A Stanford researcher has proven that the NSA can identify people based on phone records.
The National Security Agency likes to repeat that its employees only collect phone calls from millions of Americans, excluding their names from analysis (for example, they should use the so-called national security letter to the telephone company to get the FBI name). But Stanford researcher Jonathan Meyer found that he and his co-author can easily match metadata with names only through Google search.
“If several university researchers were able to cope with this so quickly, it’s hard to believe that the NSA will have any difficulty identifying the vast majority of Americans,” they write .
')
They conducted experiments on MetaPhone, a crowdsourced public database of voluntarily provided telephone records.
The science of identifying people on supposedly anonymous databases has become a kind of game for scientists. Last year, a team of researchers proved that they can identify people from a DNA database based on their open demographic information.
What is even more curious, other researchers were able to determine the sexual preferences of Facebook users based on information about the pages they “liked”.
“Even if you think you are keeping your personal information secret, we can still learn a lot about you,” writes Jennifer Golbeck, a scientist at the University of Maryland, who conducted a study similar to that in which Facebook users were identified.
From the point of view of statistics, it is not difficult to do this, each is unique in its own way. Despite the fact that the algorithm most likely cannot identify literally every person, the search capabilities are approaching the point when certain people will have no difficulty in finding information about the person of interest.
One can argue about whether government agencies should have access to our personal data, but do not pretend that they cannot find out everything they need based on the information they have.
“If several university researchers were able to cope with this so quickly, it’s hard to believe that the NSA will have any difficulty identifying the vast majority of Americans,” they write .
')
They conducted experiments on MetaPhone, a crowdsourced public database of voluntarily provided telephone records.
We randomly selected 5,000 numbers from the MetaPhone dataset and made inquiries on the Yelp, Google Places and Facebook directories. Without much effort and using only three of these (being public) sources, we compared 1,356 (27.1%) numbers. More specifically, there were 378 hits (7.6%) on Yelp, 684 (13.7%) on Google Places and 618 (12.3%) on Facebook.
What if any organization wants to do the same?
In order to conservatively evaluate how anonymous phone numbers are, we randomly selected 100 numbers from our data array, and then launched Google search for each one. Within an hour, we were able to relate to individuals or organizations 60 out of 100 numbers. When we connected our first three sources, this figure increased to 73.
The science of identifying people on supposedly anonymous databases has become a kind of game for scientists. Last year, a team of researchers proved that they can identify people from a DNA database based on their open demographic information.
What is even more curious, other researchers were able to determine the sexual preferences of Facebook users based on information about the pages they “liked”.
“Even if you think you are keeping your personal information secret, we can still learn a lot about you,” writes Jennifer Golbeck, a scientist at the University of Maryland, who conducted a study similar to that in which Facebook users were identified.
From the point of view of statistics, it is not difficult to do this, each is unique in its own way. Despite the fact that the algorithm most likely cannot identify literally every person, the search capabilities are approaching the point when certain people will have no difficulty in finding information about the person of interest.
One can argue about whether government agencies should have access to our personal data, but do not pretend that they cannot find out everything they need based on the information they have.
Source: https://habr.com/ru/post/207506/
All Articles