Musk has fired or accepted resignations from about three-fourths of Twitter’s employees since his $44 billion takeover at the end of October. He has also terminated thousands of contractors who were monitoring the site for slurs and threats.
Those cuts went deepest outside North America, where more than 75 percent of the company’s 280 million daily users live and where Twitter already had fewer moderators who understood local languages and cultural references ,and where the political landscape could be chaotic and prone to violence.
Musk also welcomed back thousands of banned accounts, including many suspended for promoting hate or violence, even as he has personally has tweeted misinformation and interacted with far-right accounts. Sensing an opportunity — if not a welcome — political operatives and attention-seeking profiteers have rushed to fill the vacuum that the drop in moderation efforts has left, employees said.
That has changed the tenor of the site in its No. 2 market, Japan, where nearly 59 million are estimated to use the site, and made it more fraught in India (nearly 24 million users) and Brazil (nearly 20 million), the third and fourth largest markets, according to current and former staff and researchers. Musk cut virtually all staff in Brazil, allowing an unmoderated surge in misinformation that helped fuel this month’s attacks on the country’s government center.
Even in the better-moderated English-speaking world, the tone of Twitter has become rougher, say those tasked with monitoring the site. Australia’s eSafety commissioner, Julie Inman Grant, who worked at Twitter from 2014 to 2016, told The Washington Post that the platform had already been like a “sewer” in her country before Musk let some of the worst users back on.
“You can’t expect them not to behave like sewer rats, and you probably should expect that further pestilence is going to expand to the user base,” said Inman Grant, who has written the company twice and reminded it that she can order abusive material to be taken down. “It’s becoming a cesspool.”
Australian Electoral Commissioner Tom Rogers said he has been taken aback by vitriolic attacks on a campaign to persuade more Indigenous people to register to vote ahead of a referendum expected next year on whether the legislature should have an advisory council of native people. “We’re watching it very closely. This has been a dry run for what we might see when that referendum occurs,” he said.
Tweets from Sydney-area accounts using two of the most common gay slurs doubled in volume over the last three weeks of November, according to Timothy Graham, a digital media expert at the Queensland University of Technology.
“It is reasonable to argue that Elon Musk’s disintegration of the Twitter safety team and major cuts to moderation is influencing these trends,” Graham said. “At the least, it strongly suggests that Elon’s claims about hate speech reduction are overblown. If anything, the trend is increasing.”
In Australia, Twitter is in the process of laying off even more of its staff.
Entire teams monitoring Asian countries, including Japan, are now gone or nearly so, including those responsible for local curation, trust and safety, and legal issues.
Twitter employees in Singapore and elsewhere in Asia worked on misinformation beyond the immediate region, monitoring content on everything including the war in Ukraine, protests in Iran and major elections in markets where Twitter has less presence. The teams were expanding efforts to identify and remove Russian misinformation around the Ukraine war when their computers were turned off Nov. 4.
“The people who were in the U.S., because of their location, were primarily dealing with the U.S. Outside, we were taking on everything else. There was already that neglect underlying in the company, and now it has been exacerbated,” a former employee who worked on misinformation in the Asia-Pacific region said, speaking on the condition of anonymity to discuss internal matters.
India was a particular focus for Twitter moderation before the layoffs because it was a priority market and because political misinformation there was rife, along with legitimate protest. Moderation for that country has fallen precipitously with staff cuts.
In the past, employees could report any misinformation or hate speech, while curation teams would de-emphasize hashtags that were especially partisan or invited hatred of minority groups.
One former Twitter employee said a Nov. 18 post written by provocateur Suresh Chavhanke would have “definitely” been taken down previously for violating platform’s policy.
The media entrepreneur and militant Hindu nationalist shared a wedding invitation for an interfaith couple that was to take place two days later at an event hall in the Mumbai suburbs — just weeks after the murder of a Hindu woman from the same neighborhood, allegedly by her Muslim boyfriend, made national news.
“How is this still happening even after the heart-rending crime?” Chavhanke demanded to know. He added the hashtag #lovejihad_ActOfTerrorism — a reference to a popular but false claim in India that Muslim men were broadly waging religious war by seducing Hindu women.
The post from Chavhanke racked up more than 5,700 retweets and 10,000 likes, with some of his followers urging local residents to stop the wedding. Activists tried to report his post by tagging and emailing Twitter employees, to no avail. The couple’s relatives were inundated with phone calls, the event hall proprietor told the media outlet Newslaundry. By evening, the families had called off the wedding.
Chavhanke has gone further since then, posting on Twitter a new email address and phone number for tipsters across India to report interfaith relationships.
Alt News and SM Hoax Slayer, two Indian fact-checking and social media advocacy groups, said they tried to alert Twitter through emails about Chavhanke’s post trolling the interfaith wedding. Nothing happened.
“There has to be somebody there to make a determination about takedowns. I’m not sure if they have the bandwidth,” said Pratik Sinha, the co-founder of Alt News.
Current and former employees in India said they worried about both moderation team cuts and the departure of lesser-known teams that improved the platform, including by debunking disinformation.
In the past, said Sinha of Alt News: “They elevated our fact checks multiple times, when the curation team was around.”
During a fraught 2021 election in India’s biggest state, Uttar Pradesh, one employee said Twitter’s content curation team promoted Indian fact-checkers who debunked fake posts alleging ballot box tampering.
That team hasn’t posted anything related to India recently, though forged photos again became an issue in the closely watched Gujarat state elections.
In Japan, Twitter’s largest market outside the United States, staff cuts and other changes have produced a different feel to the service without inflaming passions.
In the past, a small team curated topics of interest and promoted them, which made Twitter an important resource for people to discuss serious issues anonymously. That produced more varied takes on political and cultural issues than did the relatively uniform media.
But with the news curation team gone, trending topics that are surfaced automatically are less about political events and more about entertainment topics, a review of posts found.
Users said topics that disappeared from their timelines in recent weeks included the divisive issue of the U.S. military’s presence in Okinawa and socially sensitive topics like infertility and feminism.
Users reported that they no longer see tweets about issues such as human rights or criticism of the administration of Shinzo Abe, a polarizing political leader who was assassinated in July.
While that disappointed some earnest Twitter users, others were relieved.
“No more political topics trending on Twitter. How peaceful,” one posted.
Shih reported from Delhi, Miller from Sydney and Menn from San Francisco. Shibani Mahtani in Singapore and Julia Mio Inuma in Tokyo contributed to this report.