New regulations had been adopted that year to hold tech employees in India criminally liable for failing to comply with takedown requests, a provision that executives referred to as a “hostage provision.” After authorities dispatched anti-terrorism police to Twitter’s New Delhi office, Twitter whisked its top India executive out of the country, fearing his arrest, former company employees recounted.

In the past two years, the Indian government has dramatically tightened its grip on American social media companies. Silicon Valley firms that were at times defiant are now far more accepting of Indian government dictates to censor material, in particular criticism of Prime Minister Narendra Modi and his right-wing Bharatiya Janata Party (BJP).

Indian officials say they have accomplished something long overdue: strengthening national laws to bring disobedient foreign companies to heel.

This escalating censorship in the world’s largest democracy is part of a wider campaign by Modi and his Hindu nationalist allies to monopolize public discourse: tightening their grip on power, advancing their Hindu-first ideology and squeezing out critical and dissenting voices. American technology companies have increasingly fallen in line, fearing for their employees’ security and their profits.

Among the Big Tech companies, the shift has been most notable at Twitter, once seen as Silicon Valley’s flag-bearer for resisting government pressure worldwide. A company that not long ago adopted the risky strategy of fighting government censorship in the Indian courts now consistently bends to official demands. It has repeatedly taken down posts critical of Modi and his administration and accounts belonging to journalists and the BJP’s political opponents.

“The [stuff] that they’re doing in India should be freaking everybody out,” said a former U.S. Twitter policy staffer.

In January, Twitter and YouTube complied with orders to remove links in India to a BBC documentary that faulted Modi, while chief minister of Gujarat state, for allowing the spread of intercommunal riots in 2002 that left more than 1,000 people dead, most of them Muslims. Citing a legal basis for the order, a senior adviser to the Broadcasting Ministry tweeted that the documentary was “hostile propaganda and anti-India garbage.”

In October, Twitter, now renamed X, agreed in India to block the accounts of two U.S.-based groups, Hindus for Human Rights and the Indian American Muslim Council, both nonprofits advocating for pluralism and religious freedom in South Asia.

Twitter became increasingly compliant under immense Indian government pressure even before Elon Musk bought the company just over a year ago. But after he did, Musk proved even less willing to contest takedown orders and discontinued transparency reports about how the company responded to them.

In more than 50 interviews, current and former technology executives and Indian officials detailed how the government broke Twitter’s resistance through a raft of new regulations, a streamlined censorship process — and the coercive muscle of law enforcement agencies. Many spoke on the condition of anonymity to describe such private interactions as well as the 69A committee meetings, which have not been previously reported in detail. (The name 69A refers to the section of the information technology law providing for government censorship.) Some executives wanted to share their concerns about how dire the situation has become and the industry’s complicity in the increasing censorship, while government officials wanted to highlight their success in reining in what they say are irresponsible companies.

Digital and human rights advocates warn that India has perfected the use of regulations to stifle online dissent and already inspired governments in countries as varied as Nigeria and Myanmar to craft similar legal frameworks, at times with near-identical language. India’s success in taming internet companies has set off “regulatory contagion” across the world, according to Prateek Waghre, a policy director at India’s Internet Freedom Foundation.

“India is steadily becoming a norm-shaping country,” said Neeti Biyani, a researcher at the Internet Society, a Virginia-based global internet advocacy organization. “Being the strongest economy in South Asia and one of the strongest emerging economies in the Asia-Pacific, it’s considered one of the first movers on new regulations.” Bangladesh, for example, adopted internet regulations in 2022 that were a “copycat” of India’s, Biyani said.

Despite the huge size of China’s market, companies like Twitter and Facebook were forced to steer clear of the country because Beijing’s rules would have required them to spy on users. That left India as the largest potential growth market. Silicon Valley companies were already committed to doing business in India before the government began to tighten its regulations, and today say they have little choice but to obey if they want to remain there.

“We are toeing the line, not antagonizing the government, knowing very well that this is a government that can come after you,” said an industry official. “All governments in India have been intolerant. But now, they are putting in place the mechanisms and measures. They are going about it in a systematic manner.”

Neither Twitter nor Musk responded to written questions for this article.

Silicon Valley companies once believed that “ideology trumped local law. They have been moved from that delusion,” said Rajeev Chandrasekhar, the deputy technology minister in the BJP government who oversees many of the new regulations, speaking in an interview. “The shift was really simple: We’ve defined the laws, defined the rules, and we have said there is zero tolerance to any noncompliance with the Indian law.”

“You don’t like the law? Don’t operate in India,” Chandrasekhar added. “There is very little wiggle room.”

The Information Technology Ministry did not respond to a list of subsequent questions for this article.

In 2018, Jack Dorsey, Twitter’s co-founder and its chief executive at the time, flew to New Delhi for a visit that would foreshadow trouble to come.

Clad in a black hoodie and beanie, the bearded billionaire posed for a photo with Indian anti-caste and feminist activists while holding a poster that read: “Smash Brahminical Patriarchy!” The picture went viral and incensed the Indian right wing, which viewed Dorsey as the archetype of an elite American trampling over traditional Indian culture. Many denounced Twitter as “racist,” and the company apologized on behalf of its chief executive.

Twitter, Facebook and other U.S. companies were already grappling with how to moderate speech in India. They were seeing a deluge of anti-Muslim and other hateful posts by BJP leaders and their Hindu nationalist supporters, and on occasion the companies called out and pushed back against activity they considered abusive. Prominent BJP supporters accused Twitter of suppressing their online reach and denying them the blue check marks given to other high-profile individuals.

The Modi administration began to ramp up censorship requests, and the 69A committee meetings grew longer, spilling into days, executives recalled. Often, officials simply called Twitter to demand takedowns, industry and government officials said. Records published by the Indian Parliament show that annual takedown requests for posts and accounts increased from 471 to 6,775 between 2014 and 2022, with those to Twitter soaring from 224 in 2018 to 3,417 in 2022.

Government demands to block entire accounts and topic hashtags also soared. Twitter’s transparency reports show that 77 accounts were suspended in the country in 2020. In 2021, there were nearly 1,400. (The company stopped publishing transparency reports after that.)

One former IT Ministry official involved in the orders defended the blocking of accounts. “There are certain accounts that continue to spew venom,” he said. “I have to go by what content you have posted, how much of it is anti-India.”

In late 2020, hundreds of thousands of Indian farmers descended on New Delhi, demanding that Modi withdraw new laws that reduced crop subsidies and price supports and overhauled state-regulated agricultural markets. The farmers, drawn in large part from members of the Sikh religion in Punjab state, occupied highways for months. They brought food, tents, tractors and something unexpected: social media savvy.

The protesters formed an “IT cell” and connected with the Punjabi and Sikh diaspora around the world. They waged a social media campaign that attracted support from NBA players, Greta Thunberg and Rihanna.

The protest movement posed a rare — and unexpectedly serious — challenge to the BJP government. It was a “defining moment,” said Raman Jit Singh Chima, Asia policy director at Access Now, a digital rights group. “The government realized, ‘We need to cement our power in the tech sector,’” he said. “The digital authoritarian turn accelerated in 2021.”

Soon, the censorship orders flooded in. Government officials, who argued that the farmers’ movement harbored ties to separatists seeking an independent Sikh country, saw any posts that could be linked to the secessionist movement as a clear “no-go zone,” a tech executive said.

But the government’s net went far beyond that. Officials balked at posts critical of Modi and started to block journalists.

Sandeep Singh, a freelance journalist who spent months following the farmers’ protest, recalled tracking dozens of Twitter accounts that were taken down. The farmers’ “IT cell” was silenced. So were a news magazine, a Sikh politician and a poet from Canada. “We were shocked to see the scale,” Singh said.

In a public statement, Twitter declared it would not take down accounts of journalists, activists and politicians because the orders were not “consistent with Indian law” or its own “principles of defending protected speech and freedom of expression.”

Twitter’s defiance infuriated the Indian officials, who accelerated the adoption of new rules that in part required social media companies to respond to takedown requests within set time periods, and to appoint a compliance manager, a grievance officer and a liaison with law enforcement, all based in India.

While the farmers’ protest raged in the spring of 2021, a devastating delta variant of the coronavirus swept through India, ultimately killing more than 200,000 people. As criticism of the government response swelled on social media, Modi officials issued waves of new takedown orders. In response, Twitter even removed straightforward commentary, such as the tweet: “Second wave of COVID-19 in India = @narendramodi made disaster. #ResignModi”

When opposition leader Rahul Gandhi accused the prime minister of failing to distribute vaccines and shedding insincere “crocodile tears” about pandemic deaths, the speech set off a flood of anti-Modi and anti-government tweets with the hashtag “#crocodiletears.” In New Delhi, officials ordered Twitter to remove all posts with that hashtag and demanded that the company hand over information identifying users who tweeted it, two former employees recalled.

After Twitter resisted, Indian officials claimed that the hashtag was being used by terrorists and later alleged it was being used to distribute pornography, the former employees said. Indeed, some posts with the #crocodiletears hashtag contained pornographic images. But a Twitter investigator found that some of the posts came from a location near a police building, and a Twitter team concluded that the government itself was planting the images to justify social media restrictions, the former employees recalled.

Tensions between Twitter and the Modi administration reached a boiling point after the company labeled some BJP leaders’ posts as “manipulated media,” meaning they contained images that had been deceptively altered. An elite police force showed up outside Twitter’s New Delhi office in May 2021 with television cameras in tow. The promised raid, however, never materialized because the office was empty amid the pandemic.

In a separate dispute, police in the BJP-ruled state of Uttar Pradesh summoned Twitter India’s managing director, Manish Maheshwari, for questioning over accusations that the platform showed a disputed map of India that did not reflect the country’s declared borders, according to internal company documents. Although Twitter lawyers staved that off, a subsequent police visit to his home and escalating anonymous threats made Twitter’s U.S.-based leadership conclude that it had become “very, very dangerous for him and his family,” a former company executive said. Twitter hurried Maheshwari out of India in mid-2021, and he has not returned.

Maheshwari declined to comment for this article.

In a tweet that May, Twitter highlighted “the use of intimidation tactics by the police” and said it was “concerned by recent events regarding our employees in India and the potential threat to freedom of expression for the people we serve.” The IT Ministry hit back by calling Twitter’s noncompliance “an attempt to dictate its terms to the world’s largest democracy.”

But despite Twitter’s public rebuke of the Indian government, employees inside the company were growing concerned about how much it was bowing to the government’s censorship demands. At an annual presentation to the company’s leadership in 2021, Twitter’s communications team singled out India and delivered a warning about the global precedent the company was setting, according to a person familiar with the presentation.

“The concern was that if we were beginning to make exceptions for a certain government, a whole host of other governments would come to us, and it would be difficult to explain why we could not do the same for them,” the person said.

As India’s new IT rules kicked in, the government told a Delhi court in summer 2021 that Twitter did not have local officers responsible for addressing grievances, coordinating with law enforcement and other tasks mandated by law, and thus had lost its “safe harbor” status, leaving it potentially liable for content deemed illegal. These local officers had to be based in India and could face penalties of five years in jail for failing to comply with government orders.

The threat of arrest “shifts the calculus significantly” for corporate decision-making, said Waghre, of the Internet Freedom Foundation. “You can draw a line in the sand from when the IT rules 2021 went into effect. There was a sudden drop in reported instances of any sort of pushback.” Former Twitter and Facebook executives agreed, saying in interviews that they could not allow colleagues to be jailed.

To meet the government mandate, Twitter hired Vinay Prakash — whose longest previous job had been as a political analyst for Chandrasekhar, then in Parliament and now the deputy IT minister in the BJP government — to be both grievance and compliance officer.

Because of the sensitivity of the position — compliance officers typically have access to internal discussions about legal and human-resource issues as well as user information — Twitter had asked a Florida-based boutique research firm, Divine Intel, to conduct an independent appraisal and background check on Prakash. The firm raised concerns, according to two people familiar with the events.

“Our assessment identified one applicant as high-risk/high-threat and we advised against hiring the individual due to the potential for undue influence from members of Parliament” and other issues, said a Divine Intel executive who spoke on the condition of anonymity to comply with the company’s secrecy policy. The executive said that despite these findings, the person was hired. Divine Intel did not identify Twitter as its client, saying only that it was a major U.S. tech company. But the people familiar with the episode said it was referring to the vetting of Prakash.

Prakash did not respond to requests to comment for this article.

Zatko, who was fired in January 2022, never publicly named the employee he believed to be working for the Indian government. Twitter has denied Zatko’s allegations.

In July 2022, Twitter took the Indian government to court, suing over the continuing demands to censor tweets about the farmers’ protest.

Twitter lost. A judge in Karnataka state sided with the government in June, ruling that posts it wanted censored were “anti-India & seditious … designed to incite violence.”

Twitter appealed. But the company lost much of its appetite to challenge the government after Twitter was acquired a year ago by Musk, former employees said.

“Twitter doesn’t have a choice but to obey local governments,” Musk said in June to Geeta Mohan, executive editor of the India Today news channel. “If we don’t obey local government law, then we will get shut down. The best we can do is really to hew close to the law in any given country, but it’s impossible for us to do more than that or we will be blocked and our people will be arrested.”

Other U.S. companies also opted for a less confrontational approach. After the new IT law was adopted in 2021, for example, Google chief executive Sundar Pichai became one of the first tech leaders to say his company would adhere to it.

“When you comply, you make more revenue. We gained from that,” said a former Google global affairs executive.

Google spokeswoman Christa Muldoon said that the company examined removal requests to see if the content violated local laws, removing it for users in the relevant country, and that revenue was not a factor in decisions. The company reported that the number of items removed from all Google platforms in India soared from about 11,000 in 2019 to more than 23,000 in just the first half of 2023.

Lokman Tsui, who oversaw Google’s free expression program in the Asia-Pacific region in the early 2010s, said the company began a global shift away from its “moral stance” in his region, notably in India, before this spread worldwide.

Now, Indian officials say they want to tighten their regulations further. The “Digital India” bill being drafted is likely to weaken legal protections afforded to the companies for hosting content deemed illegal, according to Chandrasekhar.

“There will no longer be blanket immunity” for social media companies that do not obey, he said. He added that online anonymity and safe harbor protections have been abused to spread harmful misinformation and hate speech.

Other government officials, meanwhile, say it has become easier to make companies take down content because they have turned over many of the decisions to India-based staff, who focus more on complying with local laws and less on company policies. One former IT Ministry official praised the companies for becoming more “understanding” of the government’s perspective, noting that they increasingly comply with its orders.

Just the prospect of further regulation is being used by the government to bend the tech companies to its will, according to Chima, of the Access Now digital rights group. “It’s not only the issue of what actually will get into the law,” he said. “It’s that legal threats are used as a form of negotiation, to get companies to do or not do certain things.”

Gerry Shih contributed to this report.

About this story

Design by Anna Lefkowitz. Visual editing by Chloe Meister, Joe Moore and Jennifer Samuel. Copy editing by Gaby Morera Di Núbila and Martha Murdock. Story editing by Alan Sipress. Project editing by Jay Wang.

The article appeared in the Washington Post.