AT&T: this is how you spy on a social network
That the telecommunications giant AT&T is involved in murky relations with US state espionage in order to protect national security is already widely discussed: EFF has even filed a class action to reveal its dynamics. However, they are less well known the techniques set up by AT&T and the targets on which the monitoring focused.
Andrew Appel provides some of the details: he unearthed a six-year-old document, in which the telecommunications giant illustrates the operation of Hancock, a programming language used to translate some of the surveillance algorithms.
What is surprising is the intent of the mechanism: an excellent tool for undertaking marketing campaigns – reason for which it was developed – the algorithm would allow to map the networks of relationships between users, to clarify the relationships that exist between them, to organize them in community of interest , even when it means creating “guarded groups” such as scammers, terrorists or “threats to the nation”. All the nodes of the social network would be invested by the “guilt by association”, guilty of participating in an intertwining of intangible communication threads.
The New York Times had already highlighted the matter, browsing through the FBI documents. The detective agency, he explained, was not merely interested in monitoring individuals, but was requesting telephone companies for information on networks of people in contact with each other. Information that many phone companies had denied giving to the National Security Agency (NSA), which wanted it to target its wiretaps.
This unspeakable practice of monitoring, the Electronic Frontier Foundation warns, risks producing misleading representations of reality, trespassing into the private life of citizens, and involving individuals totally unrelated to the threats that would hang on the US. Suffice it to say that there are numerous studies that show how the distance between any two people in the world can be measured in six relational connections: what “distance” between a recognized terrorist transforms any user into a threat?
The answer exists: AT & T’s monitoring algorithm did not go beyond two degrees of separation from the goal. Information was collected only about those who spoke to who was in contact with the villain on duty. Nine GB of data could be accumulated and archived every night, explains Wired: between information extracted from short and long distance calls, logs of browsing sessions, IP addresses and movements from cell to cell of mobile users, there could not be but some interesting fact for the police.