Does social media have a censorship problem?

12 May 16 | mobile, News and features

Credit: Flickr / Jason Howie

Facebook made headlines this week over allegations by former staff that the site tampers with its “what’s trending” algorithm to remove and suppress conservative viewpoints while giving priority to liberal causes.

The news isn’t likely to shock many people. Attempts to control social media activity have been rife since Facebook and Twitter launched in 2006. We are outraged when political leaders ban access to social media, or when users face arrest or the threat of violence for their posts. But it is less clear cut when social media companies remove content they deem in breach of their terms and conditions, or move to suspend or ban users they deem undesirable.

“Legally we have no right to be heard on these platforms, and that’s the problem,” Jillian C. York, director for international freedom of expression at the Electronic Frontier Foundation, tells Index on Censorship. “As social media companies become bigger and have an increasingly outsized influence in our lives, societies, businesses and even on journalism, we have to think outside of the law box.”

Transparency rather than regulation may be the answer.

Screen Shot 2016-05-11 at 17.16.47Back in November 2015, York co-founded Online Censorship, a user-generated platform to document content takedowns on six social media platforms (Facebook, Twitter, Instagram, Flickr, Google+ and YouTube), to address how these sites moderate user-generated content and how free expression is affected online.

Online Censorship’s first report, released in March 2016, stated: “In the United States (where all of the companies covered in this report are headquartered), social media companies generally reserve the right to determine what content they will host, and they do not consider their policies to constitute censorship. We challenge this assertion, and examine how their policies (and the enforcement thereof) may have a chilling effect on freedom of expression.”

The report found that Facebook is by far the most censorious platform. Of 119 incidents, 25 were related to nudity and 16 were due to the user having a false name. Further down the list were content removed on grounds of hate speech (6 reports) and harassment (2).

“I’ve been talking with these companies for a long time, and Facebook is open to the conversation, even if they haven’t really budged on policies,” says York. If policies are to change and freedom of expression online strengthened, “we have to keep the pressure on companies and have a public conversation about what we want from social media”.

Critics of York’s point of view could say if we aren’t happy with the platform, we can always delete our accounts. But it may not be so easy.

Recently, York found herself banned from Facebook for sharing a breast cancer campaign. “Facebook has very discriminatory policies toward the female body and, as a result, we see a lot of takedowns around that kind of content,” she explains.

Even though York’s Facebook ban only lasted one day, it proved to be a major inconvenience. “I couldn’t use my Facebook page, but I also couldn’t use Spotify or comment on Huffington Post articles,” says York. “Facebook isn’t just a social media platform anymore, it’s essentially an authorisation key for half the web.”

For businesses or organisations that rely on social media on a daily basis, the consequences of a ban could be even greater.

Facebook can even influence elections and shape society. “Lebanon is a great example of this, because just about every political party harbours war criminals but only Hezbollah is banned from Facebook,” says York. “I’m not in favour of Hezbollah, but I’m also not in favour of its competitors, and what we have here is Facebook censors meddling in local politics.”

York’s colleague Matthew Stender, project strategist at Online Censorship, takes the point further. “When we’re seeing Facebook host presidential debates, and Mark Zuckerberg running around Beijing or sitting down with Angela Merkel, we know it isn’t just looking to fulfil a responsibility to its shareholders,” he tells Index on Censorship. “It’s taking a much stronger and more nuanced role in public life.”

It is for this reason that we should be concerned by content moderators. Worryingly, they often find themselves dealing with issues they have no expertise in. A lot of content takedown reported to Online Censorship is anti-terrorist content mistaken for terrorist content. “It potentially discourages those very people who are going to be speaking out against terrorism,” says York.

Facebook has 1.5 billion users, so small teams of poorly paid content moderators simply cannot give appropriate consideration to all flagged content against the secretive terms and conditions laid out by social media companies. The result is arbitrary and knee-jerk censorship.

“I have sympathy for the content moderators because they’re looking at this content in a split second and making a judgement very, very quickly as to whether it should remain up or not,” says York. “It’s a recipe for disaster as its completely not scalable and these people don’t have expertise on things like terrorism, and when they’re taking down.”

Content moderators — mainly based in Dublin, but often outsourced to places like the Philippines and Morocco — aren’t usually full-time staff, and so don’t have the same investment in the company. “What is to stop them from instituting their own biases in the content moderation practices?” asks York.

One development Online Censorship would like to see is Facebook making public its content moderation guidelines. In the meantime,the project will continue to strike at transparency by providing crowdsourced transparency to allow people to better understand what these platforms want from us.

These efforts are about getting users to rethink the relationship they have with social media platforms, say York. “Many treat these spaces as public, even though they are not and so it’s a very, very harsh awakening when they do experience a takedown for the first time.”

Pakistan web users force government backtrack on internet filtering

12 Apr 12 | Asia and Pacific, Digital Freedom

The past few months have seen the rise of a vocal and sophisticated anti-censorship campaign in Pakistan that has effectively shamed the government into shelving its plans for a national internet filtering system.

The Pakistan government’s ICT research and development fund issued a call in February for proposals from academia and companies for the development of a large-scale filter to block websites deemed “undesirable”.

According to the call, Pakistani  internet service providers (ISPs) and backbone providers had “expressed  their  inability to block millions of undesirable web sites using current manual blocking systems.”

The document goes on to specify that the system should allow for the blocking of up to 50 million URLs with a processing delay of “not more than 1 milliseconds [sic]”. Were it to succeed, such a blanket system would put Pakistan’s internet on a par with the surveillance and filtering of China’s Great Firewall.

Human rights groups Bolo Bhi and Bytes for All have called on companies not to respond to the bid for proposals. Their methods seem to have worked, with five companies, including Websense, McAfee and Cisco saying they will not bid. Websense issued the following statement last month:

Broad government censorship of citizen access to the internet is morally wrong. We further believe that any company whose products are currently being used for government-imposed censorship should remove their technology so that it is not used in this way by oppressive governments.

The grassroots campaign has also garnered international attention, with a global coalition of NGOs, including Index, Article 19 and the Global Network Initiative, calling for the withdrawal of Pakistan’s censorship plans.

For Jillian C York,  director for International Freedom of Expression at the San Francisco-based Electronic Frontier Foundation, the support of, rather than initiation by, international groups has been key. “I think that it was a combination of strong Pakistani organisations working with international organisations in tandem that made this campaign so big,” York said in an email. “Bolo Bhi and Bytes for All made the campaign a local one, using the language they preferred, but were smart enough to get the right organisations to amplify their voices while still maintaining control of the tone. I think that’s the example that they set.”

“I don’t see a company going forward with it now because there’s been public outrage and naming and shaming,” Sana Saleem, CEO of Bolo Bhi (“speak up” in Urdu), told Index. “There has been consistent effort and collaboration (…) It is tempting to shout but we said ‘let’s sit down first’. If we were reactionary it would make it hard for businesses to join us.”

In addition to appealing to companies and receiving international support, activists continued to contact the Pakistani government. Eventually a member of the National Assembly notified Bolo Bhi that the country’s secretary of IT had confirmed to her that the proposals had been shelved. Yet no official statement has been released, with Bolo Bhi and other civil society members planning to file a consititutional petition tomorrow. Saleem says the verbal commitment could be seen as a delaying tactic, arguing that now is the time to “consistently build on the campaign.”

Saleem says the Pakistani government has been looking for more control of the internet — which is accessed by 20 million of the country’s 187 million population — and that the filtering proposals could give rise to blanket surveillance. The vagueness of the terms “objectionable content” and “national security” in the terms of reference might also make the plans prone to abuse.

The proposals also threaten secure, encrypted web browsing available via https. “Something that has always annoyed intelligence agencies is not being able to access https,” Saleem said, noting that the government currently needs a court order if they wish to monitor particular users.  The proposals would essentially absolve ISPs of the responsibility of blocking content manually.

Her fear is where the filtering would stop. “If we allow the state to be our moral police, it could be pornography today and something else tomorrow,” she said, citing a case late last year in which the PTA issued directives to ISPs to block 1,000 pornographic websites.

Given its apparent backtracking, Saleem predicts that the government will now be more careful in how it approaches internet filtering and surveillance. “The government made a huge mistake in making the proposals public, so they might be more covert in the future.” She adds that more controversial issues of morality and blasphemy will continue to pose a challenge in the country. “These are very charged issues,” she said, adding: “when we talk about internet freedom and freedom of expression, the government will continue to use these [issues] as a shield to exert control.”

Saleem’s aim now is to get more stakeholders involved in a broader debate about Pakistan’s national security, starting by holding discussions with university students. “Ideally we’d want the internet to be completely free, but we do know Pakistan is a police state. This is a time when we can sit down and see what we want to do.”

Marta Cooper is an editorial researcher at Index. She tweets at @martaruco.