Online, words are a form of data and expression. From hashtags to political dissent, words have the power to build new worlds and take down old ones. At the same time, language has also become a form of data, used to create machine learning systems for profit, and it has also become an arena for automated censorship and moderation.

In the PRC, automated censorship has led to a surge of creativity as online netizens scramble to “fool the machine”, through creative use of homophones to images and new characters that bypass OCR (optical character recognition).

The Algorithmic Censorship Resistance Toolkit is a constructed and collected toolkit of different tactics that obfuscate and encode text so that it becomes machine unreadable. These tools derive inspiration from existing online practices, as well as proposed ones through research.

Background Research

Algorithmic Censorship by Social Platforms: Power and Resistance.

Jennifer Cobbe, 2019.

(Can’t) Picture This: An Analysis of Image Filtering on WeChat Moments.

Jeffrey Knockel, Lotus Ruan, Masashi Crete-Nishihata, and Ron Deibert, 2018.

Resisting the Censorship Infrastructure in China.

Yubo Kou, Yong Ming Kow, and Xinning Gui, 2017.

The effect of information controls on developers in China: An analysis of censorship in Chinese open source projects.

Jeffrey Knockel, Masashi Crete-Nishihata, and Lotus Ruan, 2018.

Tranßcripting: playful subversion with Chinese characters.

Li Wei, Zhu Hua, 2018.




An ongoing collection of
algorithmic censorship resistance tactics