So, Coke censored “Lesbian” because they treated it like a porn category instead of a sexuality. They’re not the only ones to accidentally censor something due to often auto-generated or purchased censored word lists. Bing accidentally censored searches for “Tank Man” this year. ByteDance, the owner of TikTok and Weibo, has their own “proprietary” list of censored words. This has gotten them in trouble multiple times, from banning transgender users, demoting LGBTQIA topics, or outright censorship of anything that’s against China’s human rights abuses.
Companies are censoring content. Unfortunately, it’s a homogeneous and privileged group—often lacking any meaningful diversity—setting up these censored word lists. That’s lead to accidental censorship around the world. On top of that, we have very real censorship of these topics in certain countries. In China, you can’t ever search for “Tank Man.” In Russia, most of the Middle East, China, and many other nations, you can’t find content for LGBTQIA+ people. There are so many censorship lists, all kept secret, until someone messes up and applies the wrong list to the wrong region.
But why do it at all?
In This Article:
Too Much Censorship
Recently, Apple caught some flack for censoring apps in 152 countries. Just in time for Pride month, we found out that, rather than exit the countries, Apple may have chosen to censor LGBTQ+ apps in 27 of those countries. Countries themselves can block apps, but it’s Apple who chooses to allow this, rather than leave.
YouTube is famous for targeting their LGBTQ community. A trans YouTuber proved that his videos would get demonetized if the titles contained the word “trans.” Demonetization blocks some or all ads on the video, so creators cannot make money from them. Simultaneously, YouTube allows LGBTQ hate videos, and has even allowed advertisers to pay to put these hateful videos before an LGBTQ creator’s content. LGBTQ+ words are, or were, on YouTube’s censored word list.
If you wanted to search for “Tank Man” on the anniversary of the Tienanmen Square massacre, tough luck. The Chinese government has blocked searches for their attack on a pro-democracy rally, and requires all search engines in the country adhere to their censorship guidelines. However, a mistake over at Microsoft’s Bing meant that even those outside of China were subject to China’s revisionist history.
“I don’t know how it works, but I do know that [TikTok is] not made for us; it’s made for white, straight, cis people to do dances and stuff.”
– Adea Danielle, trans TikTok user
Speaking of China, TikTok has been incredibly protective of their censored words list. However, creators have noticed that videos about LGBTQIA+ topics gain no traction or spread, while their other videos will spread. Even TikTok admitted to censoring LGBTQIA+ topics, and numerous transgender users report the network has deleted their accounts and videos on multiple occasions. They’ve done the same for videos about their Uighur concentration camps and other human rights abuses. During large scale protests over police violence and overreach and systematic racism in the United States, TikTok buried videos about protests. Often, they described this censorship as a bug, but censorship is the default for the company with close ties to the Chinese government. Like other companies, they often mix up their lists. It’s so hard to keep track of what doesn’t have to be banned as well, apparently.
Censorship Overhead
Companies that engage in this kind of censorship don’t have an easy task. In fact, it could be an engineer’s entire job to manage just one part of this process. At larger companies, the task is likely left up to multiple engineers. That’s due to the complexity of censoring content. Watch how this complexity stacks:
- Companies must maintain multiple censored word lists for words censored legally by a country.
- They must identify users in those countries either through:
- GPS location
- IP Address
- Location given during sign-up or on a profile
- Potentially (and problematically), language choice
- Then they have to apply any censors that they may want themselves for all or some users. This can include hate speech and slurs in the censored list.
- User preferences will also matter. A user may have chosen to filter out posts with topics they’re uninterested in.
- There’s also all the usual algorithms for sorting posts, profiles, search results, etc. These are built-in to those processes, but it’s an additional complexity.
This is a large undertaking. It’s why we’ve seen problems. TikTok censored LGBTQ content for users speaking Arabic, because it assumed if they were using Arabic they lived in a nation that banned LGBTQ content. Bing accidentally applied their Chinese censorship worldwide, hiding the iconic “Tank Man” photograph. Mass censorship on a per-user basis is not an easy task and even the largest companies screw it up. All it takes is the mistake of one engineer to create an international problem.
And it’s not just one engineer working on these either. Many engineers are maintaining these tools, along with perhaps the largest cost: human moderators. Human censors who are aware of trends in languages and topics and can update censor lists can also find posts that break rules through nuanced language. They can do what AI cannot. But they’re also fallible, the job is incredibly difficult. Facebook’s censors filter through violent posts, gore, and other horrors that often give employees PTSD, anxiety, and depression.
This is a lot of expensive overhead for something that is done at the behest of oppressors. Making your platform safer is a noble cause that is worth the price. It’s what allows a platform to grow, become profitable, and—outside of a capitalist lens: more importantly—help people express themselves. Yet companies have decided that forgiving bigotry and these long-term and up-front costs of censorship, are worth doing business in oppressive nations.
Leave the Hate Behind
Viral video critical of China’s treatment of Uighurs…very quickly taken down by (Chinese owned) Tik Tok. #TechnologyWithSocialistCharacteristics https://t.co/OTCXzs7BZh
— ian bremmer (@ianbremmer) November 27, 2019
The core reason companies go along with these oppressive regimes is because, frankly, there’s money to be made. They’re not losing money from boycotts in freer nations, so why not do whatever it takes to put your products everywhere? Take Disney, for example. They’ve gotten heat for a lack of LGBTQ+ representation and caving to censors in China and other heavily moderated countries. None of that has hurt Disney’s bottom line. No one’s boycotting Disney. Therefore, they’ll continue to keep LGBTQ+ stories out of the main plot line so it’s easy to censor in China, Russia, the Middle East, and wherever else they want to sell.
The same goes for Apple, Google, Microsoft, and other tech companies that take advantage of the sheer volume of consumers in these countries, without worrying about their limited rights.
But what can they really do? Leave? Would that even make a difference?
Actually, it could.
If a company like Disney or Apple left these large markets with oppressed consumers, they’d create a vacuum for their products. It would likely fuel bootlegging VPN usage, and other ways to bypass government blocks. They’d be exposed to the uncensored versions of films, media, music, articles, news, etc, with all the reality of life. This could educate and expand the mindset of a populace. That could lead to an uprising, a revolution, and free elections. As long as people are capable of pretending nothing’s wrong, they will.
On the other hand, that’s a lot to ask of people. It’s giving up comfort to engage in potentially illegal and dangerous activities. Is it ethical to create that vacuum, knowing that, while the long-term goals may be noble, the path to them could be violent? Perhaps not.
Of course, companies aren’t weighing ethical decisions. They’re not worried about either of these possible paths. Companies don’t think about spreading democracy, or acceptance for LGBTQIA+ people, or women’s rights. They’re just worried about making money. They’re willing to sacrifice free speech to aid censorship.
And Yes, Removing Hate Speech is Free Speech
It may seem counterintuitive to both be in favor of removing censorship and removing hate speech. After all, isn’t it the same? Yes and no.
Yes, it is censorship to remove hate speech. On the other hand, it’s actually pro-free speech to do so. By making platforms safer, places where people can say what they want without fear of someone attacking them (literally), they are free to fully speak their minds. Hate speech censors non-hateful speech. By keeping hate speech off a platform, diversity of opinions can flourish.
Hate speech is violence. Like shouting fire in a crowded theater, it’s speech made to inspire violence against others. Even if shouting “fire” isn’t telling people to trample one another, it causes that result. Hate speech, in all of its forms, dehumanizes groups of people to make others more comfortable with violence against them. It’s easy to see why we can’t shout fire in a theater. However, the connection between hate speech and violence isn’t so instantaneous. It’s instead something that simmers. Hate speech absolutely leads to violence, just as surely as shouting fire in a crowded theater leads to trampling.
Unfortunately, all of these sites censoring free speech are far worse at censoring hate speech.
Big Tech for Big Profits (And Nothing Else)
Long term, censorship may prove too expensive. The sheer amount of overhead for maintaining a censorship network worldwide, on top of potential for backlash, could prove more expensive than its worth. That’s the only thing that would lead companies to stand up for the rights of their customers worldwide. Unfortunately, that seems unlikely. Instead, tech and entertainment companies will continue to prop up regressive ideologies in the name of profits. Because censorship is profitable. It opens up entire countries to business. Depending on the size and wealth of those countries, companies could decide that mass censorship of their platform is a small price to pay, regardless of the human cost.
While it’s possible that these companies have the power—and therefore the responsibility—to enact global change, the question becomes whether or not pushing for that change is an issue of “the ends justify the means.” While profitable companies could still remain profitable, sacrificing very little, the strife they could push on a populace fighting for revolution could be a much higher personal cost.
As long as governments decide they have to restrict people’s civil liberties, we will have a difficult decision of whether or not to censor content. Unfortunately, that could hurt civil rights and equality even outside of censored regions. It’s difficult to measure worldwide suffering caused by companies complying with censorship, but they’re certainly not the only force to blame. They’re also not doing anything to make things better.
The censorship lists are here to stay, and we’re likely no better off for it.
Sources, Further Reading, and Referenced Articles:
- Joseph Cox, Vice
- Fight for the Future
- Marty Johnson, The Hill
- Shen Lu, Protocol, [2]
- Kim Lyons, The Verge
- Emma Powys Maurice, Pink News
- Chance Miller, 9to5Mac
- Casey Newton, The Verge
- Lauren Strapagiel, Buzzfeed News