I happened to need to consult the PressFreedom transcript of today’s Bradley Manning trial. And came across this exchange, which goes to the heart of the debate on NSA’s dragnet of Americans.
In it, Manning’s lawyer, David Coombs, questioned Chief Warrant Officeer Joshua Ehresman, the ranking officer in the SCIF Manning worked in. Ehresman describes how analysts in Manning’s role were encouraged to consult whatever sources they could get their hands on.
Q And you had earlier said the term data mining, what is data mining?
A That’s pulling everything you can from every bit of intelligence assets you’ve got to help build your products.
Q Would you expect (INAUDIBLE) list of data mining?
A Yes, sir.
Q Why is that?
A Because you can’t go off one source of intelligence to predict something to happen. You have to have other stuff that indicate that it’s going to happen. You can’t just guess.
Q Where do analysts obtain their information that they’re data mining?
A Everywhere, sir. We got them on the SIPR, we got them from the T-Drive. We got them from wherever we could, open source, anything.
Q And correct me if I’m wrong, when I think of the term data mining, what you just described, is basically an analyst looking at everything and anything that they can, at any location just to kind of figure out would this perhaps be relevant to what I’m doing, is that correct, or would you provide a different definition for it?
A Yes, sir.
Q Yes, sir, that’s —
A That’s correct. You’re trying to find out yes or no this is going to happen, and, yes, this is how it’s happened and this is why it’s happening. So you have to confirm or deny your assessment.
Q Now, with regards to I guess when you’re doing this was the any guidance put out that if your you’re data mining you can do everything but go to this particular area on SIPRnet?
A No, sir.
Q So were there any restrictions on what you would data mine on SIPRnet?
A No, sir.
Q Was it common for a soldiers or analysts to data mine?
A Yes, sir.
Q Did analysts also use open source information?
A Yes, sir.
Q And what is open source information?
A That’s regular Internet, sir.
Q How would an analyst use an open source?
A We would get on and check out the web pages or you can check out local newspaper or it’s anything that doesn’t come through our secret or higher confidential webs.
Q And, again, in kind of a general description, how would open source information help your work products?
A Sometimes some of the media had information that we didn’t find out through our patrols or something. We could get patrol report and they would have outside information or a different point of view from what happens. So we would use that in our assessment, sir.
Q And were analysts encouraged to use open source information for their work products?
A Yes, sir.
Q Was there any sort of restriction placed out by the S2 section of you can go to every place besides these sites on open source?
A There was no restriction, sir.
Ehresman would go on to testify that Manning was, “our best analyst by far when it came to developing products.” Manning was, Ehresman testified, “our go-to guy for a lot of our shops, sir.”
This has little to do, directly, with the question of whether Manning will be found guilty or innocent of the charges against him.
But it does demonstrate how impossible the goal of protecting both data mining analysis and privacy is. If you want the (then) Specialist Mannings of the world to do their job well, you need to give them as much information, in relatively unfettered form, as you can.
There’s little way to achieve this goal and, at the same time, protect the data you’re piping out, aside from the honor system. And the honor system relies, in turn, on you — the United States — matching your promises and claims.