Social Media Companies Like Instagram Are Censoring Dissent
In recent weeks, Instagram and Facebook have censored posts focused on COVID-19 in India and protests in Colombia and Palestine — with little explanation as to why.
On May 6 and 7, Instagram users in India noticed that some of their posts were starting to vanish. Gone were their COVID-19-related posts that demanded improved conditions for overworked crematorium workers, publicized volunteer-led relief efforts, and linked coronavirus deaths in the country to “abject callousness” of the government. Stranger still was the removal of private chats on the matter.
“There is a growing trend of internet shutdowns, takedown of social media content, particularly around political speech in India over the last few years,” said Vidushi Marda, global AI research and advocacy lead at ARTICLE 19, an international freedom of expression organization that has been tracking the deleted content.
In India right now, whether or not people have access to COVID-19 information on social media is a matter of life and death. Such censorship, however, is not unique to the country. Over the past month, activists and researchers have also collected numerous examples of suppressed content related to unrest in Palestine and Colombia, as well as posts related to the National Day of Awareness for Missing and Murdered Indigenous Women in the United States and Canada.
On May 7, Instagram said that “this is a widespread global technical issue not related to any particular topic” and that the issue had been “fixed.”
But the following day, the company acknowledged that there were issues with posts relating to unrest in Colombia and Palestine.
“We are so sorry this happened,” Instagram noted in a statement. “Especially to those in Colombia, East Jerusalem, and Indigenous communities who felt this was an intentional suppression of their voices and stories — that was not our intent whatsoever.”
But Instagram failed to acknowledge reports of censorship in India.
A representative of Facebook, which owns Instagram, wrote in response to questions about why dissent in India, Colombia, and Palestine seemed to have been disproportionately impacted: “This was a widespread global technical issue that affected users around the world, regardless of the topic of their Stories. We fixed it as fast as we could so users around the world could continue expressing themselves and connecting with each other through Stories.”
Despite the company’s claims that the takedowns were automatic and universal, Marda said there was “overwhelming evidence of the disproportionate impact these takedowns have had on political speech and dissent.”
In India, she noted that ARTICLE 19 observed “significant overlap between posts about activism, COVID-19 relief and government critique.” All of this, she said, points to “a significantly larger problem than just a single automation tool,” and noted “the opacity of content moderation practices” means that there are gaps in accountability.
Such digital suppression isn’t simply a matter of being able to speak freely. In each of these countries, thanks to government failures and limited media coverage, people have come to rely on social media to share information, track resources, and protect themselves from violence.
Part of the problem is automated content moderation, which uses machine learning to filter content. The systems are blunt instruments that often misunderstand context and remove too much or too little content, noted a report by the New Delhi–based Observer Research Foundation. These developments, adds the report, can negatively impact minority groups because these tools are often trained on English-language data sets, so they have trouble properly parsing dialects and rarely used languages.
“[There is] overwhelming evidence of the disproportionate impact these takedowns have had on political speech and dissent,” said Marda. “[This is] precisely why . . . human rights organizations and defenders around the world have pointed to the dangers of automated content moderation for years.”
India’s History of Digital Censorship
Because of the Indian government’s monumental failure in tackling the coronavirus, people in the country have come to rely on social media to seek and provide COVID-related help like oxygen supplies and vaccinations. Many people have also used social media to collate lists of supplies into a larger, searchable database.
Silicon Valley–driven censorship in India, therefore, has become a matter of survival, despite the fact that Instagram has yet to acknowledge it.
“Despite documented instances of censorship [in India] and Instagram users highlighting them very prominently, there was a complete lack of recognition [by Instagram] of what’s happening in India,” said Apar Gupta, executive director of Internet Freedom Foundation (IFF), a New Delhi–based organization that seeks to ensure that technology respects fundamental rights.
Digital suppression in the country isn’t new, despite the fact that the Indian Constitution guarantees the right to freedom of speech and expression.
In 2020, India had the highest number of government-instigated internet shutdowns in the world. The digital crackdowns were one of the reasons Reporters Without Borders recently ranked India 142 out of 180 countries in terms of press freedoms.
On April 28, Facebook temporarily hid posts critical of Indian prime minister Narendra Modi that included the hashtag #ResignModi for “violating its community standards.” A Facebook spokesperson later said that the posts were hidden “by mistake, not because the Indian government asked us to.”
“Silicon Valley platforms have a very natural interest in keeping governments happy in the regions that they operate,” Gupta said, pointing to the fact that India is Facebook’s biggest market.
The lack of institutionalized free speech protections is further compounded by laws and regulations in India that allow the Ministry of Electronics and Information Technology to not disclose censorship orders sent to social media companies, said Gupta.
Users are therefore often given no official explanation why their posts were suppressed.
Content Moderation in Colombia
There have also been numerous reports of censorship related to ongoing protests in Colombia over proposed tax increases and the resulting police crackdowns.
“We identified a specific problem with Instagram,” said Carolina Botero Cabrera, a researcher with Karisma, a Bogotá-based civil society organization that works on technology and human rights. “We have over 1,000 reports of censorship, around 90 percent of it was by Instagram and the content was overwhelmingly about the [ongoing] protests,” she added.
Deleted posts reportedly related to the national unrest, unemployment numbers in the country, and the death of a protestor.
For Colombia, a country with a long-lasting civil war, such automated content moderation is all the more contentious because journalists and human rights activists often find that their content is removed, their reach is diminished, or their accounts are blocked because their content is deemed too violent.
Jesús Abad Colorado, an experienced Colombian photojournalist, recently had his Twitter account blocked after he posted photographs of an armed dispute in the Chocó Department in Western Colombia. A few days later, when an independent media outlet livestreamed an interview with Colorado about the dispute, their account was blocked, too.
Another challenge, said Botero, is that the Revolutionary Armed Forces of Colombia — People’s Army (FARC), the longtime leftist guerrilla group that disarmed and became a political party in 2017, “was flagged as a terrorist organization [by social media companies at the time] even though they were in peace negotiations.”
The peace process spanned about four years, culminating in a peace agreement in 2016. “Any research about the peace process will have to deal with important problems to [understand] FARC’s position, actions, and voice,” said Botero, noting that blocked social media accounts and deleted content hamper documentation of the process.
Suppressing Palestinian Voices
As tensions escalated in Israel and Palestine, digital suppression in the region also appeared to increase.
“We have over 100 reports of censorship on Instagram,” said Alison Carmel Ramer, a researcher at 7amleh, a digital rights organization based in Haifa, Israel.
Ramer’s research and other reports found that most of the censored content was related to Israeli forces storming Jerusalem’s Al-Aqsa mosque. Other censored content was related to the eviction of Palestinians from the Sheikh Jarrah neighborhood in East Jerusalem.
Muslim, a media publication, also documented blocks on Instagram livestreams related to Palestine.
According to ِRamer, Facebook told 7amleh that a majority of the Instagram takedowns were mistakes because they did not violate community standards and that they have restored the content.
“This means there is a problem in the way content is moderated,” said Ramer. “Why is content which is not against community standards being taken down? [Facebook] also did not tell users under which policy the content was taken down.”
In general, Palestinian content is “over-moderated,” Ramer added, noting posts are often suppressed either because they are considered hate speech, or the posts appear to be connected to terrorist organizations. Many Palestinian leaders are designated as terrorists by the United States, meaning Facebook censors content related to them. Ramer also explained how hate speech in the region written in Hebrew is not censored to the same extent as hate speech in Arabic.
A March 2021 report by 7amleh that analyzed 574,000 social media conversations in 2020 showed that one out of every 1ten Israeli posts about Palestinians and Arabs contained violent speech, a 16 percent increase compared to 2019. “We have sent reports like this one to Facebook for several years and every year, [but] we find that this content just remains online,” Ramer said, adding that Facebook has not informed them of what, if any, actions it intends to take.
A recent report in the Intercept also noted how Facebook censors the word “Zionist.”
“Zionism is a political ideology,” Ramer said. “Political speech must be protected. Words like “Zionist” and “shahid” [martyr in Arabic] should be protected.” Censorship in the region is especially concerning because of the long-standing lack of transparency around Israel’s treatment of Palestinians, political activist Noam Chomsky told us.
“Israel’s brutal repression of Palestinians for many years, with strong support from the United States particularly, is a shocking crime in itself and has ominous international repercussions as well,” said Chomsky. “There have been extensive efforts to block efforts to bring the facts and their significance to the general public. These efforts amount to direct participation in the crimes.”
When asked about social media companies’ ability to freely censor content, Chomsky replied, “Their enormous power should not be tolerated.”
The Path Ahead
At ARTICLE 19, Marda said that, in order to align itself with international human rights standards, Facebook “must publicly and transparently acknowledge the reasons for recent takedowns” and “provide information for the substantive and legal reasons for takedown.”
Marda added that Facebook should also “restore all blocked content” and “publicly commit to not bowing to governmental or judicial pressure that requires it to act in violation of international human rights standards and jurisdiction-specific standards on freedom of expression.”