Bengaluru: A global foundation that combats racism and promotes climate justice found Instagram had deleted or censored its posts. An artist found her artwork criticising a ruling party member of Parliament (MP) from Bengaluru and requests for oxygen and hospital beds had disappeared from Twitter. Another found Instagram critiques of the government and a fundraiser for the trans-community in rural Tamil Nadu missing.

These events, not accompanied by intimations from social-media companies of potentially problematic online behaviour, are called shadow bans, where posts are only visible to users, invisible to everyone else. There may not be a violation of the terms of service, but an algorithm curbs the visibility or engagement with a post, largely, said experts, to shape an unofficial narrative that usually hews to the official line.

“When they (social media platforms) take down a post, they are required to follow certain procedures,” said Tanmay Singh, associate litigation counsel at Internet Freedom Foundation (IFF), an advocacy group. “But shadow-banning allows them to circumvent these legal considerations.” 

Shadow bans have unfolded in tandem with police cases (here, here and here) in the last week of June against Twitter and a host of social-media users, such as a first information report against seven Muslim journalists and The Wire, a news website, and government requests for takedown of content.

India leads the world in takedown requests, accounting for nearly a fifth of such removal requests over nine years to July 2018, rising over 200% from 6,843 requests in 2013 to 20,805 in 2018, according to this report from Comparitech, a UK-based technology research company. In 2020, Internet freedom in India “declined dramatically”, said a report from Freedom House, a global think tank.

As the government of Prime Minister Narendra Modi increases pressure on social media companies, especially Twitter, to conform to India’s new digital media rules, companies, including Facebook, Twitter and Instagram, which have 320 million, 18.8 million and 120 million active users in India, respectively, appear to be censoring content to avoid government attention.

Issued in February 2021, the new digital media rules—officially called the Information Technology (Guidelines for Intermediaries and Digital Media Ethics Code) Rules, 2021—provide the government with a raft of new powers to block content and have been criticised as a step to “digital authoritarianism”. 

These rules faced objections from the government’s own advisers, who cautioned they were beyond the scope of existing law, as Article 14 reported on 17 May 2021 and The Morning Context reported on 2 July 2021. They have now been challenged as unconstitutional in the Delhi, Karnataka and Madras high courts.

Meanwhile, social-media companies have begun the process of pre-censorship.

On 29 June, Facebook told a Parliamentary committee that, in compliance with the new rules, it would publish a report on content it had “proactively removed” between 15 May and 15 June using “automated tools”.

“This report will contain details of the content that we have removed proactively using our automated tools”, a Facebook spokesperson said in a statement. Being pro-government is not a new charge: in 2020, the company was accused of taking down pages related to the farmers' protest.

Content removals or suspending users on government request have increased over the last couple of years. 

Singh of the IFF said that the blocking rules of 2009 required the government to issue a takedown order and provide “sufficient time” and hear the defence of the “originator” or user, which it “almost never” followed.  

Article 14 sought comment from Facebook and Twitter over the allegations of shadowbans over political content, but there was no response.

The Case Of the Missing Artwork

Early in May 2021, Instagram users saw their stories and posts that amplified Covid-19 resources disappear overnight. 

Instagram user @kanmaniponmani, who requested her identity be withheld, told Article 14 that she noticed in the first week of May 2021 that her stories reposting details to a fundraiser for the trans-community in rural Tamil Nadu had gone missing. 

Initially, she assumed the originator had taken down the post, but after repeated disappearances of stories and direct messages dealing with Covid-19 resources and relief she tried to find out what had happened. 

She could not find her posts critiquing the government’s handling of the crisis. She posted about this, and found several users corroborating her claims with similar stories. “Some stories, DMs are back. I want to believe in the ‘global glitch’ narrative but as I’ve shared on my stories, Palestinian activists for instance will tell you a diff(rent) story” she wrote. 

Another user, @smishdesigns said all her stories containing SOS requests, verified leads for oxygen, hospital beds had disappeared, as had artwork criticising BJP MP Tejasvi Surya for religious profiling in Bengaluru. In the past, similar posts targeting the ruling party at the Centre were deleted. Unlike @kanmaniponmani’s, her stories have not been restored.

Instagram, which is owned by Facebook, has made no statement yet on the issue of disappearing messages and shadow bans. Article 14 sought comment from the company’s public relations managers but there was no response.

Growing Shadows Of Shadow Bans

Shadow-banning is a widening issue, as platforms pass on accountability to “sentient” algorithms. This was reflected in the supposed glitches that Instagram and Twitter encountered when hundreds of posts protesting forceful evictions of Palestinians by Israelis forces in May 2021 were removed. 

In May, 2021, Slow Factory, a nonprofit focused on antiracism and climate justice, alleged Instagram had been deleting and censoring their posts. 

“All stories related to missing and murdered indigenous women, girls, trans and two-spirit folks shared on 6 May 2021 have been deleted,” said a Slow Factory statement. “Big Tech across all social media continues to deplatform, devalue and denounce efforts of labour that don't align with their racist agenda”.  

Responding to allegations of shadowbanning, Instagram chief Adam Mosseri said in a blog on 8 June 2021 that posts or content were only taken down following “violations to community guidelines” or “misinformation”. 

“We understand people are inevitably going to come to their own conclusions about why something happened, and that those conclusions may leave people feeling confused or victimized,” said Mosseri. 

Shadow-bans enable censorship in collaboration with governments, said experts. Governments leverage their authority to compel an intermediary, such as a social media company, to censor content. 

Udbhav Tiwari, public policy advisor at non-profit Mozilla Foundation said there was “little to no transparency” whether shadow-bans were executed on government request or if the practice was regulated. “Platforms need to be more transparent about number of accounts shadowbanned or done at the behest of governments,” said Tiwari. 

In a Twitter live on 20 May 2021 organised by think-tank The Dialogue, Tiwari said content takedowns and compliance burdens “severely disincentivise” digital-media operations in India. 

As per the new rules, requests for content takedowns must be acted upon within 48 hours. “This higher threshold means platforms will comply with requests without evaluating merits of the claims,” Tiwari said. 

Intermediaries find themselves having to choose between protecting the privacy and freedom of expression of users’ and comply with government's requests. “Platforms should not be forced to make such binary calls. It discourages them from launching or offering services such in India,” Tiwari said. 

UN Special Rapporteur Criticises India’s Social-Media Policing

The Indian government’s policing of social media was evident on 24 April 2021, when it ordered Instagram, Twitter and Facebook to remove over 100 posts that criticised the government’s efforts in curbing the devastation caused by Covid-19. 

In June, social media sites were ordered to remove content that alluded to the Covid-19 delta variant as the ‘Indian variant’. As #ResignModi gained momentum, social media posts on Facebook amplifying this and other hashtags were temporarily blocked. The government claimed these hashtags were “creating panic” and spreading “misinformation”.

On 11 June 2021, the United Nations special rapporteur for freedom of expression and opinion Irene Khan said in a letter to the Indian government: “We express serious concern... parts of due diligence obligations may result in limiting or infringement of a wide range of human rights. We encourage withdrawal, review and reconsideration of certain key aspects of this legislation to ensure...compliance with ... government's international human rights obligations,” said the letter—urging a reconsideration or withdrawal of “certain key aspects”. 

In response, citing section 69A of the Information Technology Act, the government said the recent rules for intermediaries were necessary “due to... increased instances of abuse of social media and digital platforms”. 

A new rule on traceability makes it incumbent on intermediaries to identify originators of content. “India’s particular provision on traceability of end-to-end encrypted messages is the first such in the world,” Tiwari said. 

The Chilling Effect

The chilling effect of state crackdown was evident as India’s faltering health infrastructure collapsed in April 2021 during the Covid-19 pandemic’s second wave. Social media emerged as a powerful tool to address desperate searches for oxygen and frantic calls for hospital beds. 

Through April and May, social media turned into a resource hub where young volunteers coordinated, connected and collaborated nationwide to link oxygen suppliers, hospitals and patients’ families. 

Kaushik Raj, 22, of Citizens’ Aid Collective, was one such frontline volunteer. “I hadn’t imagined the magnitude of people who needed help,” he said. 

In April 2021, efforts of the collective led to a spreadsheet of verified resources and suppliers across the country. They estimated this resource benefitted more than 2,000 Indians. 

In late April, their team of 50 volunteers heard rumours of police interrogating volunteers involved in relief work and cautioning them against black marketeering. When they heard police had contacted a similar collective, Raj thought it best to slow operations. 

On 25 April 2021, UP chief minister Yogi Adityanath announced that properties of those “spreading panic” on social media would be seized, after which the collective temporarily ceased operations. 

“We thought let’s pause a minute, see what is the best way forward,” Kaushik said. 

On 25 April, content creator Andre Borges posted that an anonymous individual claimed the Delhi police threatened action against their volunteers for circulating Covid-19 resources. The police denied these allegations and accused “motivated elements on social media” of misinforming and “obstructing constructive work”. 

But the wariness was hard to shake off, as reports kept coming in, especially from UP,  of police crackdowns on people seeking help online and volunteers.   

The offline crackdown impacts online activity. Tiwari said people refrain from expressing themselves or censor their speech in conversations otherwise private. “The environment of fear that mass or targeted surveillance creates has a chilling effect” he said. 

States Launch Own Crackdowns

While the Supreme Court frowned on crackdowns, it was not the Centre alone that felt the need to restrict and curtail. 

In BJP-run Madhya Pradesh, the Indore administration on 11 May 2021 threatened prosecution under section 144 of the Code of Criminal Procedure, 1973, against those forwarding or commenting about Covid-19 on social media. 

The Jabalpur district administration curbed Covid-19 posts that were “objectionable, inciting, gory” or critical of the state. The order was repealed following an application by the IFF under section 144 (5). 

“Last night (12 May 2021), on the basis of an application we filed,” said an IFF statement on Instagram, “He (district collector) has repealed the problematic provision and replaced it with a clearer, more limited order that no longer threatens to prosecute citizens of Indore under the Epidemic Disease Act or National Disaster Management Authority.” 

How New IT Rules Pressure Social-Media Companies

The government has been unresponsive to criticism of the new intermediary rules.

 “Data collection for surveillance is legitimised by an acceptable public goal and the encroachment of the state into private lives is routinised,” columnist Ispita Chakravarty wrote in Scroll.in. 

India was found to be in the bottom five on non-European Union surveillance societies due to its “systemic failure to maintain safeguards,” according to a 2019 study. 

Delhi-based lawyer Raghav Tankha described the rules as “another tool in the hands of the government to use law enforcement agencies to go after individuals who express opinions…contrary to its ideas and interests”. 

In February 2021, when the government ushered in the rules for digital media and intermediaries, several experts said they appeared unconstitutional and that content takedowns must be subject to democratic and judicial processes, not governmental whims. 

Section 66A of the IT Act that allowed for prosecution based on “offensive, insulting” or even for content “causing annoyance” was “judicially challenged and eventually struck down in the Shreya Singhal vs Union of India case of 2015,” said Singh. 

“But the new rules are antithetical to the Supreme Court order as it allows the government a large amount of discretion in what they believe is acceptable and non-acceptable facets of speech,” he said. 

The new rules ensure greater government control over intermediaries. Non-compliance will result in the loss of a “safe harbour” promised under section 79 of the IT Act and make intermediaries legally liable for content posted by a third party. 

That is what happened when Twitter did not follow the new rules to appoint a grievance officer, a compliance officer and a nodal 24x7 contact officer. On 25 May, a government deadline passed and Twitter “lost” its safe-harbour protection. 

Within hours police in the UP district of Ghaziabad filed an FIR against Twitter and nine others over a video on Twitter. Since then, Twitter has had cases filed against it by police in Delhi—for allowing child porn and in Madhya Pradesh for showing a map with Jammu and Kashmir and Ladakh outside of India.

“This is scary. Now social media sites will take down content before the government can get wind of it,” said IFF’s Singh. “Censorship and surveillance are both shrouded by the same curtain—severe lack of transparency.” 

This loss of digital freedom and control exercised on the Internet has come at a time when the Internet has emerged as essential infrastructure. Among those who extensively used social media to help during the pandemic was Sanjeev Gupta, secretary of the inter-state council secretariat at the union home ministry. On 29 June, Gupta tweeted from @sanjg2k1: “I've done 36 years of service in IAS.... I got active on Social Media only to save lives & help poor... since Wave 1...”. 

Twitter’s resistance to comply with the new rules is being watched by industry players and users, who regard it crucial for the defence of freedom of expression. “It will be just like a set of dominoes that fall the moment the government makes a request,” said Tiwari, “Unless they resist.” 

With contributions from Flavia Lopes