Web censorship by governments for political, social or 'national security' reasons is increasing, according to a global survey by the OpenNet Initiative.
John Palfrey, executive director of the Berkman Center for Internet and Society, and clinical professor of law at Harvard Law School, believes the survey shows that online censorship is growing around the world.
"Some regulation is to be expected as the medium matures, but filtering and surveillance can seriously erode civil liberties and privacy, and stifle global communications," he said.
The survey focused on geographical areas, such as Asia, the Middle East and North Africa, and found that 26 of the 41 countries surveyed block or filter internet content.
According to the study, censorship is expanding into new countries and becoming more sophisticated over time.
Countries are not only blocking websites that show pornographic pictures, information about human rights or YouTube videos, but applications such as Skype and Google Maps.
The study found three primary rationales for filtering:
Politics and Power leading to the filtering of political opposition groups, common in many of the countries surveyed
Social Norms leading to filtering of subjects deemed offensive to social norms, such as pornography, gay and lesbian content and gambling, also common in many of the countries surveyed.
Security Concerns leading to the filtering of sites that could endanger national security, such as websites of separatist and radical groups including the Muslim Brotherhood in some countries in the Middle East.
The report claimed that Iran, China and Saudi Arabia not only filter a wide range of topics, but block a large amount of content related to those topics.
South Korea's filtering efforts are very narrow in scope, but heavily censor one topic: North Korea.
Countries engaged in substantial politically motivated filtering include Burma, China, Iran, Syria, Tunisia, and Vietnam, according to the OpenNet Initiative.
Saudi Arabia, Iran, Tunisia and Yemen engage in substantial social content filtering, while Burma, China, Iran, Pakistan and South Korea have the most encompassing national security filtering, targeting websites related to border disputes, separatists and extremists.
No evidence of filtering was found in 14 countries, including Afghanistan, Egypt, Iraq, Israel, West Bank and Gaza, Malaysia, Nepal, Venezuela and Zimbabwe, many of which might be expected to filter internet content.
The 41 countries surveyed were chosen on two criteria: where testing could be done safely (North Korea and Cuba were not included because of security concerns); and where there was the most to learn about government online surveillance.
The research spanned thousands of websites across 120 ISPs, resulting in approximately 200,000 observations.
Countries in Europe and North America were not surveyed because it is largely private companies rather than governments which filter traffic with the aim of eliminating spam, protecting copyright or protecting children from pornography.
A key finding of the study compares the amount of information on a range of topics that is censored, referred to as 'breadth', and the actual content that is blocked, referred to as 'depth'.
"Breadth of filtering represents the ambition of the censors, whereas depth measures success at filtering," said Ron Deibert, associate professor of political science and director of the Citizen Lab at the Munk Centre for internet studies at University of Toronto.
"This can reflect the sophistication of the filtering regime, including the resources devoted to filtering.
"South Korea is an interesting case. It filters very little with the exception of pro-North Korean sites, many of which are hosted in Japan.
"South Korea's thoroughness at blocking these sites shows its determination to eliminate access out of fears of national security."
The research highlighted new and emerging trends in filtering practices, including 'event-based filtering' where content is made inaccessible around elections and other politically sensitive moments.
'Supply side' filtering involves content producers denying access to material to specific geographic locales, while 'upstream filtering' involves filtering occuring outside national jurisdictions.
The OpenNet Initiative intends to investigate internet surveillance in the future, and will develop methods to test for filtering of content available through 'edge locations' such as cyber-cafés, during elections (election monitoring) and from mobile networks, including SMS.
The OpenNet Initiative is a partnership between the universities of Cambridge, Oxford, Harvard and Toronto funded by the John D. and Catherine T. MacArthur Foundation.
Global online censorship rising fast
By Andrew Charlesworth on May 21, 2007 2:00PM