadly,” the rules are “incorporating this color-blindness idea which is not in the spirit of why we have equal protection,” said Danielle Citron, a law professor and expert on information privacy at the University of Maryland. This approach, she added, will “protect the people who least need it and take it away from those who really need it.”
Thanks, Facebook. That’s working out beautifully.
There’s a lot more. It’s both interesting and frustrating.
H/t Sackbut
Calling for the killing of anyone should be beyond the pale, one would think. I wonder how long a post calling for the killing of the current occupant of the Oval Office would be allowed to stay up, given that he is only a “subset” of a protected group.
The visit by the Secret Service would just be an added bonus.
It’s permissible to call for the death of black people who are eft-handed, but not permissible to call for the death of black people? Or female college students, but not women in general? No, of course I don’t think anyone should be threatening or inciting violence against anybody, but these rules sound totally absurd.
But…I don’t see much sign that they are protecting “women” as a group, a subset, or anything else. Are death threats against women okay, then, because they are only targeting ONE woman at a time, rather than saying “all women must be raped/killed/maimed”?
And why is “black” a protected category, but not “children”? Or are they both protected categories, but then because they are both black and children, they are a subset of a category in each case? A subset of black, a subset of children? In other words, I see no substantive difference between “black children” and “white men”.
Facebook just wants some reason, no matter how stupid, for protecting white men because, face it, it’s run by white men.
Ben,
I wonder how far one can push the limitations. Will Facebook allow a poster to call for the killing of “all black people except those in the NBA,” or “all black people with two arms,” or “all black people not named Clarence Thomas”?
I’m curious as to who drafted these rules, and on what basis. They parrot the “protected class” language of antidiscrimination law, but I’m not aware of any antidiscrimination law that adopts the “but it’s ok if it’s a subset concept” that Facebook is using. Antidiscrimination laws are usually interpreted to disallow that kind of cutesy, but-technically-I’m-not-being-discriminatory gamesmanship.
iknklast,
I suspect it’s that weird — and often disingenuous — attitude that says “racism means you hate ALL people of a certain race. I like SOME members of this race, so it’s not racist when I say that the rest are all [cue tirade of slurs and stereotypes].” Or “I can’t be a misogynist — I love my [wife/daughter/mother/sister]!”
It seems as if both the practice and the theory of Facebook’s rules leave a lot to be desired, and really I would obviously prefer a scenario in which all threats and insults are beyond the pale.
But I am a bit puzzled about the criticism of a colour-blind approach here and in the original article, because (sorry to say) were it to actually be implemented consistently it seems not only fair to me but also the best case scenario. Why would it be better to e.g. ban a white guy for saying all black people are racist while not banning a black guy for saying all white people are racist? Conversely, how would it be furthering discrimination if everybody who says “group X are all and without exception horrible people” was banned regardless of the value inserted for X? (Excepting Xes that are bad by dictionary definition.) It would force people to reject ideas instead of groups of people, so … yay?
Or do I misunderstand what the alternative to the colour-blind approach would be?
Alex – no one on this site is saying that it is okay to spew hate speech at white men; they are questioning the Facebook policies that protect white men while not protecting black children or (technically) any women. So I think your question is a bit of a straw man.
iknklast,
Maybe I misunderstand the approvingly cited last citation box in the post, but it seemed to have a plain reading to me:
Facebook’s algorithm is designed to defend all races and genders equally. “Sadly,” the rules are “incorporating this color-blindness idea which is not in the spirit of why we have equal protection,” said Danielle Citron, a law professor and expert on information privacy at the University of Maryland. This approach, she added, will “protect the people who least need it and take it away from those who really need it.”
I do not understand (a) how an algorithm can protect equally without being colour blind, because if it isn’t colour blind then it does not protect equally, by definition. And (b) how banning e.g. a minority person for saying that all whites are racist is “taking away protection” from said minority. It may take away expression, but protection? How does that work?
Again, I am not claiming that Facebook does actually offer equal protection in practice. My question is about the ideal that we should aspire to.
I read the post as carefully as I could because it’s a direct answer to an earlier question, i.e. what the hell is Facebook thinking?
And, if anything, I’m now even more confused.
“White men are considered a group because both traits are protected, while female drivers and black children, like radicalized Muslims, are subsets, because one of their characteristics is not protected.”
Uh … whuuuut?
Then, logically, nobody is protected at all. Everybody has traits that are not of the group. As other commenters said, left-handed men not protected? Or, for that matter, righthanded men? So long as you threaten each group separately, it’s all A-OK? Threatening young women is cool because you haven’t included grannies?
I guess this is real proper gaslighting because it makes me feel that my mind is going.
Well this looks easy to exploit. “All female drivers are bad drivers” plus “All female non-drivers are bad drivers” equals “all females are bad drivers” except neither step was considered an attack on a protected demographic. “All black children are out of control” plus “all black adolescents are out of control” plus “all black adults are out of control” equals “all black people are out of control. So on and so forth.
Whoever put this together was a simpleton. “All drivers” is, I agree, a group not needing protections from hate speech, but it should be obvious that subdividing that into subsets along race / sex / religion / etc. means that demographic is being singled out for differential treatment. There is no material difference between “all women are bad drivers” and “all women that drive are bad drivers”.
OK, I knew Facebook standards were pretty inconsistantly applied, but I had not realised the true insanity of their rules.
My mind is officially boggled.
quixote,
Of course it is completely illogical. White men are also a subset – two, actually. They are a subset of men, and they are a subset of whites. So if men are a protected group, white men aren’t, right?
Alex SL – in general, this is a nonsense policy, not a color blind one. It gives misogynists and racists a way to game it while still protecting the class white male, which is really no different (as you pointed out above) in that it is as much a subset as is black children.
And Facebook does not apply this rule, anyway, since women in general are often attacked as a group, and it is difficult to get posts removed expressing hate speech toward protected groups (and the ones that are legally protected, because they are in need of legal protection because so many millennia of the power structure keeping them from succeeding). They have set up a way in which they can protect white males from hate speech, while allowing all sorts of other hate speech to remain on the flimsiest of grounds. This appears to me to be the basis of that statement you quoted.
And I would also note that the majority of people who cite “color blindness” usually mean “I don’t believe there should be protections for groups that have been targeted” and that loudly shout “All lives matter” when they really mean “white lives matter, how dare you give any specialized protections to this group we hate”. I suspect from the way you spell “colour” that you may not be USAian and may be reading these things the way they appear to say, rather than with all the subtle and not-too-subtle undertones that are usually the real meaning behind all the Orwell speak in this country.
@ Alex Sl
The problem with any “colour blind” policy is that to result in equality it must assume an initial level playing field. Otherwise, those with privilege retain that privilege and disadvantaged groups remain disadvantaged.
Applying it hypothetically here: the statement, “All white people are racist. Start from this reference point or you’ve already failed,” is a harshly stated but accurate account of how life is experienced by non-white people, and how problems of racism must be approached institutionally. It’s not accusing people of individual racism, or saying all white people are bad people, but identifying that simply by being the dominant culture white people must recognise they have racist “programmes” running in the background of their minds. Non-white people are exposed to this unintentional (mostly) programming all the time while white people just don’t see what they are doing or how it is happening. We’re blind to an awful lot of it.
(Yes, of course all human beings, of all ethnicities, have these programmes running. Yes, you can say – with truth – that, “All people are racist”, but that has the same problem as changing, “Black lives matter,” to, “All lives matter.” “All lives” ignores the fact that black lives appear worth less to the dominant white culture. “All people are racist,” ignores the fact that non-white people are far more disadvantaged by endemic white racism than vice versa.)
So, Facebook’s regulations are overtly supporting white racism by banning a guy who is talking about the impact of white racism on black lives. They also support white racism by allowing death threats to be made against people of colour – because apparently you’re allowed to do that, as long as you have a footnote saying, “not all Muslims. Just the nasty ones. You know.”
As Iknklast says, the whole thing is nonsensical, yet still manages to support the dominant white patriarchal culture. Funny how often that works out.
“Colour-blind” and “Everyone is equal” are code terms for “We will protect the supremacy of the group in power by defining any pushback from people they victimize as equally morally reprehensible as the abuse. In fact, we’re going to concentrate on how victims are ‘abusive’ right back to powerful people.”
Yes, I understand that that is *not* how a person of good will would define “colour-blind.” I get that, and I’m not accusing anyone here of taking that position. It’s not necessary to argue to me that you don’t believe this, because I have already extended that benefit of the doubt to everyone here.
But it is how it functions, and how it *must necessarily function* in the society we live in, because of how it is inherently exploitative.
Speaking of “color blind”…aka “I don’t see color” aka “So you’re spurning Dr King’s dream and returning to judging people by the color of their skin instead of their character?” –
https://twitter.com/michaelshermer/status/879902217397981184
inkclast,
Again, I am not claiming that Facebook does actually offer equal protection in practice. My question is about the ideal that we should aspire to.
I suspect from the way you spell “colour” that you may not be USAian and may be reading these things the way they appear to say, rather than with all the subtle and not-too-subtle undertones that are usually the real meaning behind all the Orwell speak in this country.
Yes, not a USAian. Your reply raises the interesting question, though, of how you distinguish somebody who is a classical liberal in favour of equality but uncomfortable with a strategy to achieve this by anything but equal treatment from somebody who uses the same claim as code for racism. As it goes in science, how would you know your interpretation is wrong?
Also, I am not sure how e.g. an algorithm flagging posts as abusive that have certain combinations of terms could really “mean” only white lives matter.
Steamshovelmama,
I find it interesting how you argue for a non-literal and extremely charitable interpretation of “all whites are racist” but for a non-literal and extremely uncharitable one of “we should treat everybody the same”. In a similar vein to what I wrote above, is everything anybody writes secret code? Should the possibility not at least be entertained that people mean what they say? Should that not be the null hypothesis?
—
More generally, there are two advantages to consistent application of rules. First, it would be fair, by definition. Everybody knows what to expect. Second, and this is a tactical consideration, consistency is crucial to a getting a rule socially accepted. If people get the feeling that there is a clear rule that everybody needs to follow they are more likely to follow it. Justified or not, a rule that were not applied to people from certain backgrounds makes the ones it is applied to resentful and more likely to try and break or remove it. This applies to privileges of the rich or the current unfair treatment of women on Facebook as much as it would to an envisioned situation in which US minorities can call all whites racist on FB but whites cannot call an entire US minority racist.
@ Alex SL
Most people in the real world do speak in codes and what they say normally has implications far beyond the literal meaning of the words.
I wasn’t being “liberal” in my interpretation of the “Whites are racist” comment. I was applying my experience of that kind of discourse to it. As we all do. Am I merely much more experienced than you with listening to people of colour describe their experience of society?
To the ” let’s treat everyone the same” I am applying my experience of how systems work to support a dominant paradigm and what generally happens when we fix a rule or law whilst ignoring a systemic dysparity. Charitably, it’s naive. Realistically, it’s usually a way for the guys in power to make sure they stay in power.
However, it is beginning to seem to me that you are not arguing in good faith. Either that or you grew up completely isolated from anyone different to you, and have never bothered to consider the experiences of women, people of colour, LGBT folk etc etc. It’s not surprising therefore that you don’t see what the problems are with something that apparently seems so unproblematic to you. If you are trying to discuss things in good faith – as oposed to trolling, jaqing off etc – then I suggest you actually listen to other people’s experiences of how apparently fair and equitable rules and standards can be turned against them.
Steamshovelmama,
Unfortunately I don’t really see how telling me to listen to minority experiences addresses considerations of fairness and social acceptability of a suggested rule. More generally, I am unimpressed by the meme that goes “you are ignorant, because if you had learned enough you would obviously have to agree with me”. This is not about (only) demonstrable facts that can be learned but about values, and there people may just have to agree to disagree. Consistency is important to me, both intellectually and (for reasons I tried to point out) in rule-making and -enforcement.
Unfortunately, I also don’t really see any significant difference between “realistically” labeling everybody who argues for consistent rules a stealth agent of oppression and labeling somebody who replies to the question if MTF transgender people are “real women” to the effect of “depends on context and what you mean with that, also I don’t like to be forced into a purity test” as transphobic. The logic appears to be the same: you say innocuous-sounding thing X, but clearly you must mean evil thing Y because if you were a decent person would have no alternative but say Z.
If everything is code, to be interpreted charitably or uncharitably only depending on whether the author is member of one’s own group, communication becomes impossible. The way to figure out if it is code is to look at the author’s actions. (And yes, here Facebook seems to have a lot of room for improvement.)
Only guessing of course, but I suspect MLK was referring to a future world where race had become so irrelevant that it simply did not matter. Not one where we maintain a polite veneer by pretending it doesn’t exist while all the while allowing the consequences of past and present actions to continue discrimination.
The first would be a meaningful step toward utopia. the second is just a comforting plausible deniablity.
@ Alex SL
a) I’ve actually already covered some of those “questions” – inverted commas because you aren’t really asking for information. (If you are, you certainly aren’t listening to the answers).
b) unless you don’t live in real world and are utterly isolated from any kind of social intercourse, you show yourself utterly incapable of accepting that other people’s experience of how things work might be different to yours – yet just as valid.
I also suspect you aren’t completely ignorant of linguistics and the concept of linguistic codes – which all cultures and subcultures have (you’ve heard of the term “code switching, right?”) Even if you don’t know the actual terms, the phenomenon is familiar to all of us who live in the world and are of reasonable intelligence and observation skills. Even Aspies don’t have much problem grasping this concept (I have a son who is a high-functioning Aspie).
All this leads me to the conclusion that, no, you aren’r arguing in good faith, but are just jaqing off. You think you already know the answers to your “questions” and are trying to troll people into “revealing” that they are racist by your standards.
That being so, I’m done with giving you the benefit of the doubt. I’m out and cede to you the last word
Steamshovelmama,
I can only say that if you read my very first comment you will see that the questions I originally asked (e.g. how would it further discrimination if all abusive behaviour was banned) have not been addressed at all. At best one might say that your first reply tries to do so by equating “not directly helping the minority to achieve equal representation (because a comment policy alone can’t do that anyway) but at least stopping all abuse” with “furthering discrimination against that minority”.
But perhaps it would have been more polite if I had phrased my position as statements instead of questions. I may underestimate the degree to which that rhetorical device is perceived as annoying, so I am sorry for that.
Re ‘protected characteristics’, I expect this just means the characteristics legally protected by, in the UK, the 2010 Equalities Act (and presumably something similar in the US):
age
disability
gender reassignment
marriage or civil partnership
pregnancy and maternity
race
religion and belief
sex
sexual orientation
So ‘white men’ is two protected characteristics, race and sex, while ‘women drivers’ isn’t because ‘driving status’ is not a legally protected characteristic. They’ve made a mistake with ‘black children’, though, since age is a protected characteristic as well as race–though I guess legally that’s more likely to refer to age discrimination in employment.