How Twitter and Facebook Censor Content Without Telling Anyone

When government and social media interests align, companies remove content without a second thought—and you won't see it in their transparency reports.

Nov 24 2014, 5:50pm

​Image: ​Pete Simon/Twitter

In the first half of this year, Twitter deleted 183 tweets in Turkey due to government takedown requests, according to its official numbers. It complied with just 30 percent of the removal requests it received, the company says. But what about the others?

Social media transparency reports, which summarize government requests for censorship and user data from around the world, ​tell only a partial story of what's really being taken down, civil liberties experts say.

That's because government interests often align with social network interests, and when that happens, companies are much more likely to remove content, regardless of its potential to speak truth to power. These so-called terms of service violations are nearly impossible to track and don't show up on any transparency reports anywhere. Often, the people actually alerting social networks to terms of service violations are within the government.

In Turkey, for instance, a popular Twitter celeb, Orçun Ortaç, has frequently been targeted by the government because he often tweets criticism of the government from an account called @Taci_Kalkavan (a pseudonym).

So, when his account was flagged by Twitter to be suspended for posting "private" images of a random Turkish guy (taken 20 years earlier), who alerted the company? A Turkish court, not the impersonated man himself, according to local media reports and Ortaç's lawyer, Serhat Koç.

If we're talking about a terms of service violation, there's no transparency at all

Ortaç received an email from Twitter alerting him that he needed to remove four tweets within 48 hours or have his account suspended. The email, he says, went to his spam folder, so his account was suspended anyway (it has since been reinstated). The images were old photos of Adem Aydin, an ordinary Turkish guy who Ortaç decided to use for his Taci Kalkavan persona.

Admittedly, using the photos without permission was kind of a jerk move, but the fact that governments are apparently searching for potential terms of service violations and are directly contacting social media networks is certainly concerning.

Twitter's rules bar the posting of "images or videos that are considered and treated as private under applicable laws," and it also bars "impersonation." Koç maintains that the images Ortaç posted didn't violate either rule.

"Our client's Twitter account was solely for the purpose of parody of the trending issues of Turkey, and has never used the real name of the applicant," Koç wrote to Twitter in an email he shared with me. "The court that [the plaintiff] applied to did not notify us and ask for our defense and refused to accept our plea for objections. In addition to this unlawful process, the court sent such decision to Turkish Telecommunication Authority and the Authority handed it to your company."

Ortaç has since replaced his main image with this cartoon version to comply with Twitter's rules. Image: Taci_Kalkavan

Though Twitter posts official takedown notices that it receives from courts, there's no record anywhere that Ortaç's tweets were removed, presumably because the court request also happened to be a terms of service violation. There are no public records of terms of service violations and thus, tweets (and Facebook posts) that are classified as such won't show up in any transparency reports.

"A lot of the transparency reports don't really matter at all, because both Twitter and Facebook are perfectly willing to take phone calls from government officials who are looking for content that are against their terms and services," Jillian York, director for international freedom of expression at the Electronic Frontier Foundation, told me. "If we're talking about a terms of service violation, there's no transparency at all. It's really insidious, obviously."

For the last several years, York has been running a project to attempt to document terms of service violations. Though there's no reporting of takedowns from Facebook and Twitter themselves, there's often an electronic paper trail, such as the email Ortaç got from the company.

"There are circumstances (e.g. posting of private info or parody accounts which are not totally in compliance with our parody policy) where we give people an opportunity to come into compliance," Nu Wexler, who works on Twitter's public policy communications team, told me in an email. "Our rules apply to all users in all countries. So, an account violating our impersonation policy would be suspended in Russia, Ireland or the US."

Wexler said—and the company's actions back him up—that Twitter pushes back against government attempts to censor content more than any other social media company

York's project is in its infancy and so we have no idea how often Twitter is being pinged by government to remove content that comes from accounts it doesn't like and how often a terms of service violation means someone posted child porn or something truly gruesome.

Twitter and Facebook are not the right place to put the future of peace and human rights

And that, really, is the crux of this. There is no real way of knowing. And we don't know what motivates Twitter or Facebook to fight or not fight certain takedown requests or to interpret its own terms of service broadly to satisfy a certain government. We don't know whether social networks put on a tough façade against government censorship but are actually removing content behind the scenes in this method.

"How transparent are the transparency reports? We don't know the answer to that and there's no way of figuring it out," Meg Ambrose, an international tech policy researcher at Georgetown University, told me. " don't know anything about why this stuff is being posted and the context. Asking them to understand whether it meets some expectation of privacy or propriety doesn't make any sense. They don't know the value of the content to the public."

And asking companies based in Silicon Valley with shareholders and board members to answer to to figure out what to remove and what to push back on in all parts of the world at all times is asking a lot, Ambrose added. When fighting a legal battle potentially threatens the future of a social network in a particular country, you are all of a sudden dealing with a lot more than morals and doing what's right from a free speech perspective: You're dealing with money.

"Putting these battles in the hands of these companies doesn't make any sense. They have their own interests and they employ people who have American ideologies," she said. "It's not the right place to put the future of peace and human rights. If the bottom line is threatened, then any [impetus to fight censorship] might go away."

Regardless of what's happening behind the scenes, it's unlikely we'll ever know the full extent of what's being removed and whose accounts are being suspended.