Public Comments Regarding Foreign Censorship Practices Affecting U.S. Businesses

Second, social media censorship tools developed at the behest of countries with strong procedural protections against their misuse will eventually be used by countries with much weaker protections against misuse. Thus, what are initially human rights concessions tend to become trade concessions over time. Small concessions give way to larger ones.

In the USITC’s recent hearing on foreign censorship, witnesses claimed that “China is by far the worst offender.” In one sense this is true – China has created a uniquely restrictive system of censorship. However, because so much of its great firewall is predicated on outright market exclusion, it makes fewer demands of social media platforms than many other illiberal states. Industries allowed access to the Chinese market, such as the filmmaking industry, face very different pressures.

Some of this depends upon how costs are understood. Without access to the Chinese market, American social media firms forego tremendous potential income. However, the costs of compliance with foreign censorship regimes in accessible markets are real, ongoing, and often growing expenditures.

Even if American social media platforms were allowed access to the Chinese market, they would face extreme censorship demands. Concerns about the moral and political costs of compliance these demands already complicate platform attempts to re-enter the Chinese market. In 2018, Google shuttered “Project Dragonfly,” a proposal to offer search services in China, after intense political backlash when plans for the project became public.

Microsoft offers a censored version of Bing search in China, but has still faced intermittent bans. Its compliance imposes reputational costs at home, and perceived spillover censorship threatens trust in the search product among American users.¹ In China, Bing’s tightly managed, intermittent market access is the exception that proves the rule – most platforms are shut out entirely. Elsewhere, however, this approach is becoming commonplace.

Turkey has been particularly effective in using the intermittent denial of market access to ensure compliance with its censorship demands. In May of 2007, Tukey passed Law No. 5651, or the “Internet Act,” allowing Turkish courts to order the removal of domestically illegal speech from social media platforms, and ban access to the platforms themselves.² Less than a year later, an Istanbul court ruled that YouTube videos mocking Kemal Ataturk violated laws against “anti-Turkish speech,” and ordered internet service provider TurkTelecom to suspend access to YouTube across the country. YouTube removed the offending videos, but was repeatedly blocked by other courts, effectively suspending its access to Turkish markets for two years.

In 2015, Turkey imposed similar bans on Facebook, Twitter, and Periscope.³ One easily estimable cost of foreign censorship is platforms’ lost advertising revenue during these state-imposed blackouts. Demands to work contrary to American interests and values are harder to appraise. In 2018, Turkey compelled Facebook to remove accounts associated with the Kurdish YPG, an American partner in the fight against ISIS.⁴

In October 2020, Turkey introduced Law No. 7253, requiring large social media platforms to appoint a representative stationed in Istanbul.⁵ This representative, supervised by the Information and Communication Technologies Authority (BTK) and subject to Turkish law, is required to ensure that government censorship requests are obeyed. Maintaining what is effectively a hostage office in Istanbul is costly. Three American platforms, Twitter, Periscope, and Pinterest, initially refused to appoint a representative. They were prohibited from running advertisements in Turkey until they complied.

This approach has appeal to other states skeptical of the influence of free discourse. Nigeria recently announced its intention to impose a similar local representative mandate on social media platforms operating in the country.⁶ India has required social media firms to appoint local “compliance” and “grievance” officers. India recently retaliated against Twitter’s decision to label a BJP spokesman’s tweet “manipulated media” by conducting a police raid on the company’s offices.⁷ If met with continued success, the practice of using service blackouts to win censorship concessions will spread further.

Sometimes the threat of regulation is enough to jawbone compliance, particularly when firms already have a strong local presence. Over the past decade, American social media firms have agreed to public-private-partnerships with European authorities and developed cross-platform hash databases of prohibited content in response to European threats of regulation and denial of market access. Faced with continuing demands to moderate more stringently, platforms further expanded the use of algorithmic moderation and pledged to hire more moderators. As University of Virginia law professor Danielle Citron explains, “companies have not chosen this path for efficiency’s sake or to satisfy the concerns of advertisers and advocates. Instead, European regulators have extracted private speech commitments by threatening to pass new laws making platforms liable for extremist speech.”⁸ It is difficult to respond to these measures liberally. Attempting to regulate platform compliance with foreign censorship demands, in the style of the Anti-Israeli Boycott Act, would assert the same theory of digital sovereignty that justifies foreign censorship demands.

ITIF Associate Trade Policy Director Nigel Cory’s written testimony draws a distinction between trade and human rights concerns. This distinction may hold true for some foreign limits on free expression. However, reasonable, procedurally limited, and inexpensive concessions have paved the way for more onerous and costly censorship demands.

Certain discourse and content that is legal in the United States may be illegal in other countries, such as content related to hate speech. While this may raise valid human rights concerns, it’s less of a trade issue, as these cases tend to be narrowly focused and within a broader legal framework where U.S. firms have a transparent criteria and legal redress to manage country-specific differences.”⁹

Countries that formerly applied limited, predictable restrictions have increased their expectations of platforms while imposing law enforcement duties on them. Autocratic states have demanded the use of tools developed at the behest of countries with greater procedural speech protections.

Twitter first developed its country-withheld-content tool to comply with German prohibitions on fascist symbols. The tool allows governments to request the local removal of certain content – it is still accessible by users elsewhere, but not in the country which requested the removal. This fine-grained censorship tool might have been used in a limited, procedurally legitimate fashion by the German government, but other states demanded access too. By 2018, Turkey was using the tool to request the local restriction of more than 5,000 pieces of content a year.¹⁰ While Twitter and other social media platforms do not comply with every local removal request, sorting legitimate from capricious removal requests is expensive and time consuming. Even requests backed by local law may offend the values of platforms’ western users. What was introduced as a tool with implications for human rights but not trade quickly grew to affect both categories.

While Germany’s demands were initially quite limited, they have expanded over time. In 2017, Germany passed the Netzwerkdurchsetzungsgesetz or NetzDG law, which requires platforms to remove “obviously illegal” content within 24 hours of receiving a complaint. The statute requires platforms to make judgements about the legality of speech on a shot-clock. Platforms can be fined for failing to remove illegal content, incentivizing them to remove anything that might be illegal. In order to comply with the law, platforms have had to hire moderators with significant expertise in German language and jurisprudence. Although it is an American ally, and ostensibly a liberal nation, Germany has imposed illiberal and expensive speech policing obligations on American social media platforms. Focusing exclusively on the activities of hostile or illiberal nations omits the ways in which nations closer to America, and perhaps more easily influenced by American diplomacy, have used speech regulation to impost costs on American social media firms.

While Chinese censorship of its own internet receives the lions share of attention, countries better integrated into the global share internet have much more effectively imposed censorship obligations on American social media firms. These obligations are both more difficult for platforms to avoid or ignore, and more likely to affect the speech of American users. While Americans rarely correspond with mainland Chinese on the open internet, they frequently socialize and do business with Europeans. If American firms must govern European communications more strictly, this stringency must apply to American speech with Europeans too. Because of the internationality of internet communication, local laws are rarely met by local solutions. Only so much content can be locally withheld before effectively severing lines of communication. Instead, locally particular restrictions or taboos must often be applied across platforms’ networks writ large.

Focusing exclusively on Chinese barriers to market entry or ignoring ostensibly procedurally bound censorship requirements would exclude many of the costliest and most pernicious examples of foreign censorship as a barrier to American social media platforms.