The Net Delusion - Evgeny Morozov [151]
After all, anyone can text in deliberately erroneous reports to accuse their opponents of wrongdoing or, even worse, to sow panic in their ranks (remember the Nigerian SMS that said that all food was poisoned?). But to be credible, human rights reports and, to a somewhat lesser degree, reports from election observers need to aim for a 100 percent rate of accuracy. This is because of the peculiar nature of human rights reporting, especially in conditions where an authoritarian government may dispute the validity of results. One erroneous report—submitted by mistake or deliberately—is enough to derail the credibility of the entire database. And once human rights NGOs are caught producing data of dubious quality, the government gets a good excuse to shut them down. The New York Times praised the power of Ushahidi when it reported that “as data collects, crisis maps can reveal underlying patterns of reality: How many miles inland did the hurricane kill? Are the rapes broadly dispersed or concentrated near military barracks?” But this is quite misleading. At best, such maps could give us a general idea of the scale and the nature of abuses, but the value of this as a piece of human rights data is minimal. What’s worse, that small bit of crowdsourced data can easily make all other hard-earned data about human rights violations easy to dismiss.
Nor do we want to make certain information associated with human rights abuses publicly accessible on the Internet. In many countries, there is still a significant social stigma associated with rape. Providing even the tiniest bits of evidence—say, geographic information about where rapes have occurred—may reveal the victims, making their lives even more unbearable. There has always been a certain data protection mechanism built into human rights reporting, and given the ease with which information can be collected and disseminated, including by third parties that might be working to impede the work of human rights organizations, it is important to preserve those mechanisms, regardless of the impetus to promote Internet freedom. It’s not a given that projects that rely on crowdsourcing won’t get this balance right, but we need to resist Internet-centrism and opting for a “more people, more data” kind of approach without considering needs and capabilities.
Ironically, while most of the recent efforts of the digerati have focused on liberating the data from closed databases, the focus of their future efforts may soon shift to squeezing the open data back in or at least finding ways in which to limit the mobility of that data. This is a particularly important problem for various ethnic minorities who suddenly find themselves under threat, as digitized information has publicly identified them in ways they could not anticipate. In Russia a local branch of the Movement Against Illegal Immigration (DPNI), the country’s powerful anti-immigration network, created a series of online mash-ups in which they put census data about various ethnic minorities living in the Russian city of Volgograd onto an online map. This was not done to get a better understanding of urban life in Russia but to encourage DPNI’s supporters to organize pogroms on those minorities. DPNI is an interesting example of an unabashedly racist organization that has deftly adapted to the Internet era. Not only do they make mash-ups, but their main website runs an online store selling nationalist T-shirts; features the option to translate the site’s contents into German, English, and French; and even allows anyone registered on the site