What Queries of Metadata Derived from Upstream Data Might Include

In this post, I explained that at virtually the exact moment the NSA shut down the PRTT dragnet in 2011, FISC permitted it to start querying metadata derived from upstream collection. After that happened, it started distinguishing between data that was “handled” according to minimization procedures and data that was “processed” before being intelligible.

In this post, I want to talk about what we can learn about metadata derived from FAA 702 from the opinion that authorized it and this document which based on the date, I assume pertains at least to upstream 702 derived metadata (from which the two kinds of MCTs most likely to include domestic communications would be excluded).

First, assuming that this querying document does include upstream, then it means that entirely domestic communications might be included in the querying. The opinion allows,

NSA to copy metadata from Internet transactions that are not subject tosegregation pursuant to Section 3(b) without first complying with the other rules for handlingnon-segregated transactions – i.e., without ruling out that the metadata pertained to a discretewholly domestic communication or to a discrete non-target communication to or from a U.S.person or a person inside the United States.

This means that after the data comes in to NSA and the two types of metadata most likely to include domestic MCTs are segregated, it can be made available to metadata analysis. The NSA prevented queries of segregated data via technical means.

NSA’s technical implementation will ensure that USP metadata queries of FAA 702 collection will only run against communications metadata derived from FAA 702 [redacted] and telephony collection.

The document stated that “NSA’s Technical Directorate (TD) continues to work to implement this requirement.” It’s not clear whether that language dates to December 16, 2011, when it was first written, or to August 19, 2013, when it was most recently revised.

Yet even assuming that technical protection occurred, there would still be Americans in the pool. According to John Bates’ estimate from the same year, there might be 46,000 domestic communications in there that ended up in the batch because the domestic communication that made mention of targeted selector transited internationally, which led them to get caught in filters supposedly targeted at international traffic.

The opinion mandates that, if after doing the analysis, the analyst realizes she has a completely domestic communication, she has to destroy it (though that requirement would get softer the next year). But a footnote also reveals that the means of determining if a selector was American was not failsafe.

NSA will rely on an algorithm and/or a business rule to identify queries of communications metadata derived from the FAA 702 [redacted] and telephony collection that start with a United States person identifier. Neither method will identify those queries that start with a United States person identifier with 100 percent accuracy.

Moreover, in an apparent bid to have this querying process interact relatively seamlessly with Special Procedures Communications Metadata Analysis (SPCMA — a way to query EO 12333 metadata incorporating US person identifiers), the standards were lackadaisical. As with SPCMA, an analyst had to come up with a foreign intelligence justification, but that’s just a “memory aid” in case the analyst gets questioned about it “long after the fact” in a fact check. Analysts don’t have to seek approval before they use a particular selector to query and they’re not required to attach any supporting documentation for their justification (this was in 2013, so requirements may be stronger in the wake of the PCLOB report). And SPCMA training is considered adequate to query metadata derived from 702.

In other words (again, assuming this pertains to upstream querying), there are several risks: that US person data will get thrown in the mix, that it won’t get identified by an algorithm as such, and so that that query result will lead to further spying on a US person without getting destroyed.

Still, as made clear, the alternative is SPCMA, which offers even fewer protections than 702 querying.

One more thought: the NSA report on the aftermath of Bates’ upstream decision (and the implementation of the 2012 certificates) revealed the PRISM providers incurred cost with the transition between certificates. It’s actually quite possible that the upstream metadata queries would come to constitute a critical part of the targeting process, effectively identifying what Goole or Yahoo content might be of interest at the metadata stage, only then to submit that to the provider for the content. If that’s true, it would be somewhat easy to end up targeting a US person for content collection via such upstream searches (though that presumably would be captured in the post-targeting process).

Why Susan Rice May Be a Shiny Object

A bunch of Republican propagandists are outraged that the press isn’t showing more interest in PizzaGate Mike Cernovich’s “scoop” that the woman in charge of ensuring our national security under President Obama, then National Security Advisor Susan Rice, sought to fully understand the national security intercepts she was being shown.

There are two bases for their poutrage, which might have merit — but coming from such hacks, may not.

The first is the suggestion, based off Devin Nunes’ claim (and refuted by Adam Schiff) that Rice unmasked things she shouldn’t have. Thus far, the (probably illegally) leaked details — such as that family members, perhaps like Jared Kushner (who met with an FSB officer turned head of a sanctioned Russian bank used as cover for other spying operations), Sean Hannity (who met with an already-targeted Julian Assange at a time he was suspected of coordinating with Russians), and Erik Prince (who has literally built armies for foreign powers) got spied on — do nothing but undermine Nunes’ claims. All the claimed outrageous unmaskings actually seem quite justifiable, given the accepted purpose for FISA intercepts.

The other suggestion — and thus far, it is a suggestion, probably because (as I’ll show) it’s thus far logically devoid of evidence — is that because Rice asked to have the names of people unmasked, she must be the person who leaked the contents of the intercepts of Sergey Kislyak discussing sanctions with Mike Flynn. (Somehow, the propagandists always throw Ben Rhodes’ name in, though it’s not clear on what basis.)

Let me start by saying this. Let’s assume those intercepts remained classified when they were leaked. That’s almost certain, but Obama certainly did have the authority to declassify them, just as either George Bush or Dick Cheney allegedly used that authority to declassify Valerie Plame’s ID (as some of these same propagandists applauded back in the day). But assuming the intercepts did remain classified, I agree that it is a problem that they were leaked by nine different sources to the WaPo.

But just because Rice asked to unmask the identities of various Trump (and right wing media) figures doesn’t mean she and Ben Rhodes are the nine sources for the WaPo.

That’s because the information on Flynn may have existed in a number of other places.

Obviously, Rice could not have been the first person to read the Flynn-Kislyak intercepts. That’s because some analyst(s) would have had to read them and put them into a finished report (most, but not all, of Nunes’ blathering comments about these reports suggest they were finished intelligence). Assuming those analysts were at NSA (which is not at all certain) someone would have had to have approved the unmasking of Flynn’s name before Rice saw it.

In addition, it is possible — likely even, at least by January 2017, when we know people were asking why Russia didn’t respond more strongly to Obama’s hacking sanctions — that there were two other sets of people who had access to the raw intelligence on Flynn’s conversations with Kislyak: the CIA and, especially, the FBI, which would have been involved in any FISA-related collection. Both CIA and FBI can get raw data on topics they’re working on. Likely, in this case, the multi-agency task force was getting raw collection related to their Russian investigation.

And as I’ve explained, as soon as FBI developed a suspicion that either Kislyak was at the center of discussions on sanctions or that Flynn was an unregistered agent of multiple foreign powers, the Special Agents doing that investigation would routinely pull up everything in their databases on those people by name, which would result in raw Title I and 702 FISA collection (post January 3, it probably began to include raw EO 12333 data as well).

So already you’re up to about 15 to 20 people who would have access to the raw intercepts, and that’s before they brief their bosses, Congress (though the Devin Nunes and Adam Schiff briefing, at least, was delayed a bit), and DOJ, all the way up to Sally Yates, who wanted to warn the White House. Jim Comey has suggested it is likely that the nine sources behind the WaPo story were among these people briefed secondarily on the intercepts. And it’s worth noting that David Ignatius, who first broke the story of Flynn’s chats with Kislyak but was not credited on the nine source story, has known source relationships in other parts of the government than the National Security Advisor, though he also has ties to Rice.

All of which is to say that the question of who leaked the contents of Mike Flynn’s conversations with Sergey Kislyak is a very different question from whether Susan Rice’s requests to unmask Trump associates’ names were proper or not. It is possible that Rice leaked the intercepts without declassifying them first. But it’s also possible that any of tens of other people did, most of whom would have a completely independent channel for that information.

And the big vulnerability is not — no matter what Eli Lake wants to pretend — the unmasking of individual names by the National Security Advisor. Rather, it’s that groups of investigators can access the same intelligence in raw form without a warrant tied to the American person in question.

Devin Nunes’ So-Called Bibi Netanyahu Precedent

Throughout his ongoing information operation to claim the Obama White House spied on the Trump transition team, Devin Nunes has pointed to what he claimed was a precedent: when, in December 2015, members of Congress suddenly copped on that their conversations with Bibi Netanyahu would get picked up incidentally. In his March 22 press conference, he explained,

We went through this about a year and a half ago as it related to members of Congress, if you may remember there was a report I think it was in the Wall Street Journal and but then we had to have we had a whole series of hearings and then we had to have changes made to how Congress is informed if members of Congress are picked up in surveillance and this looks it’s like very similar to that.

Eli Lake dutifully repeated it in the second of his three-post series pitching Nunes’ information operation.

A precedent to what may have happened with the Trump transition involved the monitoring of Israel’s prime minister and other senior Israeli officials. The Wall Street Journal reported at the end of 2015 that members of Congress and American Jewish groups were caught up in this surveillance and that the reports were sent to the White House. This occurred during a bitter political fight over the Iran nuclear deal. In essence the Obama White House was learning about the strategy of its domestic political opposition through legal wiretaps of a foreign head of state and his aides.

But Lake didn’t apparently think through what the implications of Nunes’ analogy — or the differences between the two cases.

Here’s the WSJ report and CBS and WaPo versions that aren’t paywalled. All make it very clear that Devin Nunes took the lead in worrying about his conversations with Bibi Netanyahu being sucked up (I don’t remember Republicans being as sympathetic when Jane Harman got sucked up in a conversation with AIPAC). They also describe that Obama’s WH, faced with the potential that their surveillance would be seen as spying on another branch of Congress, had the NSA take charge of the unmasking.

The administration believed that Israel had leaked information gleaned from spying on the negotiations to sympathetic lawmakers and Jewish American groups seeking to undermine the talks.

According to the Journal, when the White House learned that the NSA eavesdropping had collected communications with U.S. lawmakers, it feared being accused of spying on Congress and left it to the NSA to determine what information to share with the administration. The Journal said the NSA did not pass along the names of lawmakers or any of their personal attacks on White House officials.

That’s not to say they’d take the same approach here — indeed, Lake now claims, at  least, that Susan Rice requested some Trump officials’ names to be unmasked, distinguishing it from the Bibi case in that White House did not leave it up to NSA to decide what to unmask (though the underlying reporting makes the silly claim that Rice, Loretta Lynch, and John Brennan were among a very limited number of people who could request a name be unmasked).

The larger point is, even assuming the collection of conversations between your political opponents and a foreign government designed to undermine your executive branch authority was scandalous, it’d still fall under the very legitimate concern of separation of powers.

Yes, Trump’s aides are from a different party. But they are nevertheless part of the executive branch. And the entire basis of counterintelligence spying — the entire point of FISA — is to ensure that executive branch officials are not targeted by foreign countries to be spies, which is part of the reason Mike Flynn attracted attention (which is not to justify the leaking of that intercept). Add in the legitimate necessity to implement executive branch policy and this is a very different case than the Bibi case, even if you want to defend (as I do, to a point) Republican members of Congress collaborating with foreign governments to undermine Article II authorities.

Nunes’ imagined solution — from his March 22 White House press conference — is ever nuttier.

Q: You’ve said legal and incidental. That doesn’t sound like a proactive effort to spy.

Nunes: I would refer you to, we had a similar issue with members of Congress that were being picked up in incidental collection a little over a year ago, we had to spend a full year working with the DNI on the proper notification for members of Congress to be notified which comes through the Gang of Eight. I would refer you to that because it looks very similar to that, would be the best way I can describe it.

The ODNI current informs the Gang of Eight when members of Congress get spied on (which means claims that a lot of GOP candidates got spied on is likely hot air, but which also means that if Nunes were collected as a member of the transition team, he’d have been the first to learn of it). Which is an important protection for separation of powers, but which also enables corrupt members of Congress to not just learn they’re being surveilled but, potentially, to alert the foreign targets what channels we’re using.

Maybe Trump wants that standard applied to the executive branch, but if he adopts it, we’re going to have a leaking free for all. Not to mention, it would make it absolutely impossible for the government to protect against espionage related to elections.

Or perhaps Nunes is just saying something more simple. Perhaps Nunes is saying the “dozens” of intercepts where Trump officials had been unmasked (to the extent that’s true) disclosed Trump’s transition-period attempts to drum up a war with Iran at the behest of Israel. Perhaps the real stink here is that, in the very same days Mike Flynn was telling Russia sanctions would be loosened, Trump was publicly undermining US efforts to take a stand against Israeli illegal settlements.

Perhaps, ultimately, this is still about a belief that the Israelis should never be wiretapped.

Who Violated Their Designated Role: Ezra Cohen-Watnick or Susan Rice?

In the original version of the latest right wing claim — that Susan Rice requested that multiple incoming Trump figures’ names be unmasked in intercepts — Mike Cernovich describes the genesis of Devin Nunes’ concern this way:

The White House Counsel’s office identified Rice as the person responsible for the unmasking after examining Rice’s document log requests. The reports Rice requested to see are kept under tightly-controlled conditions. Each person must log her name before being granted access to them.

Upon learning of Rice’s actions, H. R. McMaster dispatched his close aide Derek Harvey to Capitol Hill to brief Chairman Nunes.

But as Eli Lake — fresh off having apologized for letting Devin Nunes use him — tells the story, close Mike Flynn associate Ezra Cohen-Watnick discovered it and brought the discovery to the White House Counsel’s office, whereupon he was told to “end his own research” on unmasking.

The pattern of Rice’s requests was discovered in a National Security Council review of the government’s policy on “unmasking” the identities of individuals in the U.S. who are not targets of electronic eavesdropping, but whose communications are collected incidentally. Normally those names are redacted from summaries of monitored conversations and appear in reports as something like “U.S. Person One.”

The National Security Council’s senior director for intelligence, Ezra Cohen-Watnick, was conducting the review, according to two U.S. officials who spoke with Bloomberg View on the condition of anonymity because they were not authorized to discuss it publicly. In February Cohen-Watnick discovered Rice’s multiple requests to unmask U.S. persons in intelligence reports that related to Trump transition activities. He brought this to the attention of the White House General Counsel’s office, who reviewed more of Rice’s requests and instructed him to end his own research into the unmasking policy.

This repeats a claim Lake had made in his earlier apology post, which he presented as one detail in the NYT version of this story that was not accurate.

Another U.S. official familiar with the affair told me that one of the sources named in the article, former Defense Intelligence officer Ezra Cohen-Watnick, did not play a role in getting information to Nunes. This official said Cohen-Watnick had come upon the reports while working on a review of recent Justice Department rules that made it easier for intelligence officials to share the identities of U.S. persons swept up in surveillance. He turned them over to White House lawyers.

But it adds the detail that Cohen-Watnick had been told to stand down. That would explain why Lake and others would want to claim that Cohen-Watnick wasn’t involved in dealing all this to Nunes: because he had already been told not to pursue it further. If the multiple accounts saying he was involved in the hand-off to Nunes, it appears he did.

The WaPo’s version of this included a detail not included by the right wingers: that Cohen-Watnick went to John Eisenberg, not Don McGahn, with his “discovery.” Eisenberg is significantly responsible, dating back to when he was at DOJ, for ensuring that ordinary Americans would be sucked up in surveillance under PRISM. For him to be concerned about the legal unmasking of Americans’ identities (to the extent that did exist — and the record is still unclear whether it did) is laughable.

The timing of Cohen-Watnick’s research — dating back to February — intersects in interesting ways with the timeline in this March 14 Politico story of H.R. McMaster’s attempt to sideline him, which was overruled by Steven Bannon.

On Friday [March 10], McMaster told the National Security Council’s senior director for intelligence programs, Ezra Cohen-Watnick, that he would be moved to another position in the organization.

The conversation followed weeks of pressure from career officials at the CIA who had expressed reservations about the 30-year-old intelligence operative and pushed for his ouster.

But Cohen-Watnick appealed McMaster’s decision to two influential allies with whom he had forged a relationship while working on Trump’s transition team — White House advisers Steve Bannon and Jared Kushner. They brought the matter to Trump on Sunday [March 12], and the president agreed that Cohen-Watnick should remain as the NSC’s intelligence director, according to two people with knowledge of the episode.

The House Intelligence Committee first asked NSA, CIA, and FBI for details on unmasking on March 15, the day after this story broke, at which point Nunes already knew of the White House effort. When Nunes first blew this up on March 22, he falsely claimed that that March 15 request had been submitted two weeks earlier.

It’s clear the right wing wants to shift this into Benghazi 2.0, attacking Susan Rice for activities that are, at least on the face of it, part of her job. But the only way the White House could be sure that she (or Ben Rhodes, who they’re also naming) were the ones to leak this would be to investigate not just those two, but also all the FBI (which would have access to this information without unmasking these names, which not a single one of these right wing scribes admit or even seem to understand). That is, the only way they could make credible, as opposed to regurgitated right wing propaganda accusations about leakers is to have spied even more inappropriately than they are accusing the Obama White House of doing.

Raw Versus Cooked: Could NSC Monitor FBI’s Investigation?

Multiple people,including Bart Gellman and Josh Marshall, are now arguing that the reason Ezra Cohen-Watnick and Michael Ellis found intercepts involving Trump’s people is that they were monitoring FBI’s investigation of the investigation.

I certainly think the Trump people would like to do that — and would be willing to stoop to that. I even believe that the response to the Russian hack last year had some counterintelligence problems, though probably not on the FBI side.

But there are some details that may limit how much the NSC can monitor the investigation.

First, Devin Nunes has always been very clear: the intercepts he was shown have nothing to do with Russia. That’s not, itself, determinative. After all, Cohen-Watnick and Ellis might have found a bunch of Russian intercepts, but only shared the non-Russian ones so Nunes could make a stink without being accused of endangering the investigation. Also, it’s possible that intercepts involving other countries — most notably Turkey, but there are other countries that might be even more interesting, including Ukraine or Syria — would impact any Russian investigation.

Also note that among the many things Nunes appears not to understand about surveillance is that there are two ways an American’s name can be visible outside the circle of analysts doing the initial review of them: their names can be put into finished intelligence reports that get circulated more broadly, with customers asking to have the name unmasked after the fact. Alternately, their names can be found off of subsequent searches of raw data. At the NSA and CIA, searches for US person content are somewhat controlled. At FBI they are not only not controlled, but they are routine even for criminal investigations. So if, say, General Flynn (or Paul Manafort) were under investigation for failing to register as a foreign agent, the FBI would routinely search their database of raw FISA material on his name. (These are the “back door searches” Ron Wyden has been screaming about for years, concerns which people like Devin Nunes have previously dismissed on national security grounds.) And we have every reason to believe that counterintelligence intercepts of Russians in the US are among the raw feeds that the FBI gets. So if Flynn had conversations with Russians (or Turks) in the US, we should assume that FBI saw them as a routine matter if Flynn became the subject of an investigation at all. We should also assume that the FBI did a search on every Sergey Kislyak intercept in their possession, so they will have read everything that got picked up, including all recorded calls with Trump aides.

On March 15, the House Intelligence Committee asked the NSA, CIA, and FBI for information on unmasking. I don’t believe that request asked about access to US person names on subsequent searches or raw material. Furthermore, at least as of last week, the FBI was not rushing to comply with that request. As I noted after the Jim Comey hearing before HPSCI, none of the Republicans concerned about these issues seemed to have any basic clue about FBI’s searches on raw data. If Nunes doesn’t know (and he appears not to), it’s unlikely Ellis knows, who was until this month Nunes’ aide.

But there’s one other thing that may prevent NSC from obtaining information about the investigation: FBI sometimes uses what are called “ad hoc databases” that include raw FISA data (and probably, post EO 12333 sharing rule changes, raw EO 12333 data) tied to particular investigations. It’s unclear what conditions might necessitate the use of an ad hoc database (see page 25ff for a discussion of them), but if security concerns would encourage their use, it would be likely to have one here, an investigation which Comey described as being so sensitive he delayed briefing the Gang of Four. Ad hoc databases are restricted to those working on investigations, and include specific records of those authorized to access the database. So if FBI were using an ad hoc database for this investigation, it would be even harder for the NSC to learn what they were looking at.

If the FBI’s investigation relies on raw intelligence — and it would be unfathomable that it does not, because it would probably receive the raw FISA data tied to such an investigation routinely, and EO 12333 sharing rules specifically envision the sharing of raw data associated with counterintelligence investigations — then the NSC’s access to finished intelligence reports would provide little insight into the investigation (Nunes was a bit unclear on whether that’s what he was looking at, but the entire premise of his complaints is that these were finished reports).

But while we’re worrying about whether and how Trump would monitor an investigation into his aides, remember that in 2002, Jay Bybee wrote a memo authorizing the sharing of grand jury information with the President and his close advisors including for counterintelligence investigations.

In addition, the Patriot Act recently amended 6(e) and Title III specifically to provide that matters involving foreign intelligence or counterintelligence or foreign intelligence information may be disclosed by any attorney for the government (and in the case of Title III, also by an investigative or law enforcement officer) to certain federal officials in order to assist those officials in carrying out their duties. Federal officials who are included within these provisions may include, for example, the President, attorneys within the White House Counsel’s Office, the President’s Chief of Staff, the National Security Advisor, and officials within the Central Intelligence Agency and the Department of Defense.

[snip]

Although the new provision in Rule 6(e) permitting disclosure also requires that any disclosures be reported to the district court responsible for supervising the grand jury, we conclude that disclosures made to the President fall outside the scope of the reporting requirement contained in that amendment, as do related subsequent disclosures made to other officials on the President’s behalf.

In other words, Trump could demand that he — or his National Security Advisor! — get information on any grand jury investigations, including those covering counterintelligence cases. And no judge would be given notice of that.

With Jeff Sessions’ recusal, that’s far less likely to happen than it might have been. But understand that the Executive Branch believes that the President can learn about the happenings in grand jury investigations of the sort that might target his aides.

Update: additional details have been added to this post after it was first posted.

The Lesson Trump Has (Thus Far) Not Taught Us: Civilian Casualties

I have a confession.

There’s something I like about the Trump Administration.

It’s the way that his unpopularity taints long-standing policies or practices or beliefs, making people aware of and opposed to them in a way they weren’t when the same policies or beliefs were widely held under George Bush or Barack Obama. Many, though not all, of these policies or beliefs were embraced unquestioningly by centrists or even avowed leftists.

I’ve been keeping a running list in my mind, which I’ll begin to lay out here (I guess I’ll update it as I remember more).

  • Expansive surveillance
  • The presumption of regularity, by which courts and the public assume the Executive Branch operates in good faith and from evidence
  • Denigration of immigrants
  • Denigration of Muslims
  • Denigration health insurance

As an example, Obama deported a huge number of people. But now that Trump has expanded that same practice, it has been made visible and delegitimized.

In short, Trump has made things that should always have been criticized are now being far more widely so.

But there’s one thing that Trump has escalated that has thus far — with the singular exception of the botched raid on Yemen — escaped widespread condemnation: the bombing of civilians. There was the Al Jineh mosque on March 16, a school sheltering families in Raqqa on March 21, and this strike last week in Mosul, not to mention continued Saudi attacks in Yemen that the US facilitates.

Again, I’m not saying such civilian strikes didn’t happen under Obama. And it’s not clear whether this spate of civilian bombings arises from a change in the rule of engagement put in place in December, the influence of James Mattis, or Trump’s announced review of rules of engagement. But civilians are dying.

And for the most part, unlike all the other horrible things happening under President Trump, they’re getting little notice and condemnation in the US.

Update: This NYT story on the Mosul strike says that the increased civilian casualties do reflect a change in rules of engagement put in place under Trump.

If Amazon Web Services Goes Down, Do the Cloud Services AWS Provides the Intelligence Community Too?

As you may have heard, Amazon has had a bad outage today, taking down many entities that rely on its cloud service.

Most of the coverage has focused on the private businesses that have been affected, from small businesses to larger ones (I suspect Office Max was broadly affected, because they were down today too), to media outlets.

I want to know if, when Amazon’s Northern Virginia cloud services go down, whether the cloud services Amazon provides to the Intelligence Community goes down too. The IC cloud is supposed to be completely separate from AWS’ commercial services. But if things are going haywire generally in Northern Virginia, those problems may extend to Amazon’s (understood to be NoVA located) IC servers.

I raise that, in part, because of a point I made in these two posts about the new EO 12333 sharing rules Obama implemented in January. The data sharing envisioned can happen in one of three places: on NSA’s own servers, on the recipient agency’s own servers, or on the cloud.

NSA may choose to make raw SIGINT available (i) through NSA’s systems; (ii) through a shared IC or other Government capability, such as a cloud-based environment; or (iii) by transferring some or all of the information to the recipient IC element’s information systems. Only information that can be afforded appropriate handling, storage, retention, and access protections by the recipient IC element will be made available.

Indeed, rolling out the IC cloud was a necessary technical precondition for this sharing process.

As I subsequently pointed out, one application for this expanded sharing was to make counterintelligence information — of the kind that would be central to the investigation into Russia’s hack of the DNC and/or other influence peddling with Trump allies — more widely available (for example, to CIA and FBI).

In the procedures, the conditions on page 7 and 8 under which an American can be spied on under EO 12333 are partially redacted. But the language on page 11 (and in some other parallel regulations) make it clear one purpose under which such surveillance would be acceptable, as in this passage.

Communications solely between U.S. persons inadvertently retrieved during the selection of foreign communications will be destroyed upon recognition, except:

When the communication contains significant foreign intelligence or counterintelligence, the head of the recipient IC element may waive the destruction requirement and subsequently notify the DIRNSA and NSA’s OGC;

Under these procedures generally, communications between an American and a foreigner can be read. But communications between Americans must be destroyed except if there is significant foreign intelligence or counterintelligence focus. This EO 12333 sharing will be used not just to spy on foreigners, but also to identify counterintelligence threats (which would presumably include leaks but especially would focus on Americans serving as spies for foreign governments) within the US.

Understand: On January 3, 2017, amid heated discussions of the Russian hack of the DNC and public reporting that at least four of Trump’s close associates may have had inappropriate conversations with Russia, conversations that may be inaccessible under FISA’s probable cause standard, Loretta Lynch signed an order permitting the bulk sharing of data to (in part) find counterintelligence threats in the US.

This makes at least five years of information collected on Russian targets available, with few limits, to both the CIA and FBI. So long as the CIA or FBI were to tell DIRNSA or NSA’s OGC they were doing so, they could even keep conversations between Americans identified “incidentally” in this data.

Certain state adversaries would have big incentives to destabilize AWS, just for shits, giggles, and the chaos it would cause. If they could get into Amazon private clients’ servers, there would be plenty of data to make such an attack worthwhile.

But if such an attack also affected the IC cloud, that might be a different thing entirely.

Robert Eatinger Brags that CIA Complies with Law Passed 2 Years Ago — But Will It Really Limit CIA?

Robert Eatinger — the former CIA lawyer deeply implicated in torture who referred the authors of the Senate Intelligence Committee report on torture to DOJ for criminal investigation — has a curious column in The Cipher Brief. Eatinger purports to rebut commentators who have described “Executive Order 12333 as a sort of mysterious, open-ended authorization for U.S. intelligence agencies to engage in secret, questionable activities.” But mostly he addresses the Agency’s new Attorney General Guidelines under EO 12333 approved by Loretta Lynch on January 17.

Eatinger doesn’t explain what led to the adoption of new procedures. He does at least admit that the CIA had been operating on procedures written in 1982, a year after EO 12333 mandated such procedures. He also admits that those procedures did not reflect, “advances in collection methods due to changes in technology and privacy interests unforeseen in 1982, which did not contemplate the ubiquitous use of mobile phones, computers, and other digital media devices or evolving views of privacy and thus did not seek to address ‘big data’ or ‘bulk’ collection.” But readers who didn’t know better might conclude from Eatinger’s piece that the CIA just decided out of the blue to start protecting Americans’ privacy.

The proximate change to the procedures was likely a desire to finally expand data sharing under Obama’s new EO 12333 sharing rules, a final step before accessing a firehose of data from the NSA (curiously, Eatinger doesn’t mention that these new procedures will probably enable the expanded intake of vast amounts of bulk data including US person information). It also (as I’ll explain) belatedly responds to a mandate from Congress.

But in reality, the change comes in response to over three years of nagging from the Privacy and Civil Liberties Oversight Board, which asked James Clapper and Eric Holder to make agencies update these procedures back in August 2013, pointing out how much technology had changed in the interim. Which is another way of saying that, for the entire time when Eatinger was a top CIA lawyer, CIA was perfectly happy to operate on 35-year old procedures not reflecting current technology.

Among the procedures limiting CIA’s (newly expanded) access to bulk data, Eatinger highlights the five year restriction on retention of information including US person data.

These sections also satisfy the requirements to create procedures that limit to five years the retention of any nonpublic telephone or electronic communication acquired without the consent of a person who is a party to the communication except in defined circumstances (Section 309).

[snip]

Section 6 creates two different types of handling requirements for unevaluated information; one for “routine” handling and one for “exceptional” handling.  Exceptional handling requirements apply to intelligence collections either of nonpublic communications that were acquired without the consent of a party to the communication, or that are anticipated to contain U.S. person identifying information that is significant in volume, proportion, or sensitivity.  The exceptional requirements include segregating the unevaluated information, limiting access to CIA employees who receive special training, creating an auditable record of activity, and importantly, requiring such information to be destroyed no later than five years after collection, permitting extensions in limited circumstances.

The five-year limit in Section 6 is but one example of how specifics in the new procedures attempt to find the right balance of intelligence and privacy interests.  Each procedure involves an effort to find the right tradeoffs to allow lawful intelligence collection and protect privacy and civil liberty rights and interests. The tradeoff was between the risk to a loss in intelligence capabilities by destroying information at five years against the risk to compromising privacy interests by keeping the information longer.

It’s not until nine paragraphs after Eatinger introduces this requirement, which he notes arises from “Section 309” in paragraph 8, that he explains where it comes from in paragraph 17, from Congress.

The five-year retention period in Section 6 was not set by the CIA, DNI, or Attorney General, however, it was set by Congress through Section 309.

Eatinger doesn’t describe when Congress passed that law, but I will. It was in the Intelligence Authorization for FY 2015. It became law on December 19, 2014.

Which is another way of saying that for over two years after Congress passed this law mandating the destruction of bulk data including US person data after five years, CIA hadn’t updated its EO 12333 procedures to reflect that requirement (this was after Eatinger left CIA, so we can’t blame him for the tardiness).

Now, Eatinger helpfully confirms something I’ve long believed but hadn’t confirmed: rather than sorting through and deleting the US person data in the collection, which would be all the law requires, the CIA instead destroys the entire data set at the five year interval, effectively extending the privacy protections passed to cover US persons to foreigners as well (you’re welcome, Europe). Eatinger does so in a passage laying out the trade-offs to deleting data after five years.

Deleting all unevaluated information specifically concerning U.S. persons has little to no intelligence downside because intelligence agencies will never want or have reason to search their intelligence holdings.  The five-year period to destroy all unevaluated information, however, will remove not only information concerning U.S. persons but also any information potentially concerning valid intelligence targets, such as international terrorists, from the intelligence agencies holdings.  In this latter case, however, intelligence agencies will want and may have a reason to search its holdings for information on these targets.  The deletion of that information could thus have an adverse intelligence impact, particularly on counterterrorism and counterproliferation intelligence reporting, as well as on the conduct of human intelligence operations, all of which are important activities of the CIA.

The CIA could be expected to search all of its holdings upon receiving intelligence identifying a previous unknown person as a suspected terrorist or proliferator.  Under the five-year retention period, when the CIA conducts the search, any unevaluated information on that person that may have been acquired during a bulk collection activity over five years ago will have been deleted; CIA’s search will not retrieve that information.  Thus, CIA might gain an incomplete or misleading understanding of the individual, his place in a terrorist network, and his contacts.  Or, CIA may send intelligence officers to conduct dangerous human intelligence operations to collect information it once had.  The loss of five-year old information could also adversely impact the spotting, assessing, recruiting, and running of human sources. [my emphasis]

This is how Eatinger introduces Congress’ role in requiring CIA to destroy data after five years: to blame them for limiting the CIA’s ability to sit on bulk data on Americans and foreigners for 25 years. To his credit, Eatinger does describe Congress as “the right body” to “impose” a “single retention period … on the entire intelligence community.” Given his direct attacks on Congressional oversight of the torture program, though, I wonder precisely in what spirit he intended this comment.

In any case, Eatinger also emphasizes that CIA doesn’t have to abide by this “single retention period …  imposed on the entire intelligence community.” After suggesting that some agencies might be able to abide by the Congressional mandate, he asserts unnamed other agencies may not be able to.

Some intelligence entities likely could accomplish their mission and destroy unevaluated information in less than five years.  Others may need to retain information longer than five years.

He then notes that Congress has given agencies an out.

Congress has provided that intelligence agency heads may retain information longer than five years if the head determines a longer retention “is necessary to protect the national security of the United States” and certifies in writing to the intelligence committees the reasons for that determination, the new retention period, the particular information to be retained; and the measures that will be taken to protect the privacy interests of U.S. persons and persons located inside the United States.

That out is laid out in CIA’s procedures at 6.2.2.2, but rather than stating the intelligence committees must get notice, the section says only that, “Upon such extension, the [CIA Director] shall complete any notifications required by statute, Executive Order, or other Presidential decree” which, given the way the Bush Administration ignored FISA based on Presidential decree, doesn’t inspire confidence that Congress would get the notice mandated under Section 309.

In any case, we have reason to believe the CIA is just one month into receiving an expanded firehose of data, including a great deal of data on Americans. And Eatinger sure seems to suggest the CIA may never give the data obtained via that firehose up.

Ninth Circuit Trims Executive’s Expansive Claims to Be Able Pixie Dust Executive Orders

As you’ve surely heard, the Ninth Circuit handed President Trump a huge loss last night, refusing to overturn the nationwide stay on his Muslim ban. The per curium opinion is particularly strong in asserting that courts do have the ability to review Presidential orders, even those that pertain to national security.

But there’s another part of the opinion I’m particularly interested in, because if it is not reversed, it creates a very important new limit on what the President can do with EOs.

One of the problems Trump created for himself was targeting Green Card holders — lawful permanent residents. That’s because LPRs have long term relations with the country and are accorded constitutional protections, both within and outside of the US. So long as LPRs remain affected by the EO, it will be legally problematic, at least as it pertains to them.

The Administration tried to undo that damage by having the White House Counsel, Don McGahn, write guidance on how to interpret the EO, basically stopping its application to LPRs. Within the hearing, the attorney representing the states noted that the Administration’s stance toward LPRs had changed about five times. But it was clear the judges were also unimpressed with changes the WHCO, as opposed to the President, made to an EO.

Here’s where they rule that a WHCO can’t just change an EO with policy guidance.

The Government has argued that, even if lawful permanent residents have due process rights, the States’ challenge to section 3(c) based on its application to lawful permanent residents is moot because several days after the Executive Order was issued, White House counsel Donald F. McGahn II issued “[a]uthoritative [g]uidance” stating that sections 3(c) and 3(e) of the Executive Order do not apply to lawful permanent residents. At this point, however, we cannot rely upon the Government’s contention that the Executive Order no longer applies to lawful permanent residents. The Government has offered no authority establishing that the White House counsel is empowered to issue an amended order superseding the Executive Order signed by the President and now challenged by the States, and that proposition seems unlikely.

Nor has the Government established that the White House counsel’s interpretation of the Executive Order is binding on all executive branch officials responsible for enforcing the Executive Order. The White House counsel is not the President, and he is not known to be in the chain of command for any of the Executive Departments. Moreover, in light of the Government’s shifting interpretations of the Executive Order, we cannot say that the current interpretation by White House counsel, even if authoritative and binding, will persist past the immediate stage of these proceedings. On this record, therefore, we cannot conclude that the Government has shown that it is “absolutely clear that the allegedly wrongful behavior could not reasonably be expected to recur.” Friends of the Earth, Inc., v. Laidlaw Envtl. Servs., Inc., 528 U.S. 167, 189 (2000) (emphasis added).

In short, they’re arguing that to make the EO legal with respect to LPRs, the President himself is going to have to change the EO, not McGahn.

As most longtime readers know, I’m obsessed by the way that John Yoo pixie dusted EO 12333 by basically saying the President doesn’t have to modify an EO he is blowing off, by blowing it off he is simply modifying it. In a 2001 opinion (and a 2002 letter to the FISC) he wrote,

[T]here is no constitutional requirement that a President issue a new executive order whenever he wishes to depart from the terms of previous executive order. In exercising his constitutional or delegated statutory powers, the President often must issue instructions to his subordinates in the executive branch, which takes the form of an executive order. An executive order does not commit the President himself to a certain course of action. Rather than “violate” an executive order, the President in authorizing a departure from an executive order has instead modified or waived it.

George Bush used that ruling to be able to disseminate Stellar Wind data even though his EO said you could not disseminate SIGINT.

While this ruling does not directly affect that interpretation, it does suggest that only a President can alter an EO (or, alternately, he must first confirm that someone else modifying it has been delegated the authority to do so). So while it doesn’t entirely shut down the possibility of further pixie dusting, it does make such things harder. It does give people reason to challenge any such changes to an EO.

As I noted the other day, I don’t think John Yoo was so much complaining about Trump’s abuses, as complaining that the way he implemented his abuses might do permanent damage to claims of expansive Executive authority. Let’s hope Trump has already done so by refusing to formally alter an EO his WHCO recognized was vulnerable to legal challenge.

Charles Cooper, Who Sanctioned Delayed Notification of Covert Ops to Congress, Reportedly to Be Solicitor General

Update: According to the Hill, Cooper has withdrawn from consideration.

Yesterday, SCOTUSblog reported that Charles Cooper is likely to be named Trump’s Solicitor General.

Cooper is a close friend of Sen. Jeff Sessions, the Alabama lawmaker whom Trump has nominated to serve as attorney general, and helped to prepare Sessions for his confirmation hearings. Like Sessions, Cooper could prove to be a controversial pick. In addition to his role defending California’s ban on same-sex marriage, Cooper also signed a brief on behalf of the federal government supporting Bob Jones University in its challenge to the constitutionality of an Internal Revenue Service policy denying tax exemptions to religious institutions that discriminate based on race. The university had argued that its ban on interracial dating was based on its sincere religious beliefs, but the Supreme Court upheld the IRS policy, by a vote of 8-1. And in 1986, as the head of the Office of Legal Counsel, Cooper signed an OLC opinion that argued that employers could reject job applicants with AIDS if they were concerned about contracting the disease.

I take this solace out of the fact that a soul mate to Jeff Sessions will be appointed to Solicitor General. In Cooper’s most public recent work, the Prop 8 case, Cooper got his ass handed to him by facts and law. If he were to argue as poorly for Trump’s policies, it might help the courts keep the President’s excesses under check.

That said, there are things I’m gravely concerned about, with regards to the Cooper pick. I’m trying to figure out whether he is the author of the OLC memo stating that a President can ignore his own Executive Orders, the original pixie dust opinion.

But he is definitely the author of this Iran-Contra related opinion, stating that the President can delay notification on covert ops to Congress.

Section 501(b) of the National Security Act of 1947 must be interpreted in the light of § 501 as a whole and in light of the President’s broad and independent constitutional authority to conduct foreign policy. The requirement that the President inform certain congressional committees “in a timely fashion” of a foreign intelligence operation as to which those committees were not given prior notice should be read to leave the President with discretion to postpone informing the committees until he determines that the success of the operation will not be jeopardized thereby. Because the recent contacts with elements of the Iranian government could reasonably have been thought to require the utmost secrecy, the President was justified in withholding § 501(b) notification during the ongoing effort to cultivate those individuals and seek their aid in promoting the interests of the United States.

Effectively, Cooper came in, after the fact, and said it was no big deal that Ronald Reagan didn’t tell Congress he was negotiating with a sworn enemy of the United States to get around funding restrictions they had put into place. The opinion has been used since to keep Congress in the dark about things it should know — including, ironically, some of Obama’s negotiations with Iran.

Hawks should be most concerned, however, that Trump would use this to free Mike Flynn to negotiate with Russia and Syria without notice to Congress. Me, I’d be more worried that it would be used (again) to hide a torture program from Congress.