Twitter has removed a total of 2,160 accounts linked to Chinese regional and state propaganda campaigns, the social network has announced as part of its latest data release on misinformation campaigns. The accounts were attempting to push back against allegations of human rights abuses by the Chinese government against the Uyghur population in Xinjiang.
Alongside it, Twitter has also detailed a campaign it discovered in Tanzania, which used copyright complaints to harass members and supporters of the FichuaTanzania human rights group.
Twitter says 2,048 of the accounts “amplified Chinese Communist Party narratives related to the treatment of the Uyghur population,” while a further 112 were connected to a private company backed by the regional government. But according to analysis from the Australian Strategic Policy Institute (ASPI), one of the three research partners which Twitter shared information with, much of the propaganda was “embarrassingly” produced.
According to research from the thinktank reported by The Guardian, each network put out over 30,000 tweets, often disputing evidence of human rights abuses, as well as attempting to push the Chinese government’s version of events. But despite the seriousness of the abuses, much of the data analyzed from the campaign was linked to pornography, Korean soap opera content, and spam accounts, likely because the network had taken over and reused existing accounts. Hundreds of the tweets were linked to an account with the handle @fuck_next, while others tried and failed to tag former US Secretary of State Mike Pompeo.
Most of the accounts had a small number of followers, or none at all, and the overwhelming majority of their tweets had seen zero engagement. The exception was when Chinese officials retweeted them, introducing them to a much broader audience. It’s content that’s unlikely to win over new supporters but is “propaganda appealing to the base,” ASPI researcher Albert Zhang tells The Guardian.
In contrast, the operation linked to Tanzania appears to have been much more sophisticated, although it involved a comparatively smaller number of 268 accounts. In a Twitter thread, a Stanford Internet Observatory researcher who worked on the report, Shelby Grossman, explained that the pro-government network would take anti-government content posted by activists, republish it on an external website with a date that predated the tweet, and then report the tweet to Twitter on copyright grounds to have it removed.
“The tactic sometimes worked,” Grossman writes, “Twitter suspended 2 activist accounts, though both were ultimately reinstated.” But it’s a difficult situation for the activists to end up in, since countering the copyright complaint might compromise the source of the anti-government material.
The treatment of Xinjiang’s Uyghur population has been referred to as a “genocide” and is said to include mass internments, reeducation, forced labor, and even sterilization. Twitter has publicly clashed with Chinese authorities about the human rights abuses before and, in January this year, locked its US embassy’s Twitter account for referring to Uyghur women as “baby-making machines” prior to government intervention. As of this writing, the account appears to still be locked and has not tweeted since January 9th.
As well as these China and Tanzania-linked operations, Twitter says it’s removed accounts related to misinformation campaigns from Mexico, Russia, Uganda, and Venezuela.