BEIJING — The most sensitive day of the year for the Chinese interwebs is drawing near, and putting China's robot censors on high alert.
June 4, 2019 marks the 30th anniversary of the Tiananmen Square protests, or the bloody massacre that never happened, as far as China is concerned.
The Guardian reports that in 1989, government tanks violently cleared student-led protests for democratic reform in a deadly crackdown. No official death toll has ever been released, but estimates range from several hundred to several thousand.
Since then, Beijing has been systematically erasing all evidence and memory of the bloody protests, and blocking any associated content with increasingly hi-tech precision.
According to a Reuters report, Chinese censors have reached a new level of accuracy, thanks to machine learning and voice and image recognition.
Censorship of the Tiananmen protests and other forbidden topics is largely automated.
Posts that include related dates, names, and images are automatically rejected.
Reuters reports that the periods leading up to a political event or anniversary, like this one, are especially sensitive. This year, social media censorship has even targeted LGBT groups, NGOs, and labor and environmental activists.
Four censors working across Chinese internet companies Beijing Bytedance, Weibo, and Baidu censor 5,000 to 10,000 pieces of information daily, or about five to seven pieces per minute.
Companies are responsible for censoring their own platforms. For news, they play it really safe by sticking to a simple rule that it's not authorized if it's not from state media.
Otherwise, they run the risk of being punished for failing to properly censor, and hit with days to weeks-long suspensions, which is what happened to the Netease news app, Tencent news app Tian Tian, and Sina Corp over the past six weeks.
It's not any better for individual internet users, who can receive penalties about sensitive issues online.