Academic

Cybersieves

By Derek E. Bambauer

Derek E. Bambauer

Your Internet is missing something.

In India, it?s pornography. In China, it?s political dissent; in France, white supremacist sites; in America, copyrighted material. Countries worldwide are using a combination of legal rules and technological tools to make targeted information disappear from their citizens? view of cyberspace. This strategy, known as Internet filtering, poses a hard question: when is it legitimate to censor the Net?

Internet filtering breaks down roughly into three eras?or, in geek lingo, ?versions.? In Version 1 of the Internet, censorship was an impossibility. As John Gilmore framed it, ?The Net interprets censorship as damage and routes around it.?¹ Terrestrial sovereigns, such as nation states, had neither power nor place in cyberspace. In Version 2 of the Internet, filtering was conceded to be possible, but only bad actors such as authoritarian states?think China, Iran, or Saudi Arabia?would engage in the practice. In Version 3, filtering becomes ubiquitous, common to democracies and dictatorships, First World and Third World states.

Internet filtering, Version 3, is upon us, and we need to adjust our thinking accordingly.

Judging whether censorship was legitimate was unnecessary in Version 1 and straightforward in Version 2. When Australia and Vietnam both engage in broad-based filtering, though, normative judgments become much more difficult. The initial instinct is to look at what content is blocked?surely it?s natural to filter child pornography but permit file sharing? (Japan and Italy would disagree, respectively.) The difficulty with judging filtering based on one country?s norms is that it doesn?t scale. Values may not apply across different cultures, making it difficult to achieve consensus on the normative justification for filtering.  Views about acceptable content vary widely?try searching for the white power group Stormfront on Google?s French engine (google.fr) and on its main U.S. site (google.com). (You?ll find that stormfront.org appears in the U.S. results, but not the French ones.) Legal scholars and Internet thinkers have struggled, unsuccessfully, to elucidate a test for measuring when it is legitimate to block access to content on the Net.

Cybersieves offers a new approach that looks not to what is blocked but at how the decision to filter is made. The Article suggests four touchstones for this analysis: openness, transparency, narrowness, and accountability. Openness probes whether a state admits to and explains its reasons for engaging in censorship; transparency measures a country?s disclosure of the types of material it blocks and how it evaluates suspect content; narrowness checks filtering?s precision in targeting information; and accountability assesses citizens? involvement in censorship policy. This methodology, which the Article calls ?the Framework,? focuses on process. It is designed to be compatible with different normative positions on filtering?whether you support or oppose blocking religious dissent, you?ll find it useful to know how forthright Saudi Arabia is in filtering that material and how involved its citizens are in that decision.

Narrowness is the criterion of the Framework that is mostly readily assessed empirically. Projects, such as the OpenNet Initiative and Herdict, document what a state?s filtering does and does not block. This can paint a telling picture: Vietnam claims to focus its filtering on pornography, yet it fails to block a single porn site. Political dissent and human rights content, however, is heavily censored. Cybersieves proposes bringing similar measurements to the other three criteria as well. Freedom House empirically measures press freedom in many countries, and Transparency International brings empirical analysis to bear on corruption. Yet analyses will differ both on how to measure a quality like transparency and on their relative rankings of states. This divergence is a benefit, not a detriment. Competition among different ranking methodologies will improve our ability to quantify openness, transparency, and accountability. The market for metrics will winnow candidates as some tools are adopted and others ignored.

Rating the legitimacy of countries? Internet censorship is useful in at least three practical ways. First, technology companies face difficult choices in deciding whether to operate in states that filter the Net. Google, for example, has struggled with the decision of whether to comply with China?s censorship demands for its search engine and related services, such as blogging and e-mail. The firm initially defended filtering results on its Chinese-language search site as preferable to forgoing the market altogether. But when confronted with unceasing demands for increased censorship and incontrovertible evidence of hacking by Chinese agents, the company resolved to end the filtering and, perhaps, exit the country altogether.² An empirical assessment of China?s practices, and how they have changed over time, could help Google both decide internally how to proceed and defend that choice publicly. Technology companies struggle with the conflict between corporate values and corporate profits. The Framework can help them decide how to balance those competing demands in a given country.

Second, governments struggle with whether, and how, to regulate via public law the choices that technology firms make. The United States bans selling crime control equipment to China, software to Iran, and just about everything to Cuba. Should America similarly prohibit selling censorship technology to Burma? Human rights activists press the government to do so, whereas technology companies generally prefer self-regulation. The debate is complicated by the dual-use nature of most information technology?as Cisco notes, the same code can block the Code Red virus and the Amnesty International website. Moreover, the increase in filtering by democratic countries further legitimizes the business of selling censorware. It is easier to defend selling firewalls that filter when Australia bids for them, too. The Framework suggests that governments should assess the pattern of sales by companies and the likely effect of additional transactions on the censorship capabilities of customers. If technology firms are selling to countries that engage in illegitimate information restrictions, a government can move to limit or ban such transactions. Moreover, governments can check for sales that move a state past the tipping point?where its newly purchased technological prowess allows the country to widen or strengthen censorship rendering its practices illegitimate?and move to interdict them.

Finally, as online censorship becomes commonplace, a panoply of groups are moving to evaluate it. Most agree at the extremes: Burma?s filtering is bad, while that of the United States is minimal (at least at the moment). But some analysts, such as Reporters Without Borders, fail to disclose the methodology behind their rankings, making it difficult to compare their output with that of OpenNet Initiative or Human Rights Watch. The Framework would improve the ability to aggregate and readily compare rankings in two ways. First, it formalizes a set of criteria for assessing legitimacy, and makes those criteria compatible with other, values-based means of examining censorship. Second, it encourages analysts to instantiate those principles through quantitative metrics and disclose how they arrived at a country?s score. In many ways, this emphasis recapitulates the core values of the Framework, such as openness and accountability. The goal of the Framework is not to generate unanimity but to help different groups translate their results for one another and to trace the path of those outcomes from the factors they emphasize in their ranking process.

The hardest, best test case for the Framework is China. Google?s decision to cease filtering google.cn, and its announcement that it may leave China altogether, have focused attention on the country?s sophisticated censorship system. Assessments of China?s filtering practices vary widely, even within the country. The state is generally loath to discuss censorship, has a welter of laws and regulations that limit what content can lawfully be disseminated online, and uses an arsenal of informal pressures and communications to keep sensitive material unavailable. Plus, China is the world?s largest Internet market and is a country that boasts a vibrant, creative technology sector. Companies?and the governments that oversee them?tend to resolve doubts in favor of transacting business there. Indeed, after Google?s announcement, Microsoft and HP reiterated their commitments to China, which is a key market for both companies.³ And China?s domestic technological capabilities may mean that outside pressures, whether from firms or from governments, are irrelevant. ⁴

Google?s withdrawal from the market will likely bolster Baidu, a domestic search engine with over 70 percent market share and with no compunctions about engaging in censorship. China is thus a hard problem?both analytically and practically. Its position in rankings of filtering countries is likely to vary greatly depending upon the methodology of the evaluator, and the heft of those rankings in guiding policy may be slight.

The cyber-libertarian perspective on Internet filtering?which holds that it is illicit under any circumstances?has lost. We connect to an Internet where what you see depends upon where you are on the network. Thus, scholars and policymakers badly need a new approach to guide them through complex issues related to online information controls. Cybersieves offers them a new tool: the Framework, a process-based approach to assessing legitimacy. By evaluating a country?s openness, transparency, narrowness, and accountability in its filtering, the Framework produces analytical rigor along with compatibility with methodologies derived from values-based judgments. Even as Internet filtering becomes increasingly ubiquitous, we should not accept all decisions to impose it as legitimate. Cybersieves seeks to show what bits are caught in the filters? mesh, and why.

Acknowledgments:

Copyright © 2010 Duke Law Journal.

Derek E. Bambauer is an Assistant Professor at Brooklyn Law School.

This Legal Workshop Editorial is based on the following article: Derek E. Bambauer, Cybersieves, 59 DUKE L.J. 377 (2010).

Full post as published by The Legal Workshop on May 03, 2010 (boomark / email).