91�Թ�

A new report from suggests companies hosting interactive online video content in China are censoring that content – from dance performances and comedy routines to songs – based on directives from government.

Their findings are already making headlines and . 

Researchers say platforms such as  , , , and are becoming increasingly popular in China. These platforms, which combine features from traditional chat rooms, online video sharing, and social networking, have collectively amassed over 1 billion registered users. 

One of their most popular uses is to broadcast dancing and singing performances. Sina Show, for example, has more than 26,000 rooms that users access daily, and approximately 270,000 users buy virtual gifts like roses or lollipops for the performers. According to iResearch Consulting Group, Tiange Interaction – the firm which produces Sina Show – leads the social video market with a 33.9 percent market share, and revenues reaching 692 million yuan (approximately 111 million US dollars) in 2014, with an annual growth of 26.3 percent. 

“Despite its enormous user base we know very little about the industry,” said Masashi Crete-Nishihata, research manager at Citizen Lab in 91�Թ�’s Munk School of Global Affairs. 

Like other social media companies operating in China, social video platforms face a complex array of regulations and are held liable for content posted. Companies are expected to invest in the staff and technology required to monitor content and ensure compliance with government regulations, or otherwise face fines or revocation of operating licenses. 

However, there is ongoing debate about the nature of content censorship. Specifically, whether or not government authorities issue directives instructing private companies on how to implement censorship.

Researchers at U of T’s Citizen Lab and the have analyzed how content filtering and monitoring operate on these platforms, in a paper entitled, “.” The paper was a collaborative effort between the Citizen Lab (Masashi Crete-Nishihata, Jason Q. Ng, and Adam Senft) and the University of New Mexico (PhD Candidate Jeffrey Knockel who is currently a research fellow at the Lab, and his supervisor Professor Jedidiah Crandall).

The paper was presented on August 10 at the 2015 USENIX Free and Open Communications on the Internet (FOCI) conference in Washington, DC, which brings together researchers and practitioners who are studying, detecting, or circumventing practices that inhibit free and open communications on the Internet.

“In our research, we reverse engineered four social video platforms (YY, 9158, Sina Show and GuaGua) and found keyword censorship in all four, as well as keyword surveillance capabilities on YY,” said Knockel.

“Each of the four platforms has its own list of censored keywords, which, when combined, result in a total of 17,547 unique keywords that trigger censorship—the largest dataset of sensitive keywords currently available to researchers—which we have translated, contextualized, and grouped into content categories.”

Researcher Ng said that while there is limited direct overlap in unique keywords between platforms, researchers see trends in the topics that are targeted across the lists, including collective action and criticism of the government. 

“This suggests that companies are given general directives from authorities and have a degree of flexibility in the implementation,” said Ng. 

These findings also serve as a counterpoint to , who said that content related to collective action is heavily censored on Chinese social media, while content critical of the government is often allowed to persist. 

“Diversity in censorship implementation leads to diversity in content restricted,” said Crete-Nishihata. “And therefore caution is needed when it comes to applying any comprehensive theory about an ecosystem as varied and fast changing as the Chinese Internet.”