NSA data collection ineffective against terrorism and dangerous for democracy, say mathematicians

'Blanket surveillance is highly unlikely to prevent a terrorist attack and is a dangerous misuse of resources'

Two American mathematicians have spoken of their concern that the mass data collection undertaken by the NSA with the aim of 'preventing terrorism' is both ineffective in achieving the stated goal and dangerous for democracy.

Writing in Notices of the AMS June/July 2014 (PDF), Keith Devlin, a mathematician at Stanford University who spent five years researching 'the area of extracting actionable information from vast amounts of data' funded by the US Department of Defense, claimed that mass data collection is an ineffective way of preventing terrorism, and that resources would be better deployed elsewhere.

"I concentrate on whether indiscriminate 'vacuuming up' of personal information that, according to the documents Edward Snowden has released, the NSA has routinely engaged in for several years, can effectively predict terrorist attacks," Devlin writes.

"I'll say up front that, based on everything I learned in those five years, blanket surveillance is highly unlikely to prevent a terrorist attack and is a dangerous misuse of resources that, if used in other ways, possibly could prevent attacks such as the 2013 Boston Marathon bombing. Anyone with a reasonable sense of large numbers could surmise a similar conclusion. When the goal is to identify a very small number of key signals in a large ocean of noise, indiscriminately increasing the size of the ocean is self-evidently not the way to go."

Writing in the same journal Andrew Odlyzko, professor of mathematics at the University of Minnesota in Minneapolis, claims that the concentration of power represented by mass collection of data by both the NSA and private organisations is dangerous for the entire political system.

"The antiterrorism mantra is driving public policy, and it is corroding the already weakened trust in democratic governance. When high-level officials feel free to give the 'least untruthful' answers or provide assurances of careful oversight and of intelligence successes that are then shown to be false, much is lost," Odlyzko says.

Odlyzko claims that this is just the beginning, and that the incipient Internet of Things will allow much greater levels of data collection and therefore potential for abuse. The NSA obtains data through private organisations, he points out, the business roadmaps of which are often predicated on collecting and processing ever larger amounts of data.

"Most of the data that the NSA has been using came from private organisations, and those are building their business cases on ever more intrusive data collection and exploitation," he writes.

"One report from the latest Consumer Electronics Show said that the 'unsettling message' of that event was that 'everything will be tracked'. What the NSA has been amassing is tiny compared to what will be available soon."

As well as the threat of officials misusing data amassed by the NSA for political purposes, for which he is careful to say there is no hard evidence, Odlyzko points to risks inherent in the way the data collected by private organisations is stored.

"Most of that [data] will be held in databases much more poorly protected than those of the NSA. Therefore we will have to worry about ... what might be done by even less trustworthy employees of the private organisations controlling that data and by all those who manage to break into those (inevitably insecure) databases," he writes.

In the interest of balance Notices of the AMS also sought mathematicians prepared to write in defence of the NSA, but said that for whatever reason "this proved difficult".