Listen Now

Outsourcing Censorship

Internet intermediaries are being used by Asian governments to control speech online argues Kieran Bergmann.

By: /
4 September, 2012
By: Kieran Bergmann
Google Policy Fellow at the Citizen Lab, Munk School of Global Affairs, University of Toronto

The growth of Internet usage in Asia is putting many governments on the defensive. They are becoming increasingly aware of the challenges that online connectivity poses to regime stability. The Internet has provided their citizens and political opponents with a new avenue to discuss contentious topics, network with like-minded individuals, and organize and mobilize opposition and protest movements. It gives the populations of these countries a powerful tool to push the limits of freedom of expression, association, and information that many of their governments set long ago.

In response, governments are getting creative in their efforts to control online content. An increasingly popular technique is to outsource content control to intermediaries, holding them accountable for third-party content. Internet intermediaries are the entities that provide the platforms and channels for online communication and access to information, including Internet service providers (ISPs), web hosting companies, blog and video hosting sites, social networks, webmasters, and any other online service provider. Disturbingly, they are the ones increasingly facing punishment, often disproportionately severe, for failing to censor content in Asia. They are being punished for content that they did not create or publish. The fact that some of Asia’s most influential economies have adopted this angle of attack makes the potential scale of this problem very serious indeed.


Made [and Censored] in China

China is certainly the most notorious Internet censor in Asia. The government’s methods of online control – which include a robust legal framework, informal practices, and strict regulation of the private sector – are amongst the most restrictive in the world. The country’s requirement that companies exercise “self-discipline” if they wish to operate in the country obliges intermediaries to censor “unacceptable” content and monitor their customers on the government’s behalf. Most companies cannot afford to be left out of the enormous Chinese market, which is home to the largest number of Internet users in the world, and therefore comply with the demands. Rather than risk being shut out by “The Great Firewall,” they accept, and thus indirectly support, the country’s repressive rules for controlling free speech online. China has proven incredibly adept at outsourcing its online content control to intermediaries, and is now home to the most advanced and efficient system of intermediary liability in Asia.

Cases like the following are now all too familiar: In 2004, Yahoo!’s Hong Kong office caved to pressure from the Chinese government to turn over personal information about one of its users, a journalist named Shi Tao. Tao had used his Yahoo! email account to leak a government memo that pressured Chinese journalists not to cover the protests commemorating the 15th anniversary of the 1989 Tiananmen Square Protests. After identifying Tao using the information provided by Yahoo!, the government charged him for leaking state secrets, sentencing him to 10 years in prison.

This phenomenon is not unique to China – governments across Asia, including those of Thailand, India, and Malaysia, are now also using intermediaries to censor online content, monitor Internet users, and restrict their citizens’ ability to access information online.

Thailand’s Trial Run

In recent months, Thailand has demonstrated the effectiveness of intermediary liability in suppressing user-generated content. On May 30, Chiranuch Premchaiporn, the director and editor of Prachatai, a popular Thai online newspaper, was found guilty of violating Thailand’s lèse-majesté laws, which protect the much-revered king and the royal family from any actions that might damage or defame them. Lèse-majesté has been an offence against national security since 1957, but is now enjoying a cyber-security-driven surge in popularity. Chiranuch was also found guilty of breaching the more modern, but equally stifling, Computer Crime Act (CCA), adopted in 2007. Combined, these two legal documents make intermediaries, such as webmasters and ISPs, liable for online content deemed to be defamatory to the monarchy. The charges against Chiranuch related to 10 alleged lèse-majesté comments posted by users on Prachatai’s web board that she, as webmaster, failed to remove in a timely fashion.

While working in Thailand this past fall, I was asked to be an international observer at Chiranuch’s trial. A number of civil-society groups organized a campaign to secure her a fair trial. Each of the 10 charges she faced carried a maximum sentence of five years imprisonment or a fine of no more than 100,000 baht (approximately CAD$3,300) – a hefty punishment for any crime, but certainly disproportionate in a case where the accused was not even the author of the offensive content. Ultimately, she was sentenced to one year in prison, which was subsequently reduced to eight months and then suspended, and a 20,000 baht (approximately CAD$644) fine.

Chiranuch’s relatively lenient sentence came at a high price: She chose to plead guilty to all the charges against her and co-operate with the prosecution. Luckily for Chiranuch, she also had a history of complying with the Ministry of Information and Communication Technology’s (MICT) requests that she delete user comments from Prachatai’s web board. Prior to charges being laid against her, the ministry sent her almost daily emails containing a list of URLs that it had judged to be illegal, and she deleted them immediately.

Monitoring a website the size of Prachatai was no small feat, however, particularly after the 2006 coup, when the web board saw a tenfold increase in visitors. The opinions and posts became more aggressive, reflecting the growing political tensions in the offline world. Prachatai was soon in serious trouble, despite instituting more stringent moderation standards, including requiring users to register with a valid email address and giving volunteer moderators the authority to remove posts. The number of posts was just too great – Chiranuch’s job as webmaster became impossible. She was expected, when traffic to the site peaked, to monitor 20 to 30,000 new posts each day, on 300 new daily topics.

On Oct. 5, when the last alleged lèse-majesté comment was posted, the web board hosted more than one million messages. The pressure from government censorship, when combined with the charges against its webmaster, overwhelmed Prachatai. The web board was shut down in July 2010.

Old Laws, New Tactics

The ambiguity of the laws in Thailand that place legal responsibility for third-party content on webmasters and ISPs is incredibly problematic. Section 112 of Thailand’s Criminal Code, which contains the lèse-majesté provisions, prohibits any act that “defames, insults or threatens the King, Queen, the Heir-apparent or the Regent.” Yet, nowhere is there a clear indication of what sort of act would constitute such an offence. This overbroad definition leaves the interpretation of the law up to ordinary citizens (who are permitted to lodge complaints), the Royal Thai Police, the MICT, and the courts. While the MICT and the national police have held meetings with ISPs and webmasters, including Chiranuch, on how to monitor websites for inappropriate content, there is no jointly agreed upon understanding of the law. In fact, to merely discuss and debate lèse-majesté laws in court or in a public forum can lead to punishment for violating said law. During Chiranuch’s trial, the actual content of the supposedly illegal comments that she was charged for failing to remove was not even discussed.

The Computer Crime Act, which extends the criminalization of lèse-majesté to Internet intermediaries by prohibiting “intentionally supporting or consenting for crimes linked to the use of computers,” is equally vague. The CCA does not provide any rules or regulations as to what constitutes a reasonable amount of time for “offensive” content to be posted online. It is thus extremely difficult to defend yourself or appeal against any violations of the act, as any interested party can interpret the law to suit its agenda.

The Thai government has also shown it is not prepared to rely only upon domestic actors to enforce its ambiguous censorship laws. In order to censor online content more efficiently, Thailand has enlisted intermediaries outside of the country, such as Google, Facebook, and Twitter. Consider Google’s most recent Transparency Report. This report details requests from governments to remove content between July and December of 2011, and reveals that in this period, Google received requests from the MICT to remove 149 YouTube videos that were allegedly insulting to the monarchy. This number represents a huge increase in the country’s removal requests from any previous reporting period. As a policy, Google must comply with local laws in countries where it operates. The corporation therefore removed 70 per cent of these videos from view in Thailand. More troubling than the sheer number of removal requests is the fact that none was accompanied by a court order – instead, Google categorized them as “other” requests from the “executive, police, etc.” This indicates that the Thai government is successfully bypassing formal and lawful processes in an effort to secure the compliance of private-sector companies in its Internet censorship activities.

The Thai government is clearly not afraid to make demands of the corporate Internet heavyweights. On Nov. 24, 2011, the minister of ICT announced that since August, the government had requested that Facebook remove the URLs of 86,000 user accounts that contained lèse-majesté comments, photos, audio clips, or videos. Unlike Google, Facebook does not publicly release its takedown requests, and it is therefore unclear how many of these requests the company complied with.

Intermediary liability has hindered free speech in Thailand. The lack of clarity surrounding lèse-majesté laws and the CCA is prompting many intermediaries to air on the side of censorship in order to protect themselves from the harsh punishments that they face. Chiranuch’s sentence appears fairly mild when compared with other recent sentences for violations under lèse-majesté laws and the CCA. A 60-year-old man named Ampon Tangnoppakul, better known as Uncle SMS, was sentenced to 20 years in prison on Nov. 23, 2011, for sending four text messages deemed to be insulting to the Queen – a charge he adamantly denied. He died in prison on May 8, 2011. Another citizen, Suwicha Thakor, was sentenced to 10 years in prison in April 2009 for posting a picture online that was deemed to be insulting to the monarchy, although he was given a royal pardon in June 2010.

While Chiranuch’s sentence may seem mild in comparison to these, her guilty verdict did send a powerful and unnerving message to other intermediaries: Censor content to whatever extent the government demands, or you will be held accountable. We are therefore likely to see an increase in pre-emptive self-censorship, leading to sweeping restrictions on freedom of speech across the country.

India’s Take on the Rules

These types of restrictions are not unique to Thailand. In 2008, India amended its 2000 Information Technology Act (ITA), apparently to strengthen its protection of intermediaries. The new provisions established a “notice and takedown” regime. Under this regime, intermediaries are protected from liability as long as they remove illegal content when they are made aware of its existence. In 2011, this framework was supplemented by the Information Technology Rules, which require intermediaries to remove “objectionable” content. The alleged objective of these provisions is to protect freedom of expression and opinion, but they in fact do the opposite by putting the onus of censorship on intermediaries.

As in the case of Thailand, the content that the IT Rules deem objectionable is left dangerously open to interpretation, covering such vague categories as “harmful,” “defamatory,” “obscene,” “harassing,” “blasphemous,” or “disparaging.” They also extend the definition of “objectionable” to anything that “threatens the unity, integrity, defence, security or sovereignty of India, friendly relations with foreign states, or public order.” Unlike in Thailand, however, India’s IT Rules set forth a precise timeframe within which intermediaries must remove the content – 36 hours from the time a complaint was lodged or authorities gave notification of an offence. This leaves only a very small window of opportunity. The likely consequence is that the intermediaries censor first, investigate later, minimizing any real ability to deem whether the content is, in fact, illegal before taking it down.

Equally problematic is the fact that intermediaries are under no obligation to inform the individual that posted the relevant content of the takedown or the reasoning behind it. Nor is there any sort of recourse for the restoration of removed content. The rights of the content creator are clearly not a consideration.

No intermediary that hosts third-party content in India is free from the IT Rules. Google, Yahoo!, Facebook, and Microsoft have all been taken to court on charges that they hosted offensive content. According to Google’s Transparency Report, in 2011, the Indian government requested that 615 items be removed from the company’s services.

Regional Transformation

Developments in Malaysia and Vietnam complete this disturbing trend. The government in Kuala Lumpur has enacted troubling amendments to the Malaysian Evidence Act 1950 under the guise of fighting cyber-crime and protecting citizens against threatening or slanderous online attacks. The changes, in short, make Internet users liable for content posted on their webpage, computer, or registered network. The potential applications of this law are particularly alarming for intermediaries, as they can now be held accountable for any offensive, libellous, or seditious third-party content posted anonymously on any web service they host, run, or are otherwise responsible for. According to the act, “any person whose name, photograph or pseudonym appears as producer, owner, host, administrator, editor or sub-editor, or who in any manner facilitates to publish or re-publish the publication is presumed to have published or re-published the contents of the publication unless the contrary is proved.” The presumption of guilt is automatic. The Malaysian government has effectively transferred the burden of proof from itself to the intermediary.

The amendments to the Evidence Act will force intermediaries to monitor the activity of third-party users 24 hours a day, seven days a week. The huge amount of traffic that many of these intermediaries – which include social networks, blog hosting platforms, and online media sites – receive in a single day will likely make the vetting process very difficult, if not impossible. We can therefore safely assume that many will take pre-emptive steps to control online content in order to protect themselves from being held responsible for a seditious comment, offensive video, or threatening post. Online news sources, for example, may remove the comment features on their websites. The threat to the future of online free speech in Malaysia is undeniable.

Vietnam is about to institute its own framework for holding intermediaries liable for third-party content. The Decree on the Management, Provision, Use of Internet Services and Information Content Online will compel intermediaries to co-operate with the government. They will be required to filter or remove any content deemed to oppose the Socialist Republic of Vietnam, undermine the unity of the people, or undermine the customs and traditions of the nation. The decree will also force these intermediaries to report any of this prohibited online activity to the authorities. Site administrators, foreign companies like Google and Facebook, and any other entity that provides online services in Vietnam will no longer be immune from the crackdown on online content that the government has imposed on bloggers and netizens for the past few years. They, too, will be forced to choose between aiding the government’s anti-free-speech agenda and risking harsh penalties for alleged Internet crimes.

Intermediaries in Deep

The benefits of intermediary liability to the governments discussed here are clear. First, it is much easier and cheaper to require online middlemen and women to police the Internet themselves. Governments who do so can avoid the cost of acquiring the technological and human resources that would be necessary to keep constant tabs on all content. Second, intermediary liability gives governments the means to assert some control over foreign actors operating within their borders. Third, and perhaps most importantly, they can control these actors, as well as domestic groups and individuals, while claiming that they themselves are not engaged in Internet censorship – they simply punish those intermediaries that fail to carry out their legal obligations.

The Internet has become an incredibly powerful democratic and inclusive tool to scrutinize the conduct of political and economic elites. It can create an alternate arena for citizens to exercise their right to freedom of expression, aid in the dissemination of information within and beyond national borders, and create domestic and global networks of like-minded people. Globally interoperable and decentralized, it makes it much more difficult for governments to preserve their hold on power by repressing free speech, access to information, and political organization. Many governments in Asia wish to control domestic debates online to keep their regimes stable, but also wish to appear tolerant, respectful of human rights, and democratic to the international community. This is likely a crucial motivation behind their decisions to outsource Internet censorship to intermediaries.

Many intermediaries are likely to drown under self-discipline or lèse-majesté-style legal regimes. The largest international companies, such as Google and Facebook, have vast resources for monitoring content. The time and money required to keep up with the demands of the Asian governments discussed may prove too great for smaller local hosts, as the cost of assessing content exceeds the profit from hosting it. Individuals like Chiranuch, who are singularly responsible for managing the content flows of high-traffic sites, will inevitably be overwhelmed, particularly if regulations tighten further.

Perhaps the most disturbing potential impact of cracking down on intermediaries is that they themselves may end up driving tighter regulations. In an effort to protect themselves from charges of neglect in content control, intermediaries may over-comply with the rules their governments set. When they pre-emptively censor content without thoroughly investigating whether it truly constitutes a violation, they cannot help but limit some lawful content. They may set new precedents, more stringent even than government-imposed rules, and thus deter debate of even the most uncontroversial societal issues. Tragically, fear of prosecution and punishment may make those who favour online freedom actively work to undermine it.

Standing Their Ground

Internet intermediaries facilitate communication and collaboration between citizens that can help them make informed choices about their political leadership. Holding intermediaries liable for all online content erodes the often already narrow space for public discussion of political issues. This erosion has already occurred in many countries in Asia and is spreading quickly. As the battle for control of the Internet continues to play out between governments and civil-society actors throughout Asia, there are some reasons for optimism. While the former will likely continue to favour greater control of online content and seek to implement more stifling controls, the latter, who have shown a remarkable ability to circumvent government controls, will continue to fight to keep the Internet free and open. Safeguarding online freedom is an integral part of wider civil-society efforts to increase government accountability and realize their human rights. Examples of this can already be witnessed across Asia as civil-society groups push back against government efforts to hold intermediaries liable for third-party content.

A number of civil-society groups in Asia have taken up the cause and are individually and collaboratively campaigning against intermediary liability. In July, a number of Southeast Asian civil-society actors attended the Asia-Pacific Regional Internet Governance Forum (APrIGF) in Tokyo to raise their concerns about the increase in Internet censorship in the region and call on governments to respect freedom of expression online.

Malaysia declared this past Aug. 14 “Internet Blackout Day.” In protest of the amendments to the Evidence Act, many websites went offline or hosted pop-up messages to raise awareness of the troubling implications. In response, Prime Minister Najib Rajak tweeted that he would ask Cabinet to discuss the amendments.

Finally, a bill proposing amendment to the country’s lèse-majesté laws was presented to the Thai parliament in June. The submission of the amendment was allowed under Article 163 of the constitution, which requires any proposed amendment to a law to be examined by parliament if at least 10,000 citizens sign a petition of support. The anger surrounding the death of Uncle SMS in a Thai prison catapulted the petition to almost 30,000 signatures, triple the amount required. Unfortunately, the bill was rejected by all but three of the 500 House of Representative members.

While it is difficult to predict how this contestation will unfold, it is clear that Internet intermediaries are increasingly finding themselves on the frontlines. Whether they defend the line demanded by the ruling elites or that of civil-society actors will be critical to the future of online free speech in Asia.

Photo courtesy of Reuters

Before you click away, we’d like to ask you for a favour … 


Open Canada is published by the Canadian International Council, but that’s only the beginning of what the CIC does. Through its research and live events hosted by its 18 branches across the country, the CIC is dedicated to engaging Canadians from all walks of life in an ongoing conversation about Canada’s place in the world.

By becoming a member, you’ll be joining a community of Canadians who seek to shape Canada’s role in the world, and you’ll help Open Canada continue to publish thoughtful and provocative reporting and analysis.

Join us