Over at Hot Air, Jazz Shaw wrote this week about the hypocritical way Twitter hands out suspensions to users over tweets critical of the Chinese Communist Party or Chairman Xi Jinping.
The story begins with Twitter’s decision to “flag” the account of New Zealand university professor Anne-Marie Brady, who posted what could best be described as a relatively mild criticism of the CCP leader and the Chinese celebration of the anniversary of the country’s communist takeover.
When the professor posted, “Xi: it’s my Party, and I’ll cry if I want to,” several of her tweets received the “unavailable” tag the company applies to tweets it wishes to censor. A group of journalists joined with the professor in questioning the censorship, and the posts were eventually restored. Twitter maintained after the fact that the flagging occurred because of “unusual activity” on the account.
Shaw cites Edward Lucas of the Associated Press in addressing the general nature of Twitter censorship and whether it occurs because of reviews by actual people or as a result of overactive algorithms.
Lucas theorized that the censorship of the professor was likely the result of an online flood of complaints set in motion by the CCP itself. Shaw finds the theory plausible, given that when Twitter receives a large number of user complaints about posts or an account, the account is sometimes automatically flagged until it is individually reviewed. Suppose the CCP has the capacity to respond to a critical tweet with a rush of bot-generated complaints. In that instance, the party may persuade Twitter to suspend or flag an account that is displeased.
Shaw decided to test the Lucas theory with a simple experiment. On July 6, he crafted a tweet intentionally designed to trigger automatic algorithms, assuming they exist and operate as expected. He wrote, “Chinese President #XiJinping is engaging in genocide against the Uyghurs and stonewalling the research into the origins of the pandemic.”
Shaw noted that the tweet was not flagged and concluded that while he didn’t trust Twitter’s claim to be dedicated to free speech, he thought that it was likely that the professor’s tweets caught the attention of CCP bots while his did not.
While this experiment, of course, is not conclusive proof of anything, the fact remains that Twitter’s methods of choosing what content to censor and what to let pass are opaque and completely unpredictable.
One can hope that President Trump’s new class-action lawsuit aimed at Big Tech censorship in violation of the First Amendment may shed some light on how social media operates.