The Net Delusion - Evgeny Morozov [87]
One obvious use of face-recognition technology would be to allow Iranian authorities to quickly learn the identity of the people photographed during street protests in Tehran. For why should the Iranian government embark on expensive investigations if they can get their computers to match the photos taken during the protests—many of them by the very activists appearing on them—with more casual photos uploaded on social networking profiles by the same activists? That said, governments and law-enforcement agencies had been using face-recognition technologies for a while before they became a commercially viable business. What is most likely to happen in the case of Iran is that widely accessible face-recognition technologies will empower various solo agents, socially and politically conservative cyber-vigilantes who do not work for the government but would like to help its cause. Just as hordes of loyal Thais surf the Web in search of websites criticizing the monarchy or hordes of pro-government Chinese are on the lookout for highly sensitive blog posts, hordes of hard-line Iranians will be checking photos from the antigovernment protests against those in massive commercial photo banks, populated by photos and names harvested from social networking sites, that are sure to pop up, not always legally, once face-recognition technology goes fully mainstream. The cyber-vigilantes may then continue stalking the dissidents, launch DDoS attacks against their blogs, or simply report them to authorities.
Search engines capable of finding photos that contain a given face anywhere on the Internet are not far on the horizon either. For example, SAPIR, an ambitious project funded by the European Union, seeks to create an audiovisual search engine that would first automatically analyze a photo, video, or sound recording; then extract certain features to identify it; and finally use these unique identifiers to search for similar content on the Web. An antigovernment chant recorded from the streets of Tehran may soon be broken down into individual voices, which in turn can then be compared to a universe of all possible voices that exist on amateur videos posted on YouTube.
Or consider Recognizr, the cutting-edge smartphone application developed by two Swedish software firms that allows anyone to point their mobile phone at a stranger and immediately query the Internet about what is known about this person (or, to be more exact, about this person’s face). Its developers are the first to point to the tremendous privacy implications of their invention, promising that strict controls would eventually be built into the system. Nevertheless, it takes a leap of faith to believe that once the innovation genie is out of the bottle, no similar rogue applications would be available for purchase and download elsewhere.
How to Lose Face on Facebook
One gloomy day in 2009, the young Belarusian activist Pavel Lyashkovich learned the dangers of excessive social networking the hard way. A freshman at a public university in Minsk, he was unexpectedly called to the dean’s office, where he was met by two suspicious-looking men who told him they worked for the KGB, one public organization that the Belarusian authorities decided not to rename even after the fall of communism (they’re a brand-conscious bunch).
The KGB officers asked Pavel all sorts of detailed questions about his trips to Poland and Ukraine as well as his membership in various antigovernment movements.
Their extensive knowledge of the internal affairs of the Belarusian opposition—and particularly of Pavel’s own involvement in them, something he didn’t believe to be common knowledge—greatly surprised him. But then it all became clear, when the KGB duo loaded his page on vkontakte.ru, a popular Russian social