A new report by The Intercept news site reveals the National Security Agency is secretly providing troves of data to nearly two dozen government agencies using a “Google-like” search engine. Documents from Edward Snowden reveal that for years the NSA has made data directly available to domestic law enforcement agencies like the Drug Enforcement Administration and the FBI. The search tool, known as ICReach, contains information on both foreigners and millions of U.S. citizens who have not been accused of wrongdoing. It’s designed to share more than 850 billion phone, email and web records—that’s more than twice the number of stars in the Milky Way.
Ryan Gallagher talking:
so our new story is exposing what is worded in the NSA’s own secret documents as a kind of Google-like search tool to sift through hundreds of billions of communications records from phone calls, emails, Internet chats, location data from cellphones, and virtually every kind of metadata you can think of and more. And not only that, this information has been made accessible to almost two dozen agencies in the United States. Most of them are intelligence community agencies, but among those it includes, you know, domestic law enforcement like the FBI and the Drug Enforcement Administration. So, the thing is, it is of a vast scope and, you know, is much larger than I think what a lot of people expected for the sort of degree of data sharing that’s been going on between different U.S. government agencies.
General Alexander was kind of the—well, we described him as the kind of mastermind, because he was the architect behind this thing. Fascinating details in these documents that go right back actually to the early ’90s, but after 9/11, the NSA basically concluded that it had to bolster metadata sharing across the U.S. government, because there was a feeling that, after 9/11, there was intelligence failures that had failed to prevent that attack, and also they were getting slammed for the bad intelligence that led up to the Iraq invasion. And so, Alexander’s solution was that how they could solve this was to build this gigantic new metadata search system and give analysts right across the government access to it so they can sift through people’s communications, obviously in a bid to identify certain threats and things like that. But obviously, when you open a system like that up with all these records to all the thousands of analysts, there’s quite a sort of concern there about the possibility for abuse of that.
there’s multiple things that are shocking about it, mostly just the vast scale of it and the scope of it and the kind of brazen way that it’s described in the documents. You know, there was no—it didn’t really seem like there was any intention to try and restrict or place limitations on it. It’s all about how much can you share. They want to share as much as possible. And again, this kind of feeds into what they described as their kind of collect-it-all mentality, where they just want more and more and more data. That’s what they think is the solution. And so, for me, yeah, I think that the scale was shocking and also the fact that domestic law enforcement are able to tap into this thing with very little oversight and few restrictions. So, yeah, I mean, the whole thing really is just quite striking.
we’ve had it confirmed by the NSA that the data that’s swept up and stored on this database en masse is collected using this Reagan-era presidential order, which is called 12333. And this thing is subject to no court oversight from the secret foreign intelligence court and minimal congressional scrutiny. Indeed, Dianne Feinstein, who’s the chief of the Senate Intelligence Committee, even her, who—she’s usually quite a bit of a defender of the NSA—she has in the past said that this executive order isn’t subject to sufficient oversight. And so that’s the authority that’s being used to put all these records on this system, that are then being funneled across the U.S. intelligence community. So, there are huge legal questions about that, the restrictions on it, how it can be used potentially in domestic criminal investigations secretly by federal agents, and stuff like that. These are all questions that people are asking now and that we hope are going to be, you know, sufficiently addressed by the government in the weeks to come.
In the early ’90s, the CIA and the DEA, according to the documents, started this program called CRISSCROSS. And basically what they were doing was, you know, gathering as much information as they could about phone calls in Latin America and, it seems, pretty much anywhere, and using that information to go after drug targets, people involved in drug trafficking. But very quickly, the scope of this thing expanded, and by the sort of mid to late ’90s the NSA was involved in it, the FBI was involved in it, the Defense Intelligence Agency was involved in it. And they were scaling it up again, and they built a new system which they called PROTON to put on all kinds of new records, so it wasn’t just your basic phone call stuff. It was location data and records from CIA reports, stuff about people’s visa applications when traveling overseas—you know, everything like that, billions of records. And so, that was the kind of precursor to the ICReach program, which again scaled the thing up massively. They talk about a 12-fold increase. It was something like, by the end of the—though PROTON, I believe, still exists today, but it had about 50 billion or so of these metadata records on it. ICReach was developed to contain 850 billion or more. So that’s like more than a 12-fold increase in capacity.
The Intercept that compares how many records are now available via ICReach. It reads, in part, “People on Earth: 7 billion; Google searches per month: 100 billion; Estimated stars in the Milky Way: 400 billion. … And estimated records available via ICReach: more than 850 billion.”
the message it gets across is it just puts these sort of crazy numbers into context and into perspective, because often you hear, you know, 10 billion metadata records. You know, people think, “Well, what does that actually mean?” So, I think it’s quite helpful just to have that 850 billion number squared against these other ones that you mentioned, because it puts it into context, and you start to realize, well, you know, that is a huge amount of information that’s on there.
parallel construction was—some excellent Reuters reporters revealed last year that some federal agents within the Drug Enforcement Administration were using data that had been obtained through covert surveillance and using it to initiate investigations against people inside the United States, American citizens. But they’re not disclosing in the—if there was a prosecution, say, not—basically covering up that evidence and inventing a false evidence trail, so that the way the surveillance data was obtained could never be challenged in the court, which to most, you know, ordinary people, that seems like a clear subversion of the basic principles of the justice system, which I think, and everyone else who’s looked at it agrees, seems to agree, as well. I believe the Justice Department is reviewing the whole thing. But one of the big issues with the story we have just put out is that it seems to be a huge part of the jigsaw puzzle in terms of showing where some of the data that is being used in this parallel construction technique by federal agents could be coming from. And indeed, the Reuters reporter who—one of the Reuters reporters who first revealed parallel construction contacted me to say he thinks this story that we’ve done is usually significant for that reason, because it is a huge piece of the puzzle that shows how NSA data is ending up in the hands of DEA agents in the United States.
U.S. Marine Corps base that directs employees not to read The Intercept. It reads, in part, “We have received information from our higher headquarters regarding a potential new leaker of classified information. Although no formal validation has occurred, we thought it prudent to warn all employees and subordinate commands. Please do not go to any website entitled ‘The Intercept’ for it may very well contain classified material. … Viewing potentially classified material (even material already wrongfully released in the public domain) from unclassified equipment will cause you long term security issues. This is considered a security violation.”
this is a continuation now of what’s happened to WikiLeaks in 2010 and then what happened to The Guardian last year, where the military has this completely absurd policy to just block any public news website that’s publishing stories based on classified information. And they’re, you know, issuing these kind of draconian warnings to their staff that if they dare to read these news reports, that they will have dire security consequences and all the rest of it. And, you know, we know for a fact that there’s people within the military who are disturbed by this. It’s based on an actual policy that the DOD has in place that says you can’t view classified information on an unclassified computer until the information is formally declassified. But, you know, that kind of policy, in the age of Manning, in the age of Snowden, just is totally archaic, and it doesn’t fit the modern world. They need to review it, because you can have a situation where an intelligence analyst in the government, with a top-secret security clearance, is in a position that they can’t read public news reports. Now, if that’s the case, how can that intelligence analyst, whose job is to make sense of the world from inside the government—how can they do that properly if they can’t even read news websites? So it just is a counterproductive policy that I think, personally, completely absurd.
— source democracynow.org
Ryan Gallagher, reporter for The Intercept, the news website recently banned by the military. His latest report is “The Surveillance Engine: How the NSA Built Its Own Secret Google.”