FBI requests for records under Patriot Act have increased 1,000% in just four years; Update: Data-mining goes deeper than thought?

posted at 7:51 pm on June 11, 2013 by Allahpundit

It’s not just the number of requests, it’s the scope of them. They’re not demanding records related to particular investigations anymore, they’re demanding huge troves of records on random Americans for data-mining purposes, the same thing Patriot Act co-author Jim Sensenbrenner complained about a few days ago but somehow didn’t foresee in 2001.

Like I said, you might want to re-poll Bush’s numbers.

“That they were using this (provision) to do mass collection of data is definitely the biggest surprise,” said Robert Chesney, a top national security lawyer at the University of Texas Law School. “Most people who followed this closely were not aware they were doing this. We’ve gone from producing records for a particular investigation to the production of all records for a massive pre-collection database. It’s incredibly sweeping.”…

[I]n the years since [2003], the FBI’s use of Section 215 quietly exploded, with virtually no public notice or debate. In 2009, as part of an annual report to Congress, the Justice Department reported there had been 21 applications for business records to the Foreign Intelligence Surveillance Court (FISC) under Section 215 – all of which were granted, though nine were modified by the court. (The reports do not explain how or why the orders were modified.)

In 2010, the number of requests jumped to 205 (all again granted, with 176 modified.) In the latest report filed on April 30, the department reported there had been 212 such requests in 2012 – all approved by the court, but 200 of them modified.

These sharp increase in the use of Section 215 has drawn little attention until now because the number of national security letters (NSLs) issued by the bureau has been so much greater — 15,229 in 2012. But FBI Director Mueller, in little-noticed written responses to Congress two years ago, explained that the bureau was encountering resistance from telecommunications companies in turning over “electronic communication transaction” records in response to national security letters.

Google, Facebook, and the gang started to resist NSLs on grounds that they were dubious legally, so Mueller shifted to Section 215 and got back to squeezing them for more and more and more data. Hence Google’s new move, announced just this afternoon: To take some of the heat off itself and push it back onto the feds where it belongs, the company’s lawyer is asking the DOJ to let it publish the number — and scope — of FBI requests for records it receives each year. Right now it’s forbidden to say because Section 215 imposes a gag order on the target of the record request. If Google can’t legally resist complying, at least it can soothe its critics by revealing just how much pressure it’s under from the Most Transparent Administration Ever.

Exit question: This WaPo story is nothing more than a textbook pro-Obama leak authorized by the White House to take some heat off of him and maybe show the virtues of aggressive cyberwarfare, right?

Update: What does this mean?

A leading Republican senator on Tuesday described controversial U.S. spy programs as looking far deeper into Americans’ phone records than the Obama administration has been willing to admit, fueling new privacy concerns as Congress sought to defend the surveillance systems.

Sen. Lindsey Graham, R-SC., says the U.S. intelligence surveillance of phone records allows analysts to monitor U.S. phone records for a pattern of calls, even if those numbers have no known connection to terrorism.

Graham says the National Security Agency then matches phone numbers against known terrorists. Graham helped draft the surveillance law that governs the surveillance program.

In other words, they’re looking for patterns among the general population that match the phone patterns of people they’ve already identified as terrorists? So they’re not limiting themselves to targeting specific terrorist-linked numbers anymore, but suspicious “patterns” too?

Parting thought:


Related Posts:

Breaking on Hot Air

Blowback

Note from Hot Air management: This section is for comments from Hot Air's community of registered readers. Please don't assume that Hot Air management agrees with or otherwise endorses any particular comment just because we let it stand. A reminder: Anyone who fails to comply with our terms of use may lose their posting privilege.

Trackbacks/Pings

Trackback URL

Comments

Comment pages: 1 2

Skip now if you don’t like long posts :)

Data mining almost certainly goes much, much deeper than most people think because most people have no idea what is involved and how it works.

Example 1:
Drug dealer DA uses phone P1 to call drug dealer DB who is using phone P2. DB then changes location and uses phone P3 to call dealer DC who is using phone P4. The change of phone and location is intended to hide the connection between DA and DC, but data mining algorithms will reveal that a call between P1 and P2 is often followed approximately two hours later by a call between P3 and P4 in a manner that suggests (but doesn’t prove) a causal, non-coincidental, relationship between DA and DC.

Example 2:
Licence plate cameras note all the vehicles passing along the roads. Simultaneously mobile phone base stations are triangulating the locations of all mobile phones (GPS not required). Clever algorithms will then be able to associate particular mobile phones with specific vehicles or, of course, to determine that a particular phone is carried by a pedestrian or, even, a wheelchair user. In the next stage of analysis it can be determined who is driving and who is a passenger picked-up along the way. This concept has long been used to monitor traffic flow and can identify traffic problems even if nobody calls 911.

Example 3:
The bank truck delivering and collecting cash changes its routes daily so that thieves cannot easily plan to intercept it. A group of thieves will use multiple vehicles (at different stages of the journey) to follow the truck, 200 yards behind, in order to find out where it is vulnerable. All the vehicles are observed at various locations by licence plate recognition cameras. A human observer would not see the relationship between the truck and the tail cars, but data mining software could. It is hugely unlikely that any other vehicle will have the same route (at a similar time and speed) as the cash truck through more than, say, half-a-dozen cameras and hence by the time the prospective thieves have tailed that truck through a few cameras it is quite likely that the police are already moving to intercept them. Even though the thieves change vehicles before the police arrive the vehicle that they change to will already have been flagged as suspicious because some weeks previously it was associated with the other vehicles because all vehicles had been associated with a particular mobile phone.

Example 4:
A lone lorry driver takes the same lonely, remote winding road between U and V each night. There are no video cameras or licence plate cameras on any part of the journey and the route is out of mobile phone coverage for much of its length. The driver travels at different times each night and has never discussed the journey with anybody. One night the vehicle slides off the road. An hour later a police patrol is dispatched to find the vehicle. The driver’s mobile phone was noted leaving location U but didn’t arrive in V according to the pattern established by previous trips. Data mining tools can ‘observe’ this anomaly and highlight it for investigation.

These are just trivially simple examples using not more than licence plate, phone location and phone calling records; even with these three data sets far more sophistication is possible, and far more obscure relationships (or patterns) can be found. When we add the data sets for facial recognition, transaction data, vibration data (from smart phones with movement sensors), email and Internet browsing then very, very obscure relationships can be found and ‘privacy’ as understood until, say, the 1970s doesn’t exist at all. Moreover, even more sophisticated sensors are being built and tested. Like any tool it can be used for good and evil; the danger comes because the tools are controlled by a small group of people and the ordinary person is merely a subject of the tool, entirely reliant on the benevolence and goodness of those who control it.

The main point is that data mining relies on having huge data sets in order to be able to identify relationships and hence it is almost certain that far more tracking and gathering is being done than people realise.

YiZhangZhe on June 12, 2013 at 5:50 AM

The more we learn about this, the more nauseated I get by Republicans calling Snowden a traitor.

DRayRaven on June 12, 2013 at 5:58 AM

The more we learn about this, the more nauseated I get by Republicans calling Snowden a traitor.

DRayRaven on June 12, 2013 at 5:58 AM

Yup. We deserved to know this was happening, even if we did’t know exactly how or when. They didn’t see it necessary to tell us, so they can fry.

alwaysfiredup on June 12, 2013 at 7:55 AM

All the knowledge in the world is of no use to fools…

ejfrench on June 12, 2013 at 8:30 AM

I heard the explanation for that. Approval only takes one of these FISA judges but if the initial judge has problems he/she/it has to write their concerns up and two other judges have to review and give an opinion. In short the system is stacked based on the idea that these judges are too damned lazy to do their job properly (overworked was the way it was framed). It is easier to attack Americans’ civil rights than it is to actually read the brief and render a judgement based on the Constitition.

Feel better about our safeguards under the Obama dictatorship?

Happy Nomad on June 11, 2013 at 9:30 PM

This is actually very insightful. Sounds to me like it should be the other way around for the FISA judges… that they should be required to write up an opinion IF THEY SUPPORT THE SURVEILLANCE to ensure that the warrants are properly read and understood. And if they are overloaded, I suggest a priority system… something like oh, a terrorism threat level assigned to each warrant.

dominigan on June 12, 2013 at 9:24 AM

The more we learn about this, the more nauseated I get by Republicans calling Snowden a traitor.

DRayRaven on June 12, 2013 at 5:58 AM

I don’t know too many libertarians or small-government conservatives that will ever cast another vote for these bastards.

MadisonConservative on June 12, 2013 at 9:51 AM

Comment pages: 1 2