The Hypocrisy of 188.8.131.52
On April 1st, Cloudflare announced their new 184.108.40.206 and 220.127.116.11 DNS resolvers. A DNS resolver an Internet service layer that tells you how to find a website online. When you type in a website URL, the DNS resolver maps that URL to a location on the Web.
Cloudflare has enjoyed widespread popularity with its privacy-respecting basic DNS resolver, 18.104.22.168. But 22.214.171.124 and 126.96.36.199 are special. 188.8.131.52 will refuse to resolve URLs for known malware-distributing domains. 184.108.40.206 will refuse to resolve URLs for known "adult content" domains. In this case, adult content largely boils down to sexually explicit material.
Essentially, 220.127.116.11 and 18.104.22.168 are content filters, marketed and maintained by one of the biggest and most powerful cloud companies on the Internet.
After 22.214.171.124 was announced, LGBTQ+ community members and advocates discovered that LGBTQ+ educational, community, and news resources were being blocked. This included mainstream resources like Glaad, the LGBT Foundation, and the Transgender Law Center. To add insult to injury, because 126.96.36.199 was focused specifically on sexual content, anti-LGBTQ+ and other bigoted resources like Stormfront were not blocked. All of this combined to create a filter that was hostile to LGBTQ+ advocates, but welcoming to LGBTQ+ opponents.
Cloudflare responded to criticism quickly, both whitelisting specific sites, and promising to be more careful in the future about what they blocked. But on a fundamental level, 188.8.131.52 is still problematic, and Cloudflare still has more to do to assure the LGBTQ+ community that it cares about their fears.
Why this matters
I don't hate filters, and I'm not against creating safe environments online — especially for children. I have and will continue to advocate that users and communities have an unalienable Right to Filter the content they're exposed to, and a Right to Delegate the maintenance of those filters to companies and third-party services. In my own communities I tag, filter, and otherwise limit child-unfriendly content of all types, including not just sexually explicit material, but also disturbing content, hate speech, and even swearing.
But while I make use of filters, I also don't ignore the broader implications and social context of the filters I consume.
Moderation that works on a small scale doesn't always work on a large scale. And painting 184.108.40.206 as a small-scale, purely optional filtering service with a limited, focused scope is deceptive. 220.127.116.11 is a globally accessible content filter designed to be used worldwide. Decisions made about this particular filter will echo across multiple cultures and continents.
18.104.22.168 is also targeted at a demographic that has very little control over its computing environment. It is irresponsible to ignore the fact that for years, child-friendly filters have been widely abused to isolate LGBTQ+ children and to prevent them from finding educational resources and communities. Companies that offer tools that allow the suppression and manipulation of vulnerable people have a greater responsibility to consider how their products will be used.
This does not mean that child-friendly filters are evil. But it does mean that child-friendly filters are inherently more dangerous than other filters that are voluntarily consumed. It means that people who create child-friendly filters need to be thoughtful about their design.
If you're unfamiliar with the history and usage of adult content filters online, it may seem strange for LGBTQ+ communities to be so worked up about what Cloudflare claims was an innocent mistake. But these communities have decades of experience seeing the words "family friendly" used as a dog whistle to signal their exclusion from mainstream spaces. There are still deep wounds in this area that should not be lightly dismissed.
On selective fears
Cloudflare is no stranger to concerns about censorship.
After The Daily Stormer became a hot-button issue, Cloudflare openly agonized about the idea of removing hateful, bigoted speech. The decision was so traumatic that it led Cloudflare's CEO, Prince, to openly question whether or not a company of Cloudflare's size was good for an Open Internet.
And to be clear, Cloudflare's deplatforming of The Daily Stormer was not a First Amendment issue. As a private business, Cloudflare has a near-total right to choose who they will and will not do business with. Freedom of association is an essential tool in protecting free expression online — there is no Freedom of Speech without the Freedom to Filter.
But even while acknowledging that they were free to sever ties with The Daily Stormer for any reason, Cloudflare still recognized that a company of its size and power choosing not to provide an essential service based on moral grounds had problematic implications for online expression. Cloudflare worried about the potential ramifications of allowing a private, unregulated company to make decisions about what ordinary people could access online.
By cutting ties with The Daily Stormer, Cloudflare was not explicitly banning the site from the Internet. But it was removing Stormer's ring of protection; a ring that only Cloudflare could provide. Cloudflare wasn't legally obligated to provide that ring of protection. But without it, The Daily Stormer quickly found itself swamped with DDoS attacks. Cloudflare didn't need to kill the site — it simple left it to its fate.
Here's the question: when Cloudflare decided to brand its 22.214.171.124 DNS service as a content filtering tool, was even a fraction of that caution and deliberation applied to censoring adult content? Did anyone involved in this project even consider for one second that there might be problematic outcomes from a company the size of Cloudflare making moral judgments about what content was and wasn't child-friendly?
Did anyone ask whether or not this was consistent with Cloudflare's brand? Did anyone at Cloudflare worry about whether or not LGBTQ-unfriendly countries like China might ask them to expand the filters? Did anyone at Cloudflare ask whether or not media industry executives might start to make noise about blocking sites based on IP violations? Did anyone at Cloudflare worry if those IP claims would be leveraged against LGBTQ-friendly spaces like fan-fiction forums? Did anyone at Cloudflare ask why adult content was being singled out over hate speech; why sites like Stormfront should be given special consideration?
I've heard Cloudflare's apologies, but I do not believe the company was seriously thinking about the impact their choices could have on LGBTQ+ communities. Because we can see how the service launched. We can see that Cloudflare waited until after launching a censorship system into production to test whether or not it was hurting marginalized groups.
No one who was seriously thinking about potential LGBTQ+ harms would have launched a service like this without checking whether or not Glaad was blocked. Sites like Glaad are the low-hanging fruit. They are the obvious sites to check. The fact that Cloudflare was willing to deploy to production without checking these sites means that it was not seeking input from advocates or seriously considering how its services might fail.
Adult content isn't special
Being an advocate for Free Speech means sometimes tolerating content we find offensive. It means worrying about hypothetical scenarios that seem silly, or unrealistic, or just far away. When we ask questions about censoring political content, we are asking the people around us to make a difficult, dangerous choice: to worry about the future ramifications of a decision that is impacting them right now.
We are asking people who are genuinely hurt by racists every day to worry about what the Internet might look like 5 years from now if we over-broadly deplatform. We are asking people who are fighting misinformation and propaganda online to worry more about theoretical future censorship than about the potential outcomes of the next election.
And when people in our camp turn around and tell vulnerable groups that it's no big deal to start censoring content around their identities, that there's no slippery slope to asking questions about what is and isn't appropriate LGBTQ+ expression — that is one of the most wildly, unbelievably hypocritical things that we could ever do.
Of all companies, for Cloudflare not to recognize the potential political and social ramifications of making globally scoped distinctions between child-friendly and adult content is beyond the pale. It shows a remarkable lack of awareness about their position online, and a complete lack of thoughtfulness about what their actions might mean for ordinary people.
What is the gosh-darned point of any Free Speech advocacy if we still end up suppressing LGBTQ+ people online? What was the point of tolerating literal nazis if we're going to turn around now and tell people that their sexual identities are uniquely dangerous to kids?
And make no mistake — for all of their intentions, Cloudflare's lack of care around this launch is a statement to the LGBTQ+ community. Cloudflare can say all it wants that blocking LGBTQ+ content was a mistake. But it was not an accident that a company with a heavily emphasized Free Speech brand decided to launch a product that uniquely focused on suppressing sexuality while simultaneously decrying the evils of political censorship.
Intentional or not, there is a statement in that launch: that sex censorship isn't real censorship. That the concerns we have about slippery slopes are real when it comes to protecting children from bigoted nazis, but not when it comes to protecting children from a penis.
We take so much care to avoid ever even accidentally censoring political speech, even at the cost of allowing horrific content to reach our children. But when we broadly censor LGBTQ+ content, that's just "teething trouble". Cloudflare would never accept this kind of collatoral damage when launching a tool for political censorship, but when it comes to sexual expression they have no such weighty fears.
Cloudflare isn't unique
If possible, I want to be even more blunt about this.
When LGBTQ+ advocates call out Free Speech advocates as hypocrites, they are usually right. Our society is broadly hypocritical about sexual censorship. We are far more nervous about censoring hate speech than we are about censoring people's sex and gender identifications. We are unwilling to draw a line between political expression and bigotry, but perfectly willing to attempt to draw the same line between identity and pornography.
There is no ideological justification for this double standard. There is no ideological justification for granting hate speech more consideration and tolerance than we grant pornography. And yet we continuously see so-called captains of Free Speech abandoning the battles that would make the biggest impacts towards empowering marginalized communities.
We hear, again and again, that some battles are worth fighting because protecting bad people allows us to protect good people. But we prove, again and again, that there are some people we care more about protecting than others.
Why is it that Steam is so concerned about banning games that advocate racism, but totally okay censoring adult content via poorly specified rules? Why is it that sites like Facebook are willing to create broad, sweeping protective policies towards political misinformation, but force users to jump through hoops for even basic queer accommodations like post-transition name changes?
When these companies have the opportunity to prove their ideological commitments to free expression when it really matters, when vulnerable people really need that protection, they consistently prove instead that their principles have selective holes.
It is long past time for Free Speech advocates to extend LGBTQ+ and other marginalized communities the same level of concern, consideration, and care that we extend to mainstream political groups — groups that already have far more resources and far more mainstream advocates with which to defend themselves.
Across the entire tech industry, we must do better than just playing lip service to these communities. The message we're broadcasting with products and rollouts like 126.96.36.199 is loud and clear: that there are some people we do not care enough about to protect.
More than an apology
Cloudflare owes LGBTQ+ communities more than an apology. It owes them tangible action. It owes them steps that will give them confidence that 188.8.131.52 will not turn into yet another tool for supression.
First, Cloudflare should publicly sever ties with the providers that supplied lists blocking LGBTQ+ sites. Merely switching to a different list from the same provider won't solve the problem. A company that is comfortable classifying Glaad as child-unfriendly content in any of their filters will have also censored dozens or even hundreds of smaller LGBTQ+ sites, and that bias will creep out into every one of the filters they provide.
Every filter decision that provider makes in the future will still have been considered through an anti-LGBTQ+ lens.
Second, Cloudflare should make its filter lists public. 184.108.40.206 is already being offered as a free service. There is no business benefit to secrecy, since users can already scrape sites to find out what is and isn't blocked. But if Cloudflare genuinely wants its users to be identifying and reporting mistakes, then it shouldn't force those users to manually crawl the web trying to figure out which sites are miscategorized. It should publicly and accessibly list which sites will be blocked.
Third, Cloudflare should try to publicly nail down more concrete rules about what it will and will not consider to be objectionable content. This is a difficult ask, but if Cloudflare can't articulate what it means by pornography, then it's worth asking whether or not it's appropriate for Cloudflare to launch a product worldwide under their widely-recognized and respected brand that claims to be able to make that distinction. A set of concrete rules will also help Cloudflare when (inevitably) more repressive governments start requesting the addition of new sites.
Finally, Cloudflare should commit to applying the same standards to hate speech that it applies to pornography. There is no justifiable argument to be made that a child will be harmed more by seeing a penis than by consuming bigoted propaganda. If Cloudflare is committed to protecting children online, then what's good for the goose should be good for the gander.
Show LGBTQ+ parents that we care just as much about protecting their children from hate speech as we care about protecting them from sexual content. Block sites like Stormfront from 220.127.116.11.
And may I humbly suggest that if the idea of Cloudflare deciding what political content is harmful for children frightens you, that if the idea of Cloudflare drawing a bright line between political affiliation and bigotry has you concerned — then maybe the rest of this filter should concern you too.