ºÝºÝߣ

ºÝºÝߣShare a Scribd company logo

Censorship by Chinese Blog-hosting Companies

?
?
rmackinnon
rmackinnon

This is the public version of a presentation I gave at the public seminar on Digital media, civic engagement and political mobilization in repressive regimes 10-11 November 2008, Copenhagen, Denmark

1 of 37
Downloaded 300 times
China?s Censorship 2.0: How companies censor bloggers Rebecca MacKinnon Journalism & Media Studies Centre, University of Hong Kong rmack@hku.hk
2 categories of Chinese Internet censorship
2 categories of Chinese Internet censorship ? ¡°Outside the great ?rewall¡± ? Filtering of websites outside of China
2 categories of Chinese Internet censorship ? ¡°Outside the great ?rewall¡± ? Filtering of websites outside of China ? ¡°Inside the great ?rewall¡± ? Deletion of content on domestic commercial websites ? Takedown of domestically hosted websites ? Shut-down of data centers
Filtering research: Open Net Initiative
Censorship by Chinese Blog-hosting Companies
The ¡°Great Firewall¡± in action
Human Rights Watch: search engine censorship research http://www.hrw.org/reports/2006/china0806/index.htm
Human Rights Watch: search engine censorship research http://www.hrw.org/reports/2006/china0806/index.htm
Human Rights Watch: search engine censorship research http://www.hrw.org/reports/2006/china0806/index.htm
Nart Villeneuve: University of Toronto See: http://www.nartv.org/writing/
Lawyer Liu Xiaoyuan tried to sue blog host for censoring posts ... his lawsuit inspired my project to test censorship by Chinese blog-hosting companies
Censorship by Chinese blog-hosting companies Blog post about ¡°Tiananmen Mothers¡±
Your post ¡°Tiananmen mothers organization publishes a website¡± has been successfully submitted! Because it contains sensitive words, please wait for the community editors to approve it. Please don?t re-post. Thank you.
Password protected blog used to manage project: posted and categorized different kinds of content for testing across 15 different blog-hosting services.
Example of one piece of content for testing: excerpt from a BBC Chinese news story. Link to original website and the full story are included below the test item.
50 subject categories 17. 3 gorges dam 34. discussion of polit reofrm 1. sudden incidents 18. Hong Kong politics 35. legal defence 2. overseas political events 19. Taiwan, general politics 36. human rights critiques 3. olympics 20. Taiwan, independence 37. political arrests 4. historical issues 21. Macau politics 38. independence movements 5. leftist critiques 22. AIDS 39. calls for regime change 6. military/security 23. health issues (non-AIDS) 40. FLG 7. foreign policy 24. crime & criminal cases 41. natural disaster 8. anti-japanese 25. city government policies 42. economic measures 9. anti-U.S. 26. provincial government policies 43. dissidents (unjailed) 10. North Korean refugees 27. national government policies 44. censorship/surveillance 11. foreign trade & investment 28. media/tech policy 45. opposition parties 12. ?nancial and econ data/info 29. national leaders 46. NPC 13. probs. in govt. ministries 30. provincial and city leaders 47. labor issues 14. corruption 31. local leaders 48. migrant workers 15. relocation (due to construction) 32. religious issues 49. economic disparity 16. environment 33. ethnic minorities 50. Tibet
Screenshots were taken at every step of the testing process for every blog post, and uploaded into a database.
15 blog hosts tested, 108 valid tests Company A 60 B 44 C 34 D 31 NOTE: Company names have been replaced with E 27 letters due to concerns that companies who censor F 26 less will be subject to repercussions. Blog services G 24 NOTES: H* 22 I 20 *H - 2 tests deleted the full post: 85 and 115 (both milk powder related); 20 others replaced sensitive words with J 19 ¡°***¡± K 17 **N - Connection reset when visiting blog from Chinese L 13 ISP only. For: 66 (TAM mothers), 126 (milk powder); 117 (¡°Chinese people and religion¡± essay). M 9 ***O: 27 (¡°Why we are suing Yahoo¡± - by dissidents and N** 3 human rights activists) O*** 1 0 27 54 81 108 Number of blog posts tested
The percentages A 55.6% B 40.7% C 31.5% D 28.7% E 25.0% NOTES: Blog services F 24.1% G 22.2% *H - 2 tests deleted the full post: 85 and 115 (both milk powder related); 20 others replaced sensitive words with H* 20.4% ¡°***¡± I 18.5% **N - Connection reset when visiting blog from Chinese J 17.6% ISP only. For: 66 (TAM mothers), 126 (milk powder); 117 (¡°Chinese people and religion¡± essay). K 15.7% L 12.0% ***O: 27 (¡°Why we are suing Yahoo¡± - by dissidents and human rights activists) M 8.3% N** 2.8% O*** 0.9% 0% 100% Percentage of blog posts tested
Baidu: short report about riot in Weng An county, Guizhou
Baidu: published, then removed after 24 hours. Also censored by: iFeng, MySpace, Netease, Tianya, YahooCN, Qzone. (Censored by 7 of 15 blog hosts)
Baidu: clash between protestors and police in Yunnan province. Error message: ¡°Sorry, your article has failed to publish. The article?s content contains inappropriate content, please check.¡±
Netease: report about explosion in Xinjiang, posted successfully in ¡°private view¡± (can be seen when author is logged in to system).
Netease: attempt to access same post from ¡°public view¡±: anybody who is not logged in to the system as the author of that blog. Error message: ¡°This post is not public, you presently cannot view it.¡±
Sina: Same report about explosion in Xinjiang, published successfully for public view...
Sina: post is removed within 24 hours. Error message at same URL: ¡°Sorry, the blog address you visited does not exist.¡± 8 OUT OF 15 BLOG SERVICES TESTED CENSORED THIS CONTENT
¡°Letter to my son¡± wishing for multi-party democracy in China: censored by nobody
Bao Tong essay blasting 1-party system (with his name removed): censored by only 2 of 15
Xinhua: Hu Jintao pep talk to Olympic athletes censored by Mop.com and Blogbus
Xinhua: Hu Jintao pep talk to Olympic athletes censored by Mop.com and Blogbus Mop: ¡°Sorry, your article has been sent to the recycling station, please revise it then publish again.¡±
Xinhua: Hu Jintao pep talk to Olympic athletes censored by Mop.com and Blogbus Mop: ¡°Sorry, your article has been sent to the recycling station, please revise it then publish again.¡±
Why so much variation? ? Instructions to companies from city or provincial State Council Information Of?ce Internet Section, interpreted differently ? Different methods devised for implementation ? What province/city company is registered in ? Manager/editor?s relationship with local State Council ? Background and priorities of individual web company managers and editors ? Relationship between company management, investors, and regulatory bodies
Conclusions ? Internet ?ltering (¡°the great ?rewall¡±) is only one part of Chinese Internet censorship. ? Domestic censorship is not centralized. ? Domestic web censorship is outsourced by government to the private sector. ? Censorship is inconsistent - it?s usually possible to post your content somewhere, for at least a while. ? The system of ¡°managing¡± user-generated web content in China follows similar logic and approach as the system for controlling professional news media.
Implications for research ¡°inside the great ?rewall¡± ? Need larger-scale studies of domestic web censorship (include chat rooms, social networking sites, instant-messaging, mobile services) ? Unlike automated ?ltering tests, these tests require manual testing and constant analysis by Chinese speakers with contextual knowledge - it is tedious work requiring attention to detail. ? Need surveys of web service company employees. ? Need surveys of users and bloggers about their experiences.
Implications for activism ? Circumvention is important but it?s not the solution to the whole censorship problem. ? Educate bloggers and netizens about strategies for successfully disseminating information online about politically sensitive subjects ? Global ¡°user rights¡± movement demanding greater transparency and accountability by Internet companies on privacy and free expression
Global questions ? Where else in the world is this kind of political censorship by web service companies of user- generated content happening? (Companies in the West already censor for child porn, copyright violations and sometimes hate speech.) ? Will the ¡°Chinese model¡± - of demanding censorship by companies - spread globally? ? What issues in this vein should the advocacy community be preparing for? ? What further research needs to be done to better understand global trends?

More Related Content

Censorship by Chinese Blog-hosting Companies

  • 1. China?s Censorship 2.0: How companies censor bloggers Rebecca MacKinnon Journalism & Media Studies Centre, University of Hong Kong rmack@hku.hk
  • 2. 2 categories of Chinese Internet censorship
  • 3. 2 categories of Chinese Internet censorship ? ¡°Outside the great ?rewall¡± ? Filtering of websites outside of China
  • 4. 2 categories of Chinese Internet censorship ? ¡°Outside the great ?rewall¡± ? Filtering of websites outside of China ? ¡°Inside the great ?rewall¡± ? Deletion of content on domestic commercial websites ? Takedown of domestically hosted websites ? Shut-down of data centers
  • 5. Filtering research: Open Net Initiative
  • 8. Human Rights Watch: search engine censorship research http://www.hrw.org/reports/2006/china0806/index.htm
  • 9. Human Rights Watch: search engine censorship research http://www.hrw.org/reports/2006/china0806/index.htm
  • 10. Human Rights Watch: search engine censorship research http://www.hrw.org/reports/2006/china0806/index.htm
  • 11. Nart Villeneuve: University of Toronto See: http://www.nartv.org/writing/
  • 12. Lawyer Liu Xiaoyuan tried to sue blog host for censoring posts ... his lawsuit inspired my project to test censorship by Chinese blog-hosting companies
  • 13. Censorship by Chinese blog-hosting companies Blog post about ¡°Tiananmen Mothers¡±
  • 14. Your post ¡°Tiananmen mothers organization publishes a website¡± has been successfully submitted! Because it contains sensitive words, please wait for the community editors to approve it. Please don?t re-post. Thank you.
  • 15. Password protected blog used to manage project: posted and categorized different kinds of content for testing across 15 different blog-hosting services.
  • 16. Example of one piece of content for testing: excerpt from a BBC Chinese news story. Link to original website and the full story are included below the test item.
  • 17. 50 subject categories 17. 3 gorges dam 34. discussion of polit reofrm 1. sudden incidents 18. Hong Kong politics 35. legal defence 2. overseas political events 19. Taiwan, general politics 36. human rights critiques 3. olympics 20. Taiwan, independence 37. political arrests 4. historical issues 21. Macau politics 38. independence movements 5. leftist critiques 22. AIDS 39. calls for regime change 6. military/security 23. health issues (non-AIDS) 40. FLG 7. foreign policy 24. crime & criminal cases 41. natural disaster 8. anti-japanese 25. city government policies 42. economic measures 9. anti-U.S. 26. provincial government policies 43. dissidents (unjailed) 10. North Korean refugees 27. national government policies 44. censorship/surveillance 11. foreign trade & investment 28. media/tech policy 45. opposition parties 12. ?nancial and econ data/info 29. national leaders 46. NPC 13. probs. in govt. ministries 30. provincial and city leaders 47. labor issues 14. corruption 31. local leaders 48. migrant workers 15. relocation (due to construction) 32. religious issues 49. economic disparity 16. environment 33. ethnic minorities 50. Tibet
  • 18. Screenshots were taken at every step of the testing process for every blog post, and uploaded into a database.
  • 19. 15 blog hosts tested, 108 valid tests Company A 60 B 44 C 34 D 31 NOTE: Company names have been replaced with E 27 letters due to concerns that companies who censor F 26 less will be subject to repercussions. Blog services G 24 NOTES: H* 22 I 20 *H - 2 tests deleted the full post: 85 and 115 (both milk powder related); 20 others replaced sensitive words with J 19 ¡°***¡± K 17 **N - Connection reset when visiting blog from Chinese L 13 ISP only. For: 66 (TAM mothers), 126 (milk powder); 117 (¡°Chinese people and religion¡± essay). M 9 ***O: 27 (¡°Why we are suing Yahoo¡± - by dissidents and N** 3 human rights activists) O*** 1 0 27 54 81 108 Number of blog posts tested
  • 20. The percentages A 55.6% B 40.7% C 31.5% D 28.7% E 25.0% NOTES: Blog services F 24.1% G 22.2% *H - 2 tests deleted the full post: 85 and 115 (both milk powder related); 20 others replaced sensitive words with H* 20.4% ¡°***¡± I 18.5% **N - Connection reset when visiting blog from Chinese J 17.6% ISP only. For: 66 (TAM mothers), 126 (milk powder); 117 (¡°Chinese people and religion¡± essay). K 15.7% L 12.0% ***O: 27 (¡°Why we are suing Yahoo¡± - by dissidents and human rights activists) M 8.3% N** 2.8% O*** 0.9% 0% 100% Percentage of blog posts tested
  • 21. Baidu: short report about riot in Weng An county, Guizhou
  • 22. Baidu: published, then removed after 24 hours. Also censored by: iFeng, MySpace, Netease, Tianya, YahooCN, Qzone. (Censored by 7 of 15 blog hosts)
  • 23. Baidu: clash between protestors and police in Yunnan province. Error message: ¡°Sorry, your article has failed to publish. The article?s content contains inappropriate content, please check.¡±
  • 24. Netease: report about explosion in Xinjiang, posted successfully in ¡°private view¡± (can be seen when author is logged in to system).
  • 25. Netease: attempt to access same post from ¡°public view¡±: anybody who is not logged in to the system as the author of that blog. Error message: ¡°This post is not public, you presently cannot view it.¡±
  • 26. Sina: Same report about explosion in Xinjiang, published successfully for public view...
  • 27. Sina: post is removed within 24 hours. Error message at same URL: ¡°Sorry, the blog address you visited does not exist.¡± 8 OUT OF 15 BLOG SERVICES TESTED CENSORED THIS CONTENT
  • 28. ¡°Letter to my son¡± wishing for multi-party democracy in China: censored by nobody
  • 29. Bao Tong essay blasting 1-party system (with his name removed): censored by only 2 of 15
  • 30. Xinhua: Hu Jintao pep talk to Olympic athletes censored by Mop.com and Blogbus
  • 31. Xinhua: Hu Jintao pep talk to Olympic athletes censored by Mop.com and Blogbus Mop: ¡°Sorry, your article has been sent to the recycling station, please revise it then publish again.¡±
  • 32. Xinhua: Hu Jintao pep talk to Olympic athletes censored by Mop.com and Blogbus Mop: ¡°Sorry, your article has been sent to the recycling station, please revise it then publish again.¡±
  • 33. Why so much variation? ? Instructions to companies from city or provincial State Council Information Of?ce Internet Section, interpreted differently ? Different methods devised for implementation ? What province/city company is registered in ? Manager/editor?s relationship with local State Council ? Background and priorities of individual web company managers and editors ? Relationship between company management, investors, and regulatory bodies
  • 34. Conclusions ? Internet ?ltering (¡°the great ?rewall¡±) is only one part of Chinese Internet censorship. ? Domestic censorship is not centralized. ? Domestic web censorship is outsourced by government to the private sector. ? Censorship is inconsistent - it?s usually possible to post your content somewhere, for at least a while. ? The system of ¡°managing¡± user-generated web content in China follows similar logic and approach as the system for controlling professional news media.
  • 35. Implications for research ¡°inside the great ?rewall¡± ? Need larger-scale studies of domestic web censorship (include chat rooms, social networking sites, instant-messaging, mobile services) ? Unlike automated ?ltering tests, these tests require manual testing and constant analysis by Chinese speakers with contextual knowledge - it is tedious work requiring attention to detail. ? Need surveys of web service company employees. ? Need surveys of users and bloggers about their experiences.
  • 36. Implications for activism ? Circumvention is important but it?s not the solution to the whole censorship problem. ? Educate bloggers and netizens about strategies for successfully disseminating information online about politically sensitive subjects ? Global ¡°user rights¡± movement demanding greater transparency and accountability by Internet companies on privacy and free expression
  • 37. Global questions ? Where else in the world is this kind of political censorship by web service companies of user- generated content happening? (Companies in the West already censor for child porn, copyright violations and sometimes hate speech.) ? Will the ¡°Chinese model¡± - of demanding censorship by companies - spread globally? ? What issues in this vein should the advocacy community be preparing for? ? What further research needs to be done to better understand global trends?