The Latest Stinky 702 Opinion Bodes Poorly for the Next One

Last night, I Con the Record released last year’s 702 opinion, approved by current presiding FISA Judge James Boasberg. It’s stinky. It shows continued violations of querying procedures (which I’ll describe below), as well as on new troubling issue at NSA (which I hope to describe in a follow-up).

Worse still, the opinion, the timing, and recent Bill Barr actions suggest we’ll see an even stinkier opinion in maybe another year.

The opinion we’re getting on September 3, 2020, was released by FISC on December 6, 2019. Not only has it taken nine months to release this opinion, but ODNI sat on it in anticipation of and in the aftermath of the DOJ IG Report on Carter Page, which was publicly released December 9, 2019. That means that the delay in releasing this led to a disproportionate focus on events that happened three or four years ago, but not on events that have persisted under Billy Barr.

But the timing is important for several other reasons: the government has to be preparing its next reapproval package now (assuming the 2019 certificates are good until December 5, it would need to submit a new package by November 5). That’s significant for several reasons. First, as laid out by the timeline below, while the FBI waited for a FISCR review of an October 2018 Boasberg decision that its querying procedures didn’t comply with a new requirement passed by Congress, there were ongoing querying problems of the same type, including both the deliberate querying of 702 information to vet sources (and cops), but also at least one mass query that ended up finding seven leads out of 16,000 Americans. There was a significant delay in reporting some of these:

  • Querying violations found in June reported September 18, 2019
  • Querying violations found in July reported September 6, 2019
  • August querying violation involved 16,000 people reported November 25, 2019

In addition, there were several more reports on querying violations, one on September 17, and another on September 20.

That is, the reports on some of these were delayed until after FISCR ruled (on July 12), and for many of them, there was a delay until around the same time as the government submitted their new reauthorization packet on September 17, 2019 (which is the package that led to this December 6 opinion).

Then, after submitting the reauthorization package, starting on October 4, 2019, the FBI asked to be excused from two reporting requirements imposed in 2018.

In one case — requiring that FBI has retained 702 information in some archival systems — the FBI waited to comply with a change in reporting requirements made in October 2018 until it was prepping the 2019 certificates, and then asked for a weaker reporting requirement (and got it, prospectively).

It must be noted, however, that the government has unjustifiably disregarded the current reporting requirement. Instead of taking concrete steps to comply even partially with the Court’s directive (or timely seeking relief from it), it chose to wait while the FBI reportedly worked on guidance to instruct its personnel on how to handle unminimized Section 702 information on these archival systems. See Letter Regarding the FBI’s Steps to Implement an Aspect of the Court’s 2018 Section 702 Opinion and Order, Sept. 27, 2019, at 3. In fact, it has taken so long to prepare this guidance that, instead of using it to instruct personnel on the October 2018 reporting requirement, which the government reports was the original plan, the FBI now intends to address only the narrower reporting requirement incorporated into the FBI’s proposed minimization procedures. See Letter Regarding the FBI’s Steps taken by the FBI to implement an aspect of the Comt’s 2018 Section 702 Opinion and Order, Nov. 20, 2019, at 4.

It should be unnecessary to state that government officials are not free to decide for themselves whether or to what extent they should comply with Court orders. The government has not sought retrospective relief from the reporting requirement imposed by the Court on October 18, 2018. Although the AG and DNI have amended the prior Section 702 certifications to authorize the FBI to apply its proposed minimization procedures to information acquired under prior certifications, that authorization only becomes “effective on October 17, 2019, or on the date upon which [this Court] issues an order concerning [the] amendments pursuant to subsection 702(j)(3) of the Act, whichever is later.”[redacted] The Court’s approval of those amendments does not have any nunc pro tune effect, nor does it excuse the government from reporting instances of retention that it is already obligated to report. With respect to those instances of retention, the October 2018 reporting requirement remains in effect.

In another — far more important — case, the FBI asked for the reporting requirement (on when an Agent conducts a criminal search and finds 702 information) to be eliminated entirely, again, after the reauthorization package was completed. This reporting requirement was designed to test the FBI’s now provably false claim that agents would never find 702 information when conducting criminal searches. It goes to the heart of concerns about Fourth Amendment violations.

Boasberg relaxed, though did not eliminate, that reporting requirement.

The government has not reported such instances in timely fashion. Rather, they have been reported to the Court belatedly, usually after they were uncovered during oversight reviews. The government now seeks relief from this reporting requirement “because the requirements in Section 702(f)(2) are a sufficient mechanism for the Court to assess the risk that the results of a query designed to elicit evidence of crimes unrelated to foreign intelligence will be viewed or otherwise used in connection with an investigation that is unrelated to national security.” October 4, 2019, Request at 8. But it would be premature to regard the government’s implementation of Section 702(f)(2) as a sufficient source of information. As discussed above, the FBI has repeatedly accessed Section 702-acquired contents under circumstances requiring a FISC order under Section 702(£)(2), but has never applied for such an order.

Closer to the mark is the government’s contention that implementing both Section 702(f)(2) and the November 2015 reporting requirement could complicate training and systems design. See October 4, 2019, Request at 8-9. For example, Section 702(f)(2) looks to whether a query involves a U.S.-person query term, while the applicability of the November 2015 reporting requirement depends on whether U.S.-person information is retrieved. And Section 702(f)(2) is implicated only when contents are accessed, while the November 2015 reporting requirement · does not distinguish between contents and non-contents information.

The Court has decided to retain a reporting requirement separate from Section 702(f)(2) because the obligation to get a FISC order under that section is limited to queries conducted in the context of a predicated criminal investigation. The FBI conducts numerous queries of Section 702 information at earlier investigative stages. See October 18, 2018, Opinion at 75. Reports about queries at those stages remain relevant to the Court’s interest in receiving information about the extent to which U.S.-person privacy interests are implicated by queries that are not designed to find and extract foreign-intelligence information. The Court has concluded, however, that it is appropriate to modify the prior reporting requirement so that it will focus on the use of U.S.-person query terms, rather than on whether U.S.-person information is accessed as a result of a query, and will be triggered only when contents information is accessed. Such modifications should make it considerably simpler for the government to implement the requirement in combination with Section 702(f)(2), while still requiring reporting in situations where Fourth Amendment concerns are likely to be implicated. See October 18, 2018, Opinion at 93 (queries that use U.S.-person query terms and result in review of contents are “the subset of queries that are particularly likely to result in significant intrusion into U.S. persons’ privacy”).

Ultimately, Boasberg approved the certifications, effectively arguing that FBI just needed time to be trained on them.

The Court has previously assessed that requiring FBI personnel to document why a query involving a U.S.-person query term is reasonably likely to have returned foreign-intelligence information or evidence of crime before examining contents returned by the query should “help ensure that FBI personnel … have thought about the querying standard and articulated why they believe it has been met” and prompt them “to recall and apply the guidance and training they have received on the querying standard.” See id. at 93; see also In re DNI/AG Certifications at 41 (that requirement may “motivate FBI personnel to carefully consider … whether a query satisfies” the standard). The recently reported querying violations suggest that some FBI personnel still need such help. That is not altogether surprising. As discussed above, the FBI is really just sta11ing to implement that documentation requirement on a comprehensive basis. For that reason, the improper queries described above do not undermine the Court’s prior determination that, with that requirement, the FBI’s querying and minimization procedures meet statutory and Fourth Amendment requirements.

I suggested when the 2018 package was released last year, we’d start learning details of back door searches that had been implicit since 2007.

Nevertheless, 12 years after this system was first moved under FISA (notably, two key Trump players, White House Associate Counsel John Eisenberg and National Security Division AAG John Demers were involved in the original passage), we’re only now going to start getting real information about the impact on Americans, both in qualitative and quantitative terms. For the first time,

  • We will learn how many queries are done (the FISC opinion revealed that just one FBI system handles 3.1 million queries a year, though that covers both US and non US person queries)
  • We will learn that there are more hits on US persons than previously portrayed, which leads to those US persons to being investigated for national security or — worse — coerced to become national security informants
  • We will learn (even more than we already learned from the two reported queries that this pertained to vetting informants) the degree to which back door searches serve not to find people who are implicated in national security crimes, but instead, people who might be coerced to help the FBI find people who are involved in national security crimes
  • We will learn that the oversight has been inadequate
  • We will finally be able to measure disproportionate impact on Chinese-American, Arab, Iranian, South Asian, and Muslim communities
  • DOJ will be forced to give far more defendants 702 notice

The thing is, 11 months after the release of that opinion, we’re still not seeing results — in the form of declassified opinions — of what FBI’s querying really looks like, once they’re forced to actually track it. The entirely of this 2019 opinion still shows what Boasberg considers the pre-implementation period for this reporting regime.

And the FBI has been trying to weaken it for two years now!

There’s one more indication that we may see troubling details once we get the next 702 opinion in a year’s time, if we do get it.

Less than a week ago, Billy Barr issued a memo imposing a new national security auditing function on the FBI.

To enhance the FBI’s existing compliance efforts, the Director of the FBI is taking steps to build a more robust and exacting internal audit capability, including the creation of an office focused on auditing the FBI’s national security activities. To support that effort, I hereby authorize the Director of the FBI to commence the process of establishing, consistent with law and policy, the Office ofInternal Auditing (“OIA”). A separate office devoted to internal auditing and headed by a senior FBI official will ensure that ri gorous and robust auditing, which is an essential ingredient to an effective compliance regime, is canied out. The FBI shall work with the Justice Management Division to make the required reorganization notifications regarding this new office. Once established, OJA shall be led by an Assistant Director who shall have the same reporting chain as the Assistant Director for OIC and the Assistant Director for INSD. The Director of the FBI shall appoint the Assistant Directors for OIC, INSD, and OIA, with the approval of the Deputy Attorney General.

OIC, INSD, and OIA shall be responsible for carrying out the internal compliance functions of the FBI as assigned by the Director of the FBI, who shall ensure that each office does not duplicate responsibilities and is adequately staffed to perform its assigned functions. The Deputy Attorney General and the Assistant Attorney General for Administration shall coordinate with the Director to ensure that those functions are resourced and funded appropriately.

Even though Barr says the newly created OIA won’t overlap with the compliance and inspection functions at FBI, it’s not clear why not. Further, Barr’s memo does not explicitly say why FBI needed a new compliance review for national security cases rather than the existing legal reviews that had conducted such review.

Don’t get me wrong, done correctly, this could be a long-needed reform. It’s not clear it is being done correctly. It seems partly timed to the elections (with a report on implementation due just before then). And DOJ IG — which has, historically, found abundant problems with the functions enumerated here — will not review the efficacy of this until around May 2022.

The Department ofJustice Inspector General has agreed to assess the implementation of this memorandum (“initial assessment”) no sooner than 18 months after the establishment of OIA and to report such assessment, consistent with the Inspector General Act, to the Attorney General, Deputy Attorney General, Director of the FBI, and Assistant Attorney General for National Security. The Inspector General has furt her agreed to conduct a subsequent assessment no later than five years after the initial assessment, and periodically thereafter as determined by the Inspector General, and to report such assessments, consistent with the Inspector General Act, to the Attorney General, Deputy Attorney General, Director of the FBI, and Assistant Attorney General for National Security.

Within 60 days of the date of the Inspector General’s initial assessment, the Director of the FBI shall provide the Attorney General and Deputy Attorney General an assessment of the implementation of this memorandum, including an assessment of the effectiveness of the FBI’s compliance structure and whether compliance functions should be consolidated under an Executive Assistant Director.

Which is to say, this initiative, while it may be long overdue, feels like Barr trying to get ahead of something or somethings.

Billy Barr is an authoritarian. He doesn’t care about surveillance (indeed, he’s the grandfather of the dragnets that Edward Snowden revealed).

But something must have led him to take action to make it look like he cares.


March 24-27, 2017: The querying of 70K facilities “associated with” persons who had access to the FBI’s facilities and systems. FBI General Counsel (then run by Jim Baker, who had had these fights in the past) warned against the query, but FBI did it anyway, though did not access the communications. This was likely either a leak or a counterintelligence investigation and appears to have been discovered in a review of existing Insider Threat queries.

December 1, 2017: FBI conducted queries on 6,800 social security numbers.

December 7-11, 2017, the same entity at FBI also queried 1,600 queries on certain identifiers, though claimed they didn’t mean to access raw data.

February 5 and 23, 2018: FBI did approximately 30 queries of potential sources.

February 21, 2018: FBI did 45 queries on people being vetted as sources.

March 27, 2018: Initial 2018 package submitted.

April 5, 2018: Extension order.

Before April 13, 2018: an unspecified FBI unit queried FISA acquired metadata using 57,000 identifiers of people who work in some place.

October 17, 2018: Order finding FBI querying procedures do not comply with FISA.

February 21, 2019: NSA submits notice of Upstream violations.

February 26, 2019: Date after which NSA fixes Upstream violations.

June 2019: Oversight review finds violations of querying rules, including to vet a source, a candidate to be a local cop, and to find information about a planned visit by foreign officials.

June 26, 2019: Notice that CIA assistance to NCTC does not comply with rules.

July 2019: Oversight review finds violations of querying rules, including of college students in a “Collegiate Academy” and individuals who visited an FBI office. 

July 12, 2019: FISCR opinion finding that FBI querying procedures do not comply with FISA.

August 2019: Query of 16,000 persons identifies seven leads. 

August 12, 2019: FBI submits new querying procedures.

August 23, 2019: NSA complains about post-tasking for some collections.

September 4, 2019: Approval of amended FBI querying procedures.

September 6, 2019: Report of July 2019 query violations.

September 13, 2019: Notice regarding 702 query response showing 100 characters of text surrounding search term.

September 17, 2019: Application submitted, including proposed improvements on targeting procedures.

September 17, 2019: Notice of at least four querying violations involving taking steps to access 702 products without getting a warrant.

September 18, 2019: Report on June 2019 query violations.

September 20, 2019: Reports of other FBI querying violations, including to vet sources, to search on complainants, and to vet potential cops.

September 26, 2019: 45-day report on fulfilling FBI query rules.

October 1, 2019: Review period extended to December 16, 2019 (because of NSA and NCTC compliance issues, not FBI ones).

October 3, 2019: FISC orders further information.

October 4, 2019: FBI requests relief from requirement to report 702 access in response to criminal search.

October 10, 2019: Notice of overly attenuated NSA queries, including content searches using 23 US person identifiers.

October 11, 2019: Notice on FBI violations tied to not opting out of including FISA in searches.

November 4, November 13, 2019: Government provides additional information.

November 8, 2019: 45-day report on fulfilling FBI query rules.

November 14, 2019: Notice on violations tied to not opting out of including FISA in searches.

November 20, 2019: Government tells FISC that they never tried to comply with reporting requirement imposed in October 2018, are instead training their new proposed compliance method.

November 25, 2019: Notice regarding August 2019 mass query.

mid-December 2019: Date FBI promised to impose new record-keeping on FBI’s queries.

January 2020: Date NSA promised to have purged improperly acquired communications.

[Photo: National Security Agency, Ft. Meade, MD via Wikimedia]

Basaaly Moalin Wins His Appeal — But Gets Nothing

Basaaly Moalin is a Somali-American prosecuted for funding Al-Shabaab in 2010 who, years later, was used by FBI to justify the phone dragnet. After Edward Snowden revealed the Section 215 dragnet, the FBI pointed to his case, claiming they would not have found him were it not for the dragnet.

He just won an appeal of his case in the 9th Circuit, which found that the Section 215 dragnet may violate the Fourth Amendment. But it doesn’t do him any good, because the 9th Circuit panel determined that the government had been lying about how central the dragnet was in identifying him in the first place. The ruling is important, however, because it affirms that if the government is going to use evidence obtained from surveillance in court — or derived from surveillance — they need to notify the defendant.

The opinion argued that the Third Party doctrine probably doesn’t apply here, because current metadata collection obtains so much more than old-style pen registers.

There are strong reasons to doubt that Smith applies here.
Advances in technology since 1979 have enabled the
government to collect and analyze information about its
citizens on an unprecedented scale. Confronting these
changes, and recognizing that a “central aim” of the Fourth
Amendment was “to place obstacles in the way of a too
permeating police surveillance,” the Supreme Court recently
declined to “extend” the third-party doctrine to information
whose collection was enabled by new technology. Carpenter
v. United States, 138 S. Ct. 2206, 2214, 2217 (2018) (quoting
United States v. Di Re, 332 U.S. 581, 595 (1948)).

Carpenter did not apply the third-party doctrine to the
government’s acquisition of historical cell phone records
from the petitioner’s wireless carriers. The records revealed
the geographic areas in which the petitioner used his cell
phone over a period of time. Id. at 2220. Citing the “unique
nature of cell phone location information,” the Court
concluded in Carpenter that “the fact that the Government
obtained the information from a third party does not
overcome [the petitioner’s] claim to Fourth Amendment
protection,” because there is “a world of difference between
the limited types of personal information addressed in Smith
. . . and the exhaustive chronicle of location information
casually collected by wireless carriers today.” Id. at 2219–

There is a similar gulf between the facts of Smith and the
NSA’s long-term collection of telephony metadata from
Moalin and millions of other Americans.


The distinctions between Smith and this case are legion
and most probably constitutionally significant. To begin
with, the type of information recorded in Smith was
“limited” and of a less “revealing nature” than the telephony
metadata at issue here. Carpenter, 138 S. Ct. at 2219. The
pen register did not disclose the “identities” of the caller or
of the recipient of a call, “nor whether the call was even
completed.” Smith, 442 U.S. at 741 (quoting United States v.
New York Tel. Co., 434 U.S. 159, 167 (1977)). In contrast,
the metadata in this case included “comprehensive communications routing information, including but not limited to session identifying information (e.g., originating and terminating telephone number, International Mobile station Equipment Identity (IMEI) number, International Mobile Subscriber Identity (IMSI) number, etc.), trunk identifier, telephone calling card numbers, and time and duration of call.” In re Application II, 2013 WL 5741573, at *1 n.2. “IMSI and IMEI numbers are unique numbers associated with a particular telephone user or communications device.” Br. of Amici Curiae Brennan Center for Justice 11. “A ‘trunk identifier’ provides information about where a phone connected to the network, revealing data that can locate the parties within approximately a square kilometer.” Id. at 11–12.

Although the Smith Court perceived a significant distinction between the “contents” of a conversation and the phone number dialed, see 442 U.S. at 743, in recent years the distinction between content and metadata “has become increasingly untenable,” as Amici point out. Br. of Amici Curiae Brennan Center for Justice 6. The amount of metadata created and collected has increased exponentially, along with the government’s ability to analyze it. “Records that once would have revealed a few scattered tiles of information about a person now reveal an entire mosaic—a vibrant and constantly updating picture of the person’s life.” Klayman v. Obama, 957 F. Supp. 2d 1, 36 (D.D.C. 2013), vacated and remanded, 800 F.3d 559 (D.C. Cir. 2015). According to the NSA’s former general counsel Stewart Baker, “[m]etadata absolutely tells you everything about somebody’s life. . . . If you have enough metadata you don’t really need content . . . .” Laura K. Donohue, The Future of Foreign Intelligence 39 (2016). The information collected here was thus substantially more revealing than the telephone numbers recorded in Smith.

Importantly, it pointed to how much more revealing Moalin’s metadata was collected in conjunction with that of millions of other people (a point I made shortly after the District Court rejected Moalin’s original challenge).

Also problematic is the extremely large number of people from whom the NSA collected telephony metadata, enabling the data to be aggregated and analyzed in bulk. The government asserts that “the fact that the NSA program also involved call records relating to other people . . . is irrelevant because Fourth Amendment rights . . . cannot be raised vicariously.” Br. of United States 58. The government quotes the FISA Court, which reasoned similarly that “where one individual does not have a Fourth Amendment interest, grouping together a large number of similarly-situated individuals cannot result in a Fourth Amendment interest springing into existence ex nihilo.” In re Application II, 2013 WL 5741573, at *2. But these observations fail to recognize that the collection of millions of other people’s telephony metadata, and the ability to aggregate and analyze it, makes the collection of Moalin’s own metadata considerably more revealing.

After suggesting that Carpenter would apply to this dragnet, the panel then concluded that it doesn’t matter, because the dragnet wasn’t all that central to obtaining a warrant against Moalin.

Having carefully reviewed the classified FISA applications and all related classified information, we are convinced that under established Fourth Amendment standards, the metadata collection, even if unconstitutional, did not taint the evidence introduced by the government at trial. See Wong Sun v. United States, 371 U.S. 471, 488 (1963). To the extent the public statements of government officials created a contrary impression, that impression is inconsistent with the contents of the classified record

This will be a working thread.

Billy Barr Signs a Memo That Wouldn’t Have Helped Carter Page

For eight months, FBI and DOJ have been diligently making changes to the way they do FISA applications, with regular reports into the FISA Court. Whether or not those changes are adequate to fix the problems that beset the Carter Page application, they represent significant effort.

Curiously, a memo Billy Barr just released purporting to enhance compliance in FISA applications appears unaware of the filings at FISC, and instead cites only changes implemented in Christopher Wray’s response to the December 9, 2019 DOJ IG Report (see PDF 466 for his letter).

Therefore, in order to address concerns identified in the report by the Inspector General of the Department of Justice entitled, “Review of Four FISA Applications and Other Aspects of the FBI ‘s Crossfire Hurricane Investigation” (December 2019), and to build on the important reforms described by the Director of the Federal Bureau of Investigation (“FBI”) in his December 6, 2019, response to the Inspector General’s report, I hereby direct that the following additional steps be taken:

Arguably (as I’ll show), at least one of the provisions in the memo is weaker than a change FISC mandated itself.

And while the memo claims to want to protect the rights of people like Carter Page, Barr’s memo would in no way apply to Page. That’s because the special protections tied to political campaigns only apply to those currently associated with campaigns.

With respect to applications for authorization to conduct electronic surveillance or physical searches pursuant to FISA targeting (i) a federal elected official or staff members of the elected official, or (ii) an individual who is a declared candidate for federal elected office or staff members or advisors of such candidate’s campaign (including any person who has been publicly announced by a campaign as a staff member or member of an official campaign advisory committee or group, or any person who is an informal advisor to the campaign),

By the time FBI applied for a FISA application targeting Page, several prominent members of the campaign had dissociated the campaign from him — for his controversial ties to Russia! — in no uncertain terms; those disavowals were included in the FISA application. Yes, Page had been announced as an informal advisor, but then the campaign made very clear he was no longer an informal advisor (and even claimed he never had been).

To be sure, some of the changes proposed — both those limited to those connected with a campaign and the more general ones — are improvements. For example:

  • ¶3(b) requires non-delegable sign-off by the Director of the FBI and the Attorney General) of any application targeting someone associated with a campaign; while requiring non-delegable sign-off may introduce some problems, this is the kind of certification recommended by the DOJ IG Report (though arguably is already incorporated in the December 6, 2019 letter Barr cited).
  • ¶3(d) and ¶3(e) institutes a shorter renewal deadline for these political FISAs, 60 days instead of 90, and requires monthly reports to FISC describing the results and affirming the continued need for such surveillance. These are arbitrary but perhaps useful improvements, not least because by increasing the paperwork required to surveil a political target, they make it more likely that such surveillance will actually be worth it (as the third and fourth applications targeting Page were not).
  • ¶3(f) requires that any political application describe whether less intrusive investigative procedures have been considered — something already required in all FISA applications — and an explanation why those procedures weren’t used. Such a requirement would have been useful in Page’s case (as I noted last year), because it would have emphasized the efforts FBI was making not to take public actions, but in practice this response would almost always point to DOJ guidelines on avoiding taking public actions that might affect an election and might actually encourage the increased reliance on informants, something Trump’s people claim equates to FISA surveillance. A requirement like this might be useful if it took place in the scope of a debate about what techniques were intrusive or not, but there’s zero evidence such a debate has happened.

The memo has two parts on defensive briefings, probably designed to placate Republicans, but which likely don’t do much in practice:

  • For political targets, ¶3(a) requires the FBI Director to consider a defensive briefing before targeting someone, and if no briefing is given, then the Director must document it in writing. FBI did consider defensive briefings for Trump’s people, but for various reasons decided not to do it, but in the case of Carter Page, he had long been wittingly sharing non-public information with known Russian intelligence officers and when FBI tried to explain why such dalliances were problematic in March 2017, he simply disagreed. A defensive briefing for Page would have been as useless as President Obama’s warnings to Trump that Mike Flynn was a problem.
  • For all counterintelligence concerns pertaining to election interference, ¶4 requires the FBI Director to “promulgate procedures, in consultation with the Deputy Attorney General, concerning defensive briefings.” Not only is this requirement utterly silent about what such procedures should do, not only did Wray commit to a similar recommendation in his December 2019 letter, but defensive briefings are precisely what Acting Director of National Intelligence John Ratcliffe is currently politicizing.

As for key review processes mandated by the memo, some are just redundant at best or stupid at worst. For example:

  • ¶1 requires FBI personnel to review the accuracy sub-file before submitting a FISA application. That process is already in place. It’s called the Woods Procedure and it’s the procedure that failed to find errors in the Page application.
  • ¶2 requires someone — it doesn’t say whether FBI or NSD bears responsibility — to report any misstatement or omission to FISC. That’s already required. Plus, this requirement twice gives NSD the authority to determine whether something amounts to a reportable incident. The ongoing DOJ IG investigation into all the errors in FISA applications suggest NSD has deemed some omissions and errors not to be worthwhile of reporting (indeed, there were multiple instances in the Page applications where NSD did not include information they knew of, in at least one case information that FBI did not have). In short, this paragraph seems more focused on ensuring NSD — and not an outside entity, like DOJ IG or the FISC — retains the ability to determine what is and is not a reportable error.
  • ¶3(c) requires an FBI Assistant Special Agent in Charge who is not involved in an investigation to review the FISA application of any defined political targets. The DOJ IG Report found that even NSD lawyers involved in an investigation don’t have enough insight into a case to identify omissions. While an ASAC might have access to case files that NSD lawyers do not, there’s zero reason to believe someone with even less insight into an investigation would better be able to spot omissions than an NSD lawyer with an ongoing role in the application. So this review is likely useless busywork.
  • ¶3(g) requires the Assistant Attorney General to review the case file of a political target within 60 days of its initial grant to make sure everything is kosher, including that the investigation was properly predicated. In conjunction with the shorter renewal timeframe of such applications (which would require DAG sign-off in any case), all this amounts to is a heightened review on first renewal (the memo does not say this is not delegable, so such a review will and probably should not be done by the AAG). But in Page’s case, it would have done nothing (indeed, at the time this would have been done for Page, he was in Russia meeting high level officials, falsely claiming to represent Trump’s interests).

In short, while some of these changes are salutary, a number are just show, and some are worthless busy work.

But my real concern about them — particularly given how Barr only invokes the first Christopher Wray letter to DOJ IG — is how they interact with other details of the FISA reform events that have transpired since last December.

For example, in the last month, the FBI and DOJ engaged in a big dog-and-pony show to claim that none of the errors DOJ IG had identified in 29 FISA applications they reviewed affected probable cause and just two were material. Effectively, that big press push amounted to having NSD pre-empt DOJ IG’s findings in an ongoing investigation, and the public details of NSD’s own review raise abundant reason to doubt the rigor of it. So Barr’s emphasis (in ¶2) on NSD’s role in deciding what is an error seems to be a reassertion of the status quo ante in the midst of an ongoing investigation that is still assessing whether NSD’s reviews are adequate. That makes this feel like another attempt to pre-empt an ongoing investigation.

Even more troubling, Barr’s memo seems unaware of — and in key respects, conflicts with — an order presiding FISA Judge James Boasberg issued in March. As I noted at the time, that order recognized something that was apparent from the DOJ IG Report but which the IG either missed, ignored, or was bureaucratically unable to address: it wasn’t just FBI that dropped the ball on the Page FISA application, NSD did so too.

According to the OIG Report, the DOJ attorney responsible for preparing the Page applications was aware that Page claimed to have had some type of reporting relationship with another government agency. See OIG Rpt. at 157. The DOJ attorney did not, however, follow up to confirm the nature of that relationship after the FBI case agent declared it “outside scope.” Id. at 157, 159. The DOJ attorney also received documents that contained materially adverse information, which DOJ advises should have been included in the application. Id. at 169-170. Greater diligence by the DOJ attorney in reviewing and probing the information provided by the FBI would likely have avoided those material omissions.

Because of that, Boasberg required that DOJ attorneys, too, sign off on all FISA applications, and suggested they get more involved earlier in the process.

As a result, reminders of DOJ’s obligation to meet the heightened duty of candor to the FISC appear warranted. The Court is therefore directing that any attorney submitting a FISA application make the following representation: “To the best of my knowledge, this application fairly reflects all information that might reasonably call into question the accuracy of the information or the reasonableness of any FBI assessments in the application, or otherwise raise doubts about the requested probable cause findings.”

DOJ should also consider whether its attorneys need more formalized guidance – e.g. , their own due-diligence checklists. Consideration should also be given to the potential benefits of DOJ attorney visits to field offices to meet with case agents and review investigative files themselves, at least in select cases – e.g. , initial applications for U.S.-person targets. Increased interaction between DOJ attorneys and FBI case agents during the preparatory process should not only improve accuracy in individual cases but also likely foster a common understanding of how to satisfy the government’s heightened duty of candor to the FISC.

There’s no mention of Boasberg’s order and suggestions in Barr’s memo, and it’s unclear whether that’s because he has no idea what has transpired with the FISC, whether he thinks he can ignore Boasberg’s order, or whether his memo is just for show. In any case, it’s notable that Barr’s memo doesn’t incorporate the key insight Boasberg made, that FISA requires increased diligence from NSD, too.

Similarly, because Boasberg deemed the role of FBI’s lawyers to be “perfunctory,” he asked for more details about their role.

But the role described in the revised Woods Form appears largely 10 perfunctory. To assess whether additional modifications to the Woods Form or related procedures may be warranted, the Court is directing the FBI to describe the current responsibilities FBI OGC lawyers have throughout the FISA process.

Here, Barr has added one more FBI person (an ASAC uninvolved in the case) to the process, whose review can only be perfunctory, rather than ensuring that those with more visibility on the process have a substantive role. Barr also doesn’t incorporate into his memo a change that came from Amicus David Kris after the Wray letter cited in Barr’s memo that case agents attest to the accuracy of FISA reviews, a recommendation FBI adopted, which might accomplish more than any review by an outside ASAC.

There’s one more reason this memo is concerning. ABC reported the other day that long-time Deputy Assistant Attorney General for Legal Policy Brad Wiegmann was reassigned two weeks ago and replaced by a far less experienced political appointee, Kellen Dwyer (though I’ve seen people vouch for his integrity — he’s not a hack). Wiegmann would likely be part of discussions about how to meet FISC’s demands for further accountability.

Though a relatively small unit of fewer than two dozen attorneys, the Office of Law and Policy participates in almost every National Security Council meeting, works with congressional staff to draft new legislation, and conducts oversight of the FBI’s intelligence-gathering activities.

“[It] has been sort of the center of gravity for the Department of Justice on national security policy, and it’s a central role,” said Olsen, who at one point ran the department’s National Security Division and later advised Hillary Clinton’s 2016 presidential campaign.

Wiegmann has led the office since the Obama administration and for almost all of the Trump administration.

In particular, Wiegmann has long been involved in efforts to meet FISC’s demands regarding surveillance it authorizes. Here, just days after Wiegmann’s removal, Barr is issuing a memo that seems unaware of and in at least a few respects, potentially inconsistent with, explicit orders from the presiding FISA Judge.

There’s nothing obviously offensive about this memo. But it would do little to prevent a repeat of the Carter Page problems. And it’s not clear that it adds anything to the very real efforts to improve the FISA process at DOJ. Indeed, it may well be an effort to pre-empt more substantive concerns about the role of NSD (as opposed to FBI) in this process.

Barr released a second memo creating an audit mechanism for national security functions that feels like an effort to get ahead of ongoing DOJ IG investigation. I welcome additional oversight of FBI’s national security functions, though the timing of this and the timing of its implementation — with a report on its creation due just days before the election but all review of its functionality years down the road — feels like an attempt to stave off real legal oversight.


DOJ’s Accounting of Its FISA Errors Cannot Be Compared to the Carter Page Report

Last year, Bill Barr adopted the stance that Inspector General Michael Horowitz’s assessment of FISA — in the report on the Carter Page FISA applications — wasn’t strict enough, because it found no evidence that the errors in the applications arose from political bias. Last week, Bill Barr’s DOJ adopted the opposite stance, that DOJ IG was too critical of FISA, finding errors in the FBI process where there were none.

It did so in the second of two filings reviewing the errors that DOJ IG had found in 29 other FISA applications. When DOJ IG released an interim report (MAM) describing those errors in March, it appeared to suggest that the level of error in the Carter Page applications — at least with respect to the Woods Files — was actually lower than what DOJ IG had found in the 25 applications.

Now, DOJ appears to be trying to claim — without basis — that that’s not the case.

Ahead of the release of the actual filing, DOJ and FBI orchestrated a press release last week, announcing that they would tell the court none of the errors identified by DOJ IG invalidated the probable cause finding for the 29 files. Predictably, both the responsible press and the frothy right (in stories that misunderstood the findings of either DOJ IG report and at times made errors about the FISA process), concluded that this review shows that Page’s application was uniquely bad.

Only after the press had jumped on that conclusion did DOJ release the filing (here’s the earlier one and here’s AAG John Demers’ statement in conjunction with last week’s release).

The filing makes it clear that it is impossible to draw any comparison between these findings about the earlier Carter Page ones (or even to declare — as many in the press have — that this filing proves DOJ’s FISA problems aren’t as bad as DOJ IG suggested).

That’s true for three reasons:

  • DOJ IG has not finished the kind of review on any of the 29 files it did for Page, and DOJ is not claiming it did either
  • DOJ used a dramatically different methodology for this Woods review than DOJ IG did for the Page review
  • DOJ effectively disagreed with DOJ IG’s findings for roughly 46% of the errors DOJ IG identified — and it’s not clear they explained to the FISA Court why they did so

Before I explain these, there’s a more important takeaway.

In giving itself a clean bill of health, DOJ judged that it doesn’t matter that a 2016 FISA application claimed that one of their sources accused a person of sympathizing with a particular terrorist organization when in fact the source said the person had become sympathetic to radical Muslim causes. For the purposes of FISA, this is a huge distinction, because a terrorist organization counts as a foreign power for the sake of FISA, but radical Muslim causes do not. It’s the difference between targeting someone as a suspected agent of a foreign power and targeting them for First Amendment protected activities. DOJ said this error didn’t matter because there was so much other derogatory information against the target; whether that’s true or not, it remains the case that DOJ’s self-congratulation nevertheless admits to a key First Amendment problem in one of the applications.

Woods violations are different from significant inaccuracies are different from material inaccuracies are different from probable cause

As I explained in this post, the IG Report on Carter Page found two types of problems: 17 “significant inaccuracies” that were mostly errors of omission (see PDF 12 and 14-15 for a list), and Woods file errors (PDF 460ff) for which an assertion made in the application did not have or match the back-up in the accuracy file that is supposed to prove it. The “significant inaccuracies” are the more serious of the two, but a number of those were overblown and in a few cases, dubious, in the DOJ IG Report.

Both of those categories are different from material misstatements, of which DOJ admitted to a number by the time they withdrew the probable cause claim from the third and fourth, but not the first two, Page applications. Before the conclusion of the DOJ IG Report they had told the court of the following material misstatements:

  • July 12, 2018: Cover stories Papadopoulos gave to informants that FBI accurately assessed in real time as false, statements Bruce Ohr made that (in the slightly misrepresented form included in the DOJ IG Report) call into question Christopher Steele’s motives, admissions that Steele himself had spoken to the press
  • October 25, 2019 and November 27, 2019: Details about the actions of Kevin Clinesmith — first not disclosing and then altering a document to hide Page’s relationship with the CIA that covered some but not all of his willful sharing of non-public information with known Russian intelligence officers

It’s not clear the government specified which aspects of the DOJ IG Report it submitted to Rosemary Collyer in December 2019 it deemed material, but she focused on:

  • Statements made by Steele’s primary sub-source that undermined key claims about Page
  • Page’s denials (some proven true, some of still undetermined veracity) of details in the Steele dossier
  • Steele’s derogatory comments about Sergei Millian

On the scale of severity, the material misstatements are the ones that matter, because they’re the ones that will affect whether someone gets wiretapped or not. But the Woods file errors in the Carter Page report identified by DOJ IG describe just four (arguably, three) details even related to things ultimately deemed material which, in turn, led to the withdrawal of two of the applications. None directly described the core issues that led to the withdrawal of the two applications (though the Page denials in conjunction with the sub-source comments did).

Indeed, one key conclusion of this entire process — one that DOJ, DOJ IG, and FISC have all agreed with — is that the Woods files process is not very useful at finding the more important errors of omission of the kind that were the most serious problems in the Page application.

And that’s important because all three of these reports — the March DOJ IG MAM and the June and July responses to FISA — stem from, and only explicitly claim to address, Woods file errors. In its MAM, DOJ IG described what it called its “initial” review this way:

During this initial review, we have not made judgments about whether the errors or concerns we identified were material. Also, we do not speculate as to whether the potential errors would have influenced the decision to file the application or the FISC’s decision to approve the FISA application. In addition, our review was limited to assessing the FBI’s execution of its Woods Procedures, which are not focused on affirming the completeness of the information in FISA applications.

For its part, DOJ calls DOJ IG’s report “preliminary” (seemingly ignoring that the IG claimed in that MAM and claims on its website to be continuing this part of what it calls a preliminary part of a larger review of FISA). DOJ’s Office of Intelligence did do materiality reviews of both the errors DOJ IG found and some that it found in the process of compiling these reports (in addition to the CT material misstatement described above, it found what sounds like the omission of exculpatory statements in a CI case).

But all this amounts to the more basic of the two kinds of reviews that DOJ IG did in the Carter Page case.

For these reports, DOJ continued to use the accuracy review methodology it now agrees is inadequate

As noted, all parties now agree that the Woods procedure wasn’t doing what it was supposed to do. One reason it wasn’t is because the FBI has always given agents a few weeks notice before they review one of their Woods files, allowing them to scramble to fill out the accuracy file.

But DOJ IG (perfectly reasonably) didn’t give the Crossfire Hurricane team or any of the people involved in the 29 FISA applications it reviewed here that same notice. It conducted its Woods file assessment on what was actually in the accuracy file. In the case of the Carter Page review, they found a placeholder for a 302 that said exactly what DOJ IG faulted FBI for not having evidence for, an observation about how much Stefan Halper has been paid, and publicly available details about Gazprombank, among other true claims that were nevertheless not backed up in the Woods file. It would have been child’s play — but take some work — to get proof of those and most other claims in the file. The Woods file review that DOJ IG did in the Page case — and almost certainly, the review of the 29 files — tested whether the Woods procedures were being adhered to at all, not whether the Woods procedure effectively ensured only documented claims made it into a FISA application.

If you’re going to rely on the Woods procedure as an accuracy tool, that’s what reviews need to do, because otherwise they’re doing nothing to test the accuracy of the reports.

And DOJ now agrees. In its June filing, DOJ committed to changing how it does accuracy reviews starting in September (maybe). Starting then, agents will get no notice of a review before it happens, and the accuracy rate of that no-notice review will be tracked along with the accuracy once an agent is given time to chase down the documentation he didn’t include the first time.

NSD has determined that commencing with accuracy reviews starting after September 30, 2020, it will not inform the FBI field offices undergoing NSD oversight reviews which applications will be subjected to accuracy reviews in advance of those reviews. This date is subject to current operational limitations the coronavirus outbreak is imposing. NSD would not apply this change in practice to accuracy reviews conducted in response to a request to use FISA information in a criminal proceeding, given the need to identify particular information from particular collections that is subject to use. NSD also would not apply this change in practice to completeness reviews ( discussed further below); because of the pre-review coordination that is contemplated for those reviews.

NSD will expect that the relevant FBI field offices have ready, upon NSD’s arrival, the accuracy sub-files for the most recent applications for all FISAs seeking electronic surveillance or physical search. NSD will then, on its arrival, inform the FBI field office of the application(s) that will be subject to an accuracy review. If the case will also be subject to a completeness review, pre-coordination, as detailed below, will be necessary. The Government assesses that implementing this change in practice will encourage case agents in all FISA matters to be more vigilant about applying the accuracy procedures in their day-to-day work.

In addition, although NSD’s accuracy reviews allow NSD to assess individual compliance with the accuracy procedures, NSD’s historical practice has been to allow agents to obtain documentation during a review that may be missing from the accuracy sub-file. NSD only assesses the errors or omissions identified once the agent has been given the opportunity to gather any additional required documentation. While the Government believes that, in order to appropriately assess the accuracy of an application’s content, it should continue to allow agents to gather additional documentation during the accuracy review, it assesses that this historical practice has not allowed for the evaluation of how effective agents have been at complying with the requirement to maintain an accuracy sub-file, complete with all required documentation.

As a result, NSD will tally and report as a part of its accuracy review process all facts for which any documentation, or appropriate documentation, was not a part of the accuracy sub-file at the time the accuracy review commenced.

That said, that’s not how DOJ did these reviews. In fact, John Demers emphasized this fact in his statement claiming victory over these reviews.

In addition, when the OIG found a fact unsupported by a document in the Woods file, the OIG did not give the FBI the opportunity to locate a supporting document for the fact outside the file.

Indeed, that’s not the only thing that DOJ did to help DOJ clean up DOJ’s shitty performance on DOJ IG’s review of their work. After FBI Field Office lawyers got the DOJ IG assessment, they pulled together the existing documentation, then DOJ’s OI worked with agents to fill in what wasn’t there. In fact, DOJ even got an extension on the second report because DOJ and FBI agents were still working through the files, suggesting it took up to three months of work to get the files to where DOJ was willing to tell FISC about them.

In other words, whereas the Crossfire Hurricane team got judged — by Bill Barr’s DOJ — on what was in the Woods file when DOJ IG found it, Bill Barr’s DOJ is judging Bill Barr’s DOJ on what might be in a Woods file after agents have up to three months to look for paperwork to support claims they made as long as six years ago.

DOJ disagreed with DOJ IG’s finding of error about 46% of the time

Finally, DOJ and DOJ IG did not use the same categories of information to track errors on the Woods file reviews, and one of the most common ways they dismissed the import of an error was by saying that DOJ IG was wrong.

The MAM divides the errors it found into three categories: claims not supported by any documentation, claims not corroborated by the supposed back-up, and claims that were inconsistent with the supporting documentation.

[W]e identified facts stated in the FISA application that were: (a) not supported by any documentation in the Woods File, (b) not clearly corroborated by the supporting documentation in the Woods File, or (c) inconsistent with the supporting documentation in the Woods File.

In addition to the two material errors they found, DOJ claims the errors they found fall into five categories (described starting on page 10):

  • Non-material date errors
  • Non-material typographical errors
  • Non-material deviations from the source documentation
  • Non-material misidentified sources of information
  • Non-material facts lacking supporting documentation

But to get to that number, DOJ also weeded out a number of other problems identified by DOJ IG via three other categories of determination reflected in the up to three month back and forth with OI:

  • Claims made that were substantiated by documents added to the file after DOJ IG’s review
  • Claims that, after reviewing additional information, OI “determined that the application accurately stated or described the supporting documentation, or accurately summarized other assertions in the application that were supported by the accuracy subfile”
  • Claims not backed by any document, but for which “the supporting documentation taken as a whole provided support for the fact in the application”

DOJ doesn’t count those instances in its overview — as distinct from individual narratives — of the report (indeed, the scope of added documentation is not qualified at all). And while the DOJ fillings say FBI described that it added documentation to the file in the redacted FBI declaration for FISC, it’s not clear whether it told FISC what it added and how much and where and when it came from (FBI has been known to write 302s long after the fact to document events not otherwise documented in real time).

Here’s what all this looks like in one table (FBI did what is probably a similar table, but it’s classified). Note that DOJ IG used still different categories for the Carter Page review: “Supporting document does not state this fact,” which is probably the same as their “not clearly corroborated” category. In my table, I’ve counted that as a “lacking documentation error.”

There are several takeaways from this table.

First, the numerical discrepancy provides some idea of how many errors DOJ IG found that DOJ made go away either by finding documentation for them, or by deciding that DOJ IG was wrong. DOJ IG said it found an average of 20 errors in the 25 applications it was able to review, or 500 total. DOJ says it found 63 errors in the June report and 138 errors in the July Report, over a total of 29 applications (they did a review of the four files for which DOJ IG was provided with no Woods file, so had 4 more files than DOJ IG).

My numbers are off by 3 from theirs, which might be partly accounted for recurrent errors in a reauthorized application or lack of clarity on DOJ’s narrative. Or maybe like DOJ, I subtracted 48 from 138 and got 91.

Approximately 48 of these 138 non-material errors reflect typographical errors or date discrepancies between an assertion in an application and a source document. Of the remaining 91 non-material errors or unsupported facts, four involve nonmaterial factual assertions that may be accurate, but for which a supporting document could not be located in the FBI’s files; 73 involve non-material deviations between a source document and an application; and 13 involve errors in which the source of an otherwise accurate factual assertion was misidentified.

But my count shows that DOJ simply declared DOJ IG to be wrong 151 times in its assessment that something was an error, with an amazing 35 examples of that in one application, and of which 14 across all applications were instances where DOJ couldn’t find a document to support a claim (not even with three months to look), but instead said the totality of the application supported a claim.

Claiming that the totality of an application supports a claim, while being unable to find documentation for a discrete fact, sure sounds like confirmation bias.

And in the up to three months of review, FBI found documentation to support upwards of 130 claims that originally were not supported in the Woods file. In other words, these weren’t errors of fact — they were just instances of FBI not following the Woods procedure.

We know that if the Crossfire Hurricane team had been measured by the standard DOJ did in these filings, it would have done better than most of these applications (again, only with respect to the Woods file). That’s because, aside from the four claims that rely on intercepted information (which is not public), there is public documentation to support every claim deemed unsupported in the report but three: the one claiming that James Clapper had said that Russia was providing money in addition to the disinformation to help Trump.

The DNI commented that this influence included providing money to particular candidates or providing disinformation.

And the two claiming that Christopher Steele’s reporting had been corroborated, something the DOJ IG Report lays out at length was not true in the terms FBI normally measured. Except, even there, Steele handler Mike Gaeta’s sworn testimony actually said it had been. He described jumping when Steele told him he had information because he was a professional,

And at that time there were a number of instances when his information had borne out, had been corroborated by other sources.

He also provided a perfectly reasonable explanation for why Steele’s reporting was not corroborated in the way DOJ IG measured it in the report: because you could never put Steele on a stand, so his testimony would never be used to prosecute people.

From a criminal perspective and a criminal investigative kind of framework, you know, Christopher Steele and [redacted] were never individuals who were going to be on a witness stand.

In other words, while it appears that DOJ cleaned up many of the errors identified by DOJ IG by finding the documentation to back it over the course of months, the public record makes it clear that Crossfire Hurricane would have been able to clear up even more of the Page Woods file.

The exceptions prove the rule. There are, as my table notes, two or three claims that do not accurately describe what the underlying document says, claiming:

  • That Page never refuted the claims against him (he had, and in many cases, was telling the truth in his refutations)
  • That Steele told the FBI he never shared information with anyone outside his “business associate” [Fusion] and the FBI (he also shared it with State, as other parts of FBI had been told)
  • That in his first FBI interviews Papadopoulos admitted he had met with Australian officials but not that he discussed Russia during those meetings (it’s unclear how accurate this claim is)

Assume the last bullet (used just once) reflects the redacted parts of Papadopoulos’ 302s even though it does match his current statements, that nevertheless leaves you with an error rate on arguably the worst category — misrepresenting your evidence — of 2 or 3 per application. The first two of these are the Woods file errors that turned out to have a tie (a significant one in the first bullet) with the material reasons why some of the files were withdrawn. They’re the two errors in the Woods file that most directly tied to omitted evidence in the application that would lead to their withdrawal.

Of the 29 applications reviewed by DOJ, 12 of them have 3 or more “deviations from the source” material. One has 14 and another has 15.

So on the worst measure that this review actually did measure, the one that on Page’s application tied most directly to reasons to withdraw the application, Page’s application actually was within the norm.

It may well be that when all the reviews are done, DOJ will have proof that Carter Page’s application was an exceptionally bad application. Certainly, the material misstatements may end up being worse.

But the only thing this apples to oranges comparison of the Page methodology and the traditional DOJ methodology has proven is that — as a matter of the Woods file reviews — Bill Barr has used a different standard for Bill Barr’s DOJ than he has with Crossfire Hurricane. And that if the Page file had been treated as all the others were, from a Woods file perspective, it actually wouldn’t look that bad.

It also shows that when Bill Barr’s DOJ wants to continue spying on Americans who don’t happen to be associated with Donald Trump, he’s happy to argue that Michael Horowitz’s very legalistic reviews of the sort that did Andrew McCabe in are wrong.

Updated for clarity.

Task and Countertask: The Interview of Christopher Steele’s Primary Subsource

According to the interview report from Christopher Steele’s Primary Subsource, the PSS confirmed that he had two sources behind the reporting that Carter Page met with Igor Sechin. He said one of those two sources — whom he described having ties to FSB — told him that Russia was sitting on kompromat against Trump (and Hillary). He described that his source for all the Michael Cohen reporting came from an old friend whom he trusted 100%. Steele’s Primary Subsource even took credit for some of the specific phrases in the Steele dossier — such as the one describing Michael Cohen’s efforts to sweep the Carter Page and Paul Manafort scandals “under the carpet.”

Even the Primary Subsource’s interactions with a person he believed to be Sergei Millian tracked most of the report based off the call.

[PSS] recalls that this 10-15 minute conversation included a general discussion about Trump and the Kremlin, that there was “communication” between the parties, and that it was an ongoing relationship. [PSS] recalls that the individual believed to be [Millian] said that there was an “exchange of information” between Trump and the Kremlin, and that there was “nothing bad about it,” Millian said that some of the information exchange could be good for Russian, and some could be damaging to Trump, but deniable. The individual said that the Kremlin might be of help to get Trump elected, but [PSS] did not recall any discussion or mention of Wikileaks.

The passage shows how badly DOJ IG over-read the interview when it first published the report and affirmatively stated that PSS “had no discussion” or “made no mention at all of” WikiLeaks.

On pages xi, 242, 368, and 370, we changed the phrase “had no discussion” to “did not recall any discussion or mention.” On page 242, we also changed the phrase “made no mention at all of” to “did not recall any discussion or mention of.” On page 370, we also changed the word “assertion” to “statement,” and the words “and Person 1 had no discussion at all regarding WikiLeaks directly contradicted” to “did not recall any discussion or mention of WikiLeaks during the telephone call was inconsistent with.” In all instances, this phrase appears in connection with statements that Steele’s Primary Sub-source made to the FBI during a January 2017 interview about information he provided to Steele that appeared in Steele’s election reports. The corrected information appearing in this updated report reflects the accurate characterization of the Primary Sub-source’s account to the FBI that previously appeared, and still appears, on page 191, stating that “[the Primary SubSource] did not recall any discussion or mention of Wiki[L]eaks.”

To be sure, the provenance of that claimed Millian conversation is an utter shitshow — consisting of a call with someone the Primary Subsource believed, but had no way of confirming, was Millian. But Steele’s Primary Subsource did confirm that most of that report tracked the call, whoever it was from.

Still, you wouldn’t know that the Primary Subsource described the multiple sources behind key allegations in the dossier from the way the DOJ IG Report described what was a raw intelligence report. For example, this passage doesn’t reveal that the Primary Subsource heard details on Page’s trip from people with high level connections, including the meeting with Sechin (remember, the FBI had another source report that he had heard rumors about the Sechin meeting, which probably partly explains why Mueller concluded that Page’s whereabouts in Russia were still uncertain).

A second example provided by the Primary Sub-source was Report 134’s description of a meeting allegedly held between Carter Page and Igor Sechin, the President of Rosneft, a Russian energy conglomerate. 337 Report 134 stated that, according to a “close associate” of Sechin, Sechin offered “PAGE/ TRUMP’s associates the brokerage of up to a 19 percent (privatized) stake in Rosneft” in return for the lifting of sanctions against the company. 338 The Primary Sub-source told the FBI that one of his/her subsources furnished information for that part of Report 134 through a text message, but said that the sub-source never stated that Sechin had offered a brokerage interest to Page. 339 We reviewed the texts and did not find any discussion of a bribe, whether as an interest in Rosneft itself or a “brokerage. ” 340

The IG Report also repeats uncritically stuff from both the PSS and his sources that is pretty obviously bullshit, such as the claim from the PSS — who had been paid full time by Orbis for years to collect this intelligence — that he didn’t expect his reporting to show up in written reports.

The Primary Subsource also stated that he/she never expected Steele to put the Primary Subsource’s statements in reports or present them as facts. According to WFO Agent 1, the Primary Sub-source said he/she made it clear to Steele that he/she had no proof to support the statements from his/her sub-sources and that “it was just talk.” WFO Agent 1 said that the Primary Sub-source explained that his/her information came from “word of mouth and hearsay;” “conversation that [he/she] had with friends over beers;” and that some of the information, such as allegations about Trump’s sexual activities, were statements he/she heard made in “jest.”341 The Primary Sub-source also told WFO Agent 1 that he/she believed that the other sub-sources exaggerated their access to information and the relevance of that information to his/her requests.

Or the claim from a subsource who would be the key source of disinformation in the dossier if such disinformation exists that nothing in the dossier was attributable to her.

FBI documents reflect that another of Steele’s sub-sources who reviewed the election reporting told the FBI in August 2017 that whatever information in the Steele reports that was attributable to him/her had been “exaggerated” and that he/she did not recognize anything as originating specifically from him/her. 347

Nor would you know that from the reporting on the interview report of the Primary Subsource, released last night by Lindsey Graham.

Ultimately, the belated assessment of the Supervisory Intel Analyst probably appropriately attributes blame for problems with the dossier to multiple sources; a lot of the problems with this dossier stem from communication breakdowns and exaggerations from multiple people trying to make a buck.

According to the Supervisory Intel Analyst, the cause for the discrepancies between the election reporting and explanations later provided to the FBI by Steele’s Primary Sub-source and sub-sources about the reporting was difficult to discern and could be attributed to a number of factors. These included miscommunications between Steele and the Primary Sub-source, exaggerations or misrepresentations by Steele about the information he obtained, or misrepresentations by the Primary Sub-source and/or sub-sources when questioned by the FBI about the information they conveyed to Steele or the Primary Sub-source.

Let me be very clear: none of this means these allegations are true, nor does this excuse the failures to alert the FISA Court to key problems in the dossier. I was one of the first people to raise doubts about some of the problems with the allegations in the dossier, and I stand by that.

Operational security

What’s more interesting about the interview are the hints of all the ways the dossier could have gone so badly wrong. The interview report describes multiple ways that Russia’s spooks might have found out about the project and fed it with disinformation (the footnotes declassified earlier this year describes that several Russian spooks knew of the project after what would have been the PSS’ first trip to Russia to do the reporting).

Steele’s PSS was an analyst by training that Steele increasingly used in an operational role (including by getting him hired at some kind of consulting company that seems to have served as a kind of cover for his travel to Russia). The arrangement seems to have had spotty operational security. For better and worse, PSS said that he rarely took substantive notes.

[PSS] was asked if he takes notes on the information he is collecting from his sources, or if he keeps any kind of records. He was told by Steele that it is a security risk to take notes; he hasn’t kept notes or electronic records. He occasionally makes scribbles and/or chicken scratch notes here and there, but gives verbal debriefs in [redacted] following his trips [to Russia].

PSS would then share the information with Steele, whom he always briefed alone (making misunderstandings more likely). He had no communications with Steele while in Russia. PSS described that his debriefings with Steele were always at the Orbis office, which meant if Steele himself were surveilled, PSS’ ties to Steele would become obvious.

PSS was originally tasked to investigate Manafort (which he had little success on), at a time when Fusion was still being paid by Paul Singer, meaning this interview seems to confirm, once and for all, that not just Fusion’s reporting, but Steele’s, was initially paid for by a Republican. PSS specified for that reporting he did some of his reporting to Steele via an encrypted app.

But his communications with Steele included many insecure methods. He first met Steele in a Starbucks. Early on, he communicated with him via email and Skype, and Steele would task him, at least in part, via email. He described discussing Page’s trip to Russia with Source 3 on some kind of voice call, possibly a phone, while he was at a public swimming pool, though he also described talking in an opaque way about election interference. Likewise, the most problematic December 13 report was based on a conversation with the same source, which was also a phone call.

In short, while Steele and PSS and PSS’ sources made some efforts to protect their communications from the Russians that surely considered Steele a target, those efforts were inconsistent.

PSS described making three trips to Russia for his election year reporting. On the second trip, he got grilled suspiciously at the border. On his third, “nothing bad happened,” which made PSS suspicious about how perfectly everything had gone.

PSS repeatedly described being uncomfortable with the election year tasking, and he seems to have had suspicions in real time that Russia had taken note of it.

Ties to intelligence

Meanwhile, for all the reports that PSS was “truthful and cooperative,” the interview report describes that he “balked, meandered in the conversation, and did not really answer the question” about whether he used other sources for his election year reporting aside from the six he described to the FBI. And, as laid out in the interview report, it became increasingly clear over the three days of interviews that PSS was not entirely forthcoming about any interactions he had had with Russian intelligence.

This started with his lawyers’ careful caveat at the beginning of the process that PSS did not have any contacts with people he knew to be part of the Russian intelligence services (the interview as a whole was conducted under a proffer).

[PSS] indicated, to his knowledge, he has not had any contacts with the Russian intelligence or security services. [ANALYST NOTE: His attorney emphasized “to his knowledge” during this part of the discussion.]

PSS said he had contact with Russian government officials, but — “as far as he … knew,” not with anyone in SVR, GRU, or FSB.

On day three, however, PSS described a friend (whose experience he drew on for a report on how Russia coerces criminal hackers to work for the intelligence services) who had had been busted for involvement with online pornography and pressured to work with the FSB. The Senior Intel Analyst noted that conflicted with his earlier claim to have no known ties to Russian spooks.

[ANALYST NOTE: This is in contradiction to [PSS’s] statement the first day, at which time he indicated that he did not have any contacts associated with the Russian intelligence and security services.]

Later that same day, PSS seemed to acknowledge that a Russian official and a Russian journalist he interacted with were spooks. The FBI noted,

[ANALYST NOTE: This contradicted [PSS’s] earlier statements regarding having no contact with Russia’s intelligence and security services, and it also contradicted regarding not really knowing if [a Russian official] was actually connected to Russia’s intelligence and security services.]

The EC goes on to describe PSS “brush[ing] aside the idea of being approached by the intelligence and security services” while he was a student.

This squirreliness about his own ties with Russian spooks was probably just self-preservation, an effort to avoid any exposure on 18 USC 951, but it is probably the key issue where the FBI questioned his candor in real time.


Meanwhile, PSS described at least three of his sources — Source 1, Source 2, and Source 3 — in such a way that led the FBI to wonder whether PSS was being tasked by his own sources. S1, for example — who has a close relationship to a Russian intelligence officer (probably FSB) —  always asks PSS to do projects together.

[S1] is always trying to get [PSS] to start projects and make money together — [PSS] related how [S1], like others, is always asking questions like, “Can you get us some projects?” or “Can you get us financing?” or “Let’s do something together dealing with [redacted]!” [PSS] doesn’t consider this as his source “tasking him” but as simply the normal course and scope of networking in these circles. [PSS] did help [S1] with an academic book about [redacted].

And both Source 2 and Source 3 — the sources for some of the more problematic information in the Steele dossier — knew PSS brokered intelligence. Both also discussed brokering information in Russia.

[S3] is one of the individuals who knows that [PSS] works for due diligence and business intelligence. [As an aside at this point, [PSS] insisted that [S2] probably has a better idea about this than does [S3] because [S2] is always trying to monetize his relationship with [PSS]. [PSS] reiterated again to interviewers that [S2] will often pitch money-making ideas or projects — “Let’s work together. I [S2] can try and get [redacted] to answer a question, but I’ll need some money to do it.”] [S3] has an understanding that [PSS] is “connected.” In fact, either [redacted] morning or [redacted] morning, [S3] reached out to [PSS] and asked him for help in [redacted] on how [redacted] living in the United States are viewing the Trump administration. She is asking him [redacted] by the weekend, probably so she can sell it to a friend in Moscow.

And because PSS asked Orbis to help S1 — the guy with close ties to an FSB officer — get a scholarship for language study in the UK, S1 presumably knows what Orbis and who Steele is.

In addition to S1, Source 5 also has ties to Russian intelligence. This showed up in footnote 339, which was partly declassified earlier this year.

This is to be expected, of course. Indeed, the dossier prominently touts the intelligence sourcing of its allegations, as I noted the first day the dossier was published. If the person on whose source network Steele was relying didn’t have ties to spooks, it would be as problematic.

The thing, though, is that it’s certain now that many of the allegations in the dossier are not true or were rumor, particularly virtually all the allegations sourced to Source 3 (the source for all the Michael Cohen reporting), PSS’s childhood friend whom he trusts 100%. That’s true even though generally the reports were sourced to people with at least indirect access to senior level officials.

All the huffing and puffing aside, that should be the takeaway from this. Steele was definitely not collecting this intelligence in optimal fashion, and sharing it with the press made things far worse. But in January 2017, it looked like raw intelligence, of varying quality, which is precisely what it was billed at. Yet, well before any pitches Steele made to the press, it seems some really well-connected people in Russia were feeding Steele’s PSS information that distracted from the real events going on and focused it elsewhere.

The Government Argues that Edward Snowden Is a Recruiting Tool

As I noted in my post on the superseding indictment against Julian Assange, the government stretched the timeline of the Conspiracy to Hack count to 2015 by describing how WikiLeaks helped Edward Snowden flee to Russia. DOJ seems to be conceiving of WikiLeaks’ role in helping Snowden as part of a continuing conspiracy designed to recruit more leakers.

Let me make clear from the onset: I am not endorsing this view, I am observing where I believe DOJ not only intends to head with this, but has already headed with it.

Using Snowden as a recruitment tool

After laying out how Chelsea Manning obtained and leaked files that were listed in the WikiLeaks Most Wanted list (the Iraq Rules of Engagement and Gitmo files, explicitly, and large databases more generally; here’s one version of the list as entered into evidence at Manning’s trial), then describing Assange’s links to LulzSec, the superseding Assange indictment lays out WikiLeaks’ overt post-leak ties and claimed ties to Edward Snowden.

83. In June 2013, media outlets reported that Edward J. Snowden had leaked numerous documents taken from the NSA and was located in Hong Kong. Later that month, an arrest warrant was issued in the United States District Court for the Eastern District of Virginia, for the arrest of Snowden, on charges involving the theft of information from the United States government.

84. To encourage leakers and hackers to provide stolen materials to WikiLeaks in the future, ASSANGE and others at WikiLeaks openly displayed their attempts to assist Snowden in evading arrest.

85. In June 2013, a WikiLeaks association [Sarah Harrison, described as WLA-4 in the indictment] traveled with Snowden from Hong Kong to Moscow.

86. On December 31, 2013, at the annual conference of the Chaos Computer Club (“CCC”) in Germany, ASSANGE, [Jacob Appelbaum] and [Harrison] gave a presentation titled “Sysadmins of the World, Unite! A Call to Resistance.” On its website, the CCC promoted the presentation by writing, “[t]here has never been a higher demand for a politically-engaged hackerdom” and that ASSANGE and [Appelbaum] would “discuss what needs to be done if we re going to win.” ASSANGE told the audience that “the famous leaks that WikiLeaks has done or the recent Edward Snowden revelations” showed that “it was possible now for even a single system administrator to … not merely wreck[] or disabl[e] [organizations] … but rather shift[] information from an information apartheid system … into the knowledge commons.” ASSANGE exhorted the audience to join the CIA in order to steal and provide information to WikiLeaks, stating, “I’m not saying doing join the CIA; no, go and join the CIA. Go in there, go into the ballpark and get the ball and bring it out.”

87. At the same presentation, in responding to the audience’s question as to what they could do, [Appelbaum] said “Edward Snowden did not save himself. … Specifically for source protection [Harrison] took actions to protect [Snowden] … [i]f we can succeed in saving Edward Snowden’s life and to keep him free, then the next Edward Snowden will have that to look forward to. And if look also to what has happened to Chelsea Manning, we see additionally that Snowden has clearly learned….”

The following section describes how, “ASSANGE and WikiLeaks Continue to Recruit,” including two more paragraphs about the Most Wanted Leaks:

89. On May 15, 2015, WikiLeaks tweeted a request for nominations for the 2015 “Most Wanted Leaks” list, and as an example, linked to one of the posts of a “Most Wanted Leaks” list from 2009 that remained on WikiLeaks’s website.


92. In June 2015, to continue to encourage individuals to hack into computers and/or illegaly obtain and disclose classified information to WikiLeaks, WikiLeaks maintained on its website a list of “The Most Wanted Leaks of 2009,” which stated that documents or materials nominated to the list must “[b]e likely to have political, diplomatic, ethical or historical impact on release … and be plausibly obtainable to a well-motivated insider or outsider,” and must be “described in enough detail so that … a visiting outsider not already familiar with the material or its subject matter may be able to quickly locate it, and will be motivated to do so.”

Effectively, Snowden is included in this indictment not because the government is alleging any ties between Snowden and WikiLeaks in advance of his leaks (Snowden’s own book lays out reasons to think there was more contact between him and Appelbaum than is publicly known, but the superseding Assange indictment makes no mention of any contacts before Snowden’s first publications), but because WikiLeaks used their success at helping Snowden to flee as a recruiting pitch.

Snowden admits Harrison got involved to optimize his fate

This is something that Snowden lays out in his book. First, he addresses insinuations that Assange only helped Snowden out of selfish reasons.

People have long ascribed selfish motives to Assange’s desire to give me aid, but I believe he was genuinely invested in one thing above all—helping me evade capture. That doing so involved tweaking the US government was just a bonus for him, an ancillary benefit, not the goal. It’s true that Assange can be self-interested and vain, moody, and even bullying—after a sharp disagreement just a month after our first, text-based conversation, I never communicated with him again—but he also sincerely conceives of himself as a fighter in a historic battle for the public’s right to know, a battle he will do anything to win. It’s for this reason that I regard it as too reductive to interpret his assistance as merely an instance of scheming or self-promotion. More important to him, I believe, was the opportunity to establish a counterexample to the case of the organization’s most famous source, US Army Private Chelsea Manning, whose thirty-five-year prison sentence was historically unprecedented and a monstrous deterrent to whistleblowers everywhere. Though I never was, and never would be, a source for Assange, my situation gave him a chance to right a wrong. There was nothing he could have done to save Manning, but he seemed, through Sarah, determined to do everything he could to save me.

This passage is written to suggest Snowden believed these things at the time, describing what “seemed” to be true at the time. But it’s impossible to separate it from Appelbaum’s explicit comparison of Manning and Snowden at CCC in December 2013.

Snowden then describes what he thinks Harrison’s motive was.

By her own account, she was motivated to support me out of loyalty to her conscience more than to the ideological demands of her employer. Certainly her politics seemed shaped less by Assange’s feral opposition to central power than by her own conviction that too much of what passed for contemporary journalism served government interests rather than challenged them.

Again, this is written to suggest Snowden believed it at the time, though it’s likely what he has come to believe since.

Then Snowden describes believing, at that time, that Harrison might ask for something in exchange for her help — some endorsement of WikiLeaks or something.

As we hurtled to the airport, as we checked in, as we cleared passport control for the first of what should have been three flights, I kept waiting for her to ask me for something—anything, even just for me to make a statement on Assange’s, or the organization’s, behalf. But she never did, although she did cheerfully share her opinion that I was a fool for trusting media conglomerates to fairly guard the gate between the public and the truth. For that instance of straight talk, and for many others, I’ll always admire Sarah’s honesty.

Finally, though, Snowden describes — once the plane entered into Chinese airspace and so narratively at a time when there was no escaping whatever fate WikiLeaks had helped him pursue — asking Harrison why she was helping. He describes that she provided a version of the story that WikiLeaks would offer that December in Germany: WikiLeaks needed to be able to provide a better outcome than the one that Manning suffered.

It was only once we’d entered Chinese airspace that I realized I wouldn’t be able to get any rest until I asked Sarah this question explicitly: “Why are you helping me?” She flattened out her voice, as if trying to tamp down her passions, and told me that she wanted me to have a better outcome. She never said better than what outcome or whose, and I could only take that answer as a sign of her discretion and respect.

Whatever has been filtered through time and (novelist-assisted) narrative, Snowden effectively says the same thing the superseding indictment does: Assange and Harrison went to great lengths to help Snowden get out of Hong Kong to make it easier to encourage others to leak or hack documents to share with WikiLeaks. I wouldn’t be surprised if these excerpts from Snowden’s book show up in any Assange trial, if it ever happens.

Snowden’s own attempt to optimize outcomes

Curiously, Snowden did not say anything in his book about his own efforts to optimize his outcome, which is probably the most interesting new information in Bart Gellman’s new book, Dark Mirror (the book is a useful summary of some of the most important Snowden disclosures and a chilling description of how aggressively he and Askhan Soltani were targeted by foreign governments as they were reporting the stories). WaPo included the incident in an excerpt, though the excerpt below is from the book.

Early on in the process, Snowden had asked Gellman to publish the first PRISM document with a key, without specifying what key it was. When WaPo’s editors asked why Gellman’s source wanted them to publish a key, Gellman finally asked.

After meeting with the Post editors, I remembered that I could do an elementary check of the signature on my own. The result was disappointing. I was slow to grasp what it implied.

gpg –verify PRISM.pptx.sig PRISM.pptx

gpg: Signature made Mon May 20 14:31:57 2013 EDT

using RSA key ID ⬛⬛⬛⬛⬛⬛⬛⬛

gpg: Good signature from “Verax”

Now I knew that Snowden, using his Verax alter ego, had signed the PowerPoint file himself. If I published the signature, all it would prove to a tech-savvy few was that a pseudonymous source had vouched for his own leak. What good would that do anyone?

In the Saturday night email, Snowden spelled it out. He had chosen to risk his freedom, he wrote, but he was not resigned to life in prison or worse. He preferred to set an example for “an entire class of potential whistleblowers” who might follow his lead. Ordinary citizens would not take impossible risks. They had to have some hope for a happy ending.

To effect this, I intend to apply for asylum (preferably somewhere with strong Internet and press freedoms, e.g. Iceland, though the strength of the reaction will determine how choosy I can be). Given how tightly the U.S. surveils diplomatic outposts (I should know, I used to work in our U.N. spying shop), I cannot risk this until you have already gone to press, as it would immediately tip our hand. It would also be futile without proof of my claims—they’d have me committed—and I have no desire to provide raw source material to a foreign government. Post publication, the source document and cryptographic signature will allow me to immediately substantiate both the truth of my claim and the danger I am in without having to give anything up. . . . Give me the bottom line: when do you expect to go to print?

Alarm gave way to vertigo. I forced myself to reread the passage slowly. Snowden planned to seek the protection of a foreign government. He would canvass diplomatic posts on an island under Chinese sovereign control. He might not have very good choices. The signature’s purpose, its only purpose, was to help him through the gates.

How could I have missed this? Poitras and I did not need the signature to know who sent us the PRISM file. Snowden wanted to prove his role in the story to someone else. That thought had never occurred to me. Confidential sources, in my experience, did not implicate themselves—irrevocably, mathematically—in a classified leak. As soon as Snowden laid it out, the strategic logic was obvious. If we did as he asked, Snowden could demonstrate that our copy of the NSA document came from him. His plea for asylum would assert a “well-founded fear of being persecuted” for an act of political dissent. The U.S. government would maintain that Snowden’s actions were criminal, not political. Under international law each nation could make that judgment for itself. The fulcrum of Snowden’s entire plan was the signature file, a few hundred characters of cryptographic text, about the length of this paragraph. And I was the one he expected to place it online for his use.

Gellman, Poitras, and the Post recognized this would make them complicit in Snowden’s flight and go beyond any journalistic role.

After some advice from WaPo’s lawyers, Gellman made it clear to Snowden he could not publish the key (and would not have, in any case, because the slide deck included information on legitimate targets he and the WaPo had no intent of publishing).

We hated the replies we sent to Snowden on May 26. We had lawyered up and it showed. “You were clear with me and I want to be equally clear with you,” I wrote. “There are a number of unwarranted assumptions in your email. My intentions and objectives are purely journalistic, and I will not tie them or time them to any other goal.” I was working hard and intended to publish, but “I cannot give you the bottom line you want.”

This led Snowden to withdraw his offer of exclusivity which — as Gellman tells the story — is what led Snowden to renew his efforts to work with Glenn Greenwald. The aftermath of that decision led to a very interesting spat between Gellman and Greenwald — to read that, you should buy the book.

To be clear, I don’t blame Snowden for planning his first releases in such a way as to optimize the chances he wouldn’t spend the rest of his life in prison. But his silence on the topic in his own account, even while he adopted the WikiLeaks line about their goal of optimizing his outcome, raises questions about any link between Harrison’s plans and Snowden’s.

The government is using Snowden as inspiration in other cases

The superseding Assange indictment is the first place I know of where the government has specifically argued that WikiLeaks’ assistance to Snowden amounted to part of a criminal conspiracy (though it is totally unsurprising and I argued that it was clear the government was going there based on what they had argued in the Joshua Schulte case).

But it’s not the first place they have argued a tie between Snowden as inspiration and further leaks.

The indictment for Daniel Everette Hale, the guy accused of sharing documents on the drone program with Jeremy Scahill, makes it clear how Hale’s relationship with Scahill blossomed just as the Snowden leaks were coming out (and this detail makes it clear he’s the one referred to in Citizenfour as another source coming forward).

15. On or about June 9, 2013, the Reporter sent HALE an email with a link to an article about Edward Snowden in an online publication. That same day. Hale texted a friend that the previous night he had been hanging out with journalists who were focused on his story. Hale wrote that the evening’s events might provide him with “life long connections with people who publish work like this.”

Hale launched a fairly aggressive (and if it weren’t in EDVA, potentially an interesting) challenge to the Espionage Act charges against him. It included (but was not limited to) a Constitutional motion to dismiss as well as a motion to dismiss for selective prosecution. After his first motions, however, both the government’s response and Hale’s reply on selective prosecution were (and remain, nine months later) sealed.

But Hale’s reply on the Constitutional motion to dismiss was not sealed. In it, he makes reference to what remains sealed in the selective prosecution filings. That reference makes it clear that the government described searching for leakers who had been inspired “by a specific individual” who — given the mention of Snowden in Hale’s indictment — has to be Snowden.

Moreover, as argued in more detail in Defendant’s Reply in support of his Motion to Dismiss for Selective or Vindictive Prosecution (filed provisionally as classified), it appears that arbitrary enforcement – one of the risks of a vague criminal prohibition – is exactly what occurred here. Specifically, the FBI repeatedly characterized its investigation in this case as an attempt to identify leakers who had been “inspired” by a specific individual – one whose activity was designed to criticize the government by shedding light on perceived illegalities on the part of the Intelligence Community. In approximately the same timeframe, other leakers reportedly divulged classified information to make the government look good – by, for example, unlawfully divulging classified information about the search for Osama Bin Laden to the makers of the film Zero Dark Thirty, resulting in two separate Inspector General investigations.3 Yet the investigation in this case was not described as a search for leakers generally, or as a search for leakers who tried to glorify the work of the Intelligence Community. Rather, it was described as a search for those who disclosed classified information because they had been “inspired” to divulge improprieties in the intelligence community.

Hale argued, then, that the only reason he got prosecuted after some delay was because the FBI had a theory about Snowden’s role in inspiring further leaks.

Judge Liam O’Grady denied both those motions (and most of Hale’s other motions), though without further reference to Snowden as an inspiration. But I’m fairly sure this is not the only case where they’re making this argument.

In a Bid to Remain Relevant, PCLOB Will Treat Carter Page as a Suspected Terrorist

It takes until paragraph 19 of this story on the decision by the Privacy and Civil Liberties Oversight Board to examine Title I FISA processes before it explains why the decision is such an obvious political game.

[PCLOB Chair Adam] Klein said the board plans only to examine counterterrorism matters, which would preclude any review of wiretap applications for Page or any investigation by the FBI of the Trump campaign.

PCLOB’s mandate is limited to counterterrorism. There were efforts to expand its mandate to include counterintelligence as part of Section 215 reauthorization that failed, so Congress has expressed an intent in recent days to limit PCLOB’s mandate to counterterrorism. Which means PCLOB has no mandate to investigate the Carter Page investigation.

But in spite of that limit on PCLOB’s mandate, PCLOB’s Republicans have decided to examine what the story calls DOJ IG’s “findings.”

Adam I. Klein, the chairman of the privacy board, said that the issues Horowitz surfaced were precisely those that the board was established to examine.

“This is at the heartland of our jurisdiction,” said Klein, a lawyer and prominent researcher of FISA and other national security laws. “The IG found systemic compliance problems. At a minimum, we have a duty to inform ourselves.”

Let’s review the posture of DOJ IG’s investigations into FISA-related functions. DOJ IG did an investigation into the Carter Page FISA applications, and found significant problems, both Woods Procedure compliance problems and lack of disclosure of material facts to the court. The way in which FBI first validated and then fact-checked an informant — long cited as a problem by defense attorneys representing counterterrorism defendants — was among the most egregious problems in the Page applications.

The Page investigation is the only finished investigation. That investigation is into a counterintelligence case, and therefore well outside of PCLOB’s mandate.

Based on the findings in that report, DOJ IG set out on an investigation into whether the problems evinced in the Page report are more systematic. As originally scoped, however, that review focused on whether the Woods Procedures–failures in which were not the most urgent or egregious aspect of the Carter Page problems–works. After three months, DOJ IG decided to issue a Management Advisor Memorandum to formally reveal its interim results that show that the Woods Procedures, and the National Security Division’s associated Accuracy Reviews, don’t work.

As a result of these findings, in December 2019, my office initiated an audit to examine more broadly the FBI’s execution of, and compliance with, its Woods Procedures relating to U.S. Persons covering the period from October 2014 to September 2019. As an initial step in our audit, over the past 2 months, we visited 8 FBI field offices of varying sizes and reviewed a judgmentally selected sample of 29 applications relating to U.S. Persons and involving both counterintelligence and counterterrorism investigations. This sample was selected from a dataset provided by the FBI that contained more than 700 applications relating to U.S. Persons submitted by those 8 field offices over a 5-year period. The proportion of counterintelligence and counterterrorism applications within our sample roughly models the ratio of the case types within that total of FBI FISA applications. Our initial review of these applications has consisted solely of determining whether the contents of the FBI’s Woods File supported statements of fact in the associated FISA application; our review did not seek to determine whether support existed elsewhere for the factual assertion in the FISA application (such as in the case file), or if relevant information had been omitted from the application. For all of the FISA applications that we have reviewed to date, the period of courtauthorized surveillance had been completed and no such surveillance was active at the time of our review.


As a result of our audit work to date and as described below, we do not have confidence that the FBI has executed its Woods Procedures in compliance with FBI policy.


During this initial review, we have not made judgments about whether the errors or concerns we identified were material. Also, we do not speculate as to whether the potential errors would have influenced the decision to file the application or the FISC’s decision to approve the FISA application. In addition, our review was limited to assessing the FBI’s execution of its Woods Procedures, which are not focused on affirming the completeness of the information in FISA applications.

The statistics provided in the MAM reveal that, with respect to Woods Procedures, Carter Page’s FISA applications were actually far better than all but one of the applications DOJ IG reviewed.

But the MAM is not a finished review and, aside from a passing reference to FBI’s failures to document informant reliability, hasn’t focused on issues known to be problematic in FISA applications targeting counterterrorism suspects.

Meanwhile, PCLOB plans to use its mandate to review counterterrorism programs to demand a list of prominent individuals targeted under FISA for the period of the DOJ IG review, 2015 to 2019.

The board will also request the number of investigations touching on prominent individuals in which the FBI sought an order from the surveillance court between 2015 and 2019. Those investigations, which the bureau defines as sensitive investigative matters, may include public officials or candidates for office, according to Justice Department guidelines.

As far as is public there have been zero prominent individuals known to be targeted under FISA. Carter Page — an unknown advisor with no institutional affiliation in DC — certainly didn’t qualify when he was targeted. (I can think of one person investigated as part of the Russian investigation who is a key influence peddler in DC who might have been targeted, but the person is not nationally known outside of political circles.)

There have, however, been key leaders in the Muslim community — who are virtually unknown outside of the Muslim or civil liberties community — targeted under FISA, per one of the most important reports to come out of the Snowden leaks (though before the period of PCLOB’s review).

• Faisal Gill, a longtime Republican Party operative and one-time candidate for public office who held a top-secret security clearance and served in the Department of Homeland Security under President George W. Bush;

• Asim Ghafoor, a prominent attorney who has represented clients in terrorism-related cases;

• Hooshang Amirahmadi, an Iranian-American professor of international relations at Rutgers University;

• Agha Saeed, a former political science professor at California State University who champions Muslim civil liberties and Palestinian rights;

• Nihad Awad, the executive director of the Council on American-Islamic Relations (CAIR), the largest Muslim civil rights organization in the country.

PCLOB probably can’t access this list because its members all have clearance, but this is where you’d start to understand the First Amendment impact of FISA on counterterrorism subjects, not by asking for a list of all the prominent people more likely to be targeted under counterintelligence.

Don’t get me wrong. If this PCLOB review were credible, I’d welcome it. If PCLOB’s mandate actually matched the scope of FISA, it could be a welcome new check on the authority.

But, as I noted in a post on some of the efforts to reform FISA legislatively, because PCLOB’s mandate does not cover some of the FISA practices of most concern, it is useless as an oversight body.

One would imagine that Carter Page, whom the Republicans think was targeted because he volunteered for the Trump campaign, would be among the people bill drafters had in mind for First Amendment protect activities.

Except he wouldn’t be included, for two reasons.

First, PCLOB’s mandate is limited to counterterrorism programs. That didn’t matter for their very good Section 215 report, because they were examining only the CDR program, which itself was limited to terrorism (and Iran).

But it did matter for the Section 702 report. In fact, PCLOB ignored some of the most problematic practices under Section 702, conducted under the guise of cybersecurity, because that’s outside their mandate! It also didn’t explore the impact of NSA’s too-broad definition of targeting under the Foreign Government certificate.

In this case, unless you expand the scope of PCLOB, then this report would only report on the targets of terrorism FISA activity, not foreign intelligence FISA activity, and so not people like Carter Page.

I was told by a key congressional negotiator that expanding PCLOB’s mandate to match FISA (that is, to include counterintelligence and foreign cyber investigations) would kill the bill. Mind you, the bill died overnight anyway, in part because Trump and his supporters want something that more directly feels like a response to the Carter Page applications.

Particularly given that FISA remains under active legislative debate, then, PCLOB would be much better served by arguing that their mandate needs to be expanded to cover all national security investigations, citing their inability to review what happened to Carter Page without overstepping their mandate.

Instead, they appear intent on overstepping their mandate.

Update: In a response to some questions from PCLOB’s press person, it appears PCLOB may misunderstand the results of DOJ IG’s interim findings. PCLOB appears to believe that DOJ IG has found material problems with the 29 files it reviewed, rather than Woods Procedures violations that it has not yet determined to be material.

As you’re aware, the most recent DoJ IG examination found problems with all 29 FISA applications it examined, many of which were for counterterrorism. Of these 29, the Board has requested only those applications that were related to counterterrorism.

The IG’s findings are troubling and suggest systematic shortcomings, with serious implications for Americans’ privacy and civil liberties.

It also appears to believe the FISA mandate to involve PCLOB would permit PCLOB to meaningfully address First Amendment issues even though it could not address many of the problems disproportionately affecting Americans.

Finally, as you may know, the House draft of the USA FREEDOM Act reauthorization bill includes a provision that directs the Board to examine whether activities protected under the First Amendment have any impact on the FISA process.  Should the bill ultimately pass Congress and be signed into law, the forum would help inform Board members on that project as well.

Adam Schiff Makes Clear FBI Is Using Section 215 Like the 2014 Exception

For months, Congress has been debating the reauthorization of Section 215 of the PATRIOT Act. The House passed a compromise bill before COVID shut-downs really halted everything in Congress, though did so in such a way as to prevent Zoe Lofgren from offering any amendments. After the Senate failed to act, the provision (and two related ones lapsed). Then, a few weeks ago, the Senate passed a version that added an amendment from Mike Lee and Patrick Leahy that strengthened the amicus to the previously passed House bill. But an amendment offered by Ron Wyden and Steve Daines failed by one vote after Tom Carper said that Pelosi had warned him its passage would gut FISA (and after Bernie Sanders and Patty Murray didn’t make it for the vote). The operative language of their amendment read,

(C) An application under paragraph (1) may not seek an order authorizing or requiring the production of internet website browsing information or internet search history information.

Zoe Lofgren and Warren Davidson tried to pass that amendment in the House. Over a weekend of heated negotiations, they limited the Wyden-Daines language to apply just to US persons.

(C) An application under paragraph (1) may not seek an order authorizing or requiring the production of internet website browsing information or internet search history information of United States persons.

At first, Wyden endorsed the Lofgren-Davidson language. Except then Adam Schiff gave Charlie Savage a statement that suggested the amendment would only prevent the government from seeking to obtain Americans’ internet information, not prevent it altogether.

But in his own statement, Mr. Schiff put forward a narrower emphasis. Stressing the continued need to investigate foreign threats, he described the compromise as banning the use of such orders “to seek to obtain” an American’s internet information.

That led Ron Wyden to withdraw his support. Leadership withdrew that amendment from the Rule.

Schiff’s ploy seems to suggest one way the government is using Section 215.

Wyden had previously asked how each of three applications for Section 215 would appear in counts:

  • An order in which an IP address used by multiple people is the target
  • An order collecting all the people who visit a particular website
  • An order collecting all the web browsing and internet searches of a single user

I’ve argued in the past that the FBI wouldn’t go to the trouble of a Section 215 order for a person who was not otherwise targeted, the last bullet. Schiff’s willingness to limit collection to foreigners is consistent with that (because targeting non-US persons has a lower probable cause level), meaning that’s not the function the government is so intent on preserving.

Which leaves Wyden’s IP address used by multiple people and a website, what I have suggested might be VPNs and WikiLeaks. Those are the applications that Schiff (and Pelosi) are going to the mat to protect.

That makes something that happened in 2014 important. That year, FISC permitted the government to remain tasked on a selector under 702 (which can only target foreigners) even after finding that Americans were using the selector, provided the US person content was purged after the fact. Except ODNI made a list of enumerated crimes — virtually all of which exploit the Dark Web — that Section 702 content could be used to prosecute. Richard Burr codified that principle when the law was reauthorized in 2017.

Schiff has invoked the same principle — allowing the FBI to target a URL or IP, and in the name of obtaining foreign intelligence, obtaining the US person activity as well. Because this is not treated as “content,” the government may not be limited to instances where the US person activity is location obscured (though it’s possible this is just about obtaining VPN traffic, and not something like WikiLeaks).

Wyden called the resulting practice (remember, this is status quo), as “dragnet surveillance.”

“It is now clear that there is no agreement with the House Intelligence Committee to enact true protections for Americans’ rights against dragnet collection of online activity, which is why I must oppose this amendment, along with the underlying bill, and urge the House to vote on the original Wyden-Daines amendment,” Wyden said.

So once again — still — the government is using a foreign targeted law to obtain leads of Americans to investigate. That, apparently, is what Pelosi considers the key part of FISA: honey pots to identify Americans to investigate.

Meanwhile, DOJ doesn’t even like the changes Lee and Leahy implemented, falsely claiming that the law — which requires DOJ to meet the standards laid out voluntarily by FBI’s response to the DOJ IG Report — does nothing to address the problems identified by the IG Report.

The Department worked closely with House leaders on both sides of the aisle to draft legislation to reauthorize three national security authorities in the U.S.A. Freedom Act while also imposing reforms to other aspects of FISA designed to address issues identified by the DOJ Inspector General. Although that legislation was approved with a large, bipartisan House majority, the Senate thereafter made significant changes that the Department opposed because they would unacceptably impair our ability to pursue terrorists and spies. We have proposed specific fixes to the most significant problems created by the changes the Senate made. Instead of addressing those issues, the House is now poised to further amend the legislation in a manner that will weaken national security tools while doing nothing to address the abuses identified by the DOJ Inspector General.

Accordingly, the Department opposes the Senate-passed bill in its current form and also opposes the Lofgren amendment in the House. Given the cumulative negative effect of these legislative changes on the Department’s ability to identify and track terrorists and spies, the Department must oppose the legislation now under consideration in the House. If passed, the Attorney General would recommend that the President veto the legislation.

Trump, meanwhile, is opposing the bill because it doesn’t go far enough.


Republicans are inventing reasons to oppose it after supporting it in March.

Back in March, Billy Barr said he could do what he needed to with EO 12333. It’s unclear how he’d coerce providers.

But Schiff’s efforts to defeat Wyden make it clear this is a function designed to identify Americans.

Update: I had thought a current vote was on FISA, but is on China sanctions, so I’ve deleted.

Ron Wyden Hints at How the Intelligence Community Hides Its Web Tracking Under Section 215

Ron Wyden had an amendment to Section 215 that would have limited the use of that provision to obtain web traffic information that fell one vote short in the Senate, partly because Nancy Pelosi whipped Tom Carper against it and partly because two Senators (Bernie Sanders and Patty Murray) didn’t get back for a vote. In an effort to resuscitate the amendment in the House under Zoe Lofgren and Warren Davidson’s leadership (which would surely pass if Section 215 got bounced back to the Senate), Ron Wyden released a letter to Ric Grenell trying to force some transparency about how the IC hides the scope of the use of Section 215 to get web search and Internet traffic information.

The letter asks Grenell to explain how Section 215 orders served on IP addresses, rather than email addresses, might get counted in transparency provisions.

How would the government apply the public reporting requirements for Section 215 to web browsing and internet searches? In this context, would the target or “unique identifier” be an IP address?

If the target or “unique identifier” is an IP address, would the government differentiate among multiple individuals using the same IP address, such as family members and roommates using the same Wi-Fi network, or could numerous users appear as a single target or “unique identifier”?

If the government were to collect web browsing information about everyone who visited a particular website, would those visitors be considered targets or “unique identifiers” for purposes of the public reporting? Would the public reporting data capture every internet user whose access to that website was collected by the government?

If the government were to collect web browsing and internet searches associated with a single user, would the public reporting requirement capture the scope of the collection? In other words, how would the public reporting requirement distinguish between the government collecting information about a single visit to a website or a single search by one person and a month or a year of a person’s internet use?

Wyden here lays out three use cases for how the IC might (one should assume does) use Section 215 to get web traffic.

  • An order in which an IP address used by multiple people is the target
  • An order collecting all the people who visit a particular website
  • An order collecting all the web browsing and internet searches of a single user

The government is required to report:

(5)the total number of orders issued pursuant to applications made under section 1861(b)(2)(B) of this title and a good faith estimate of—

(A)the number of targets of such orders; and

(B)the number of unique identifiers used to communicate information collected pursuant to such orders;

Taking each of his three scenarios, here’s what I believe the government would report.

An order in which an IP address used by multiple people is the target

In the first scenario, the government is trying to obtain everyone who “uses” a particular IP address. The scenario laid out by Wyden is a WiFi router used by family or friends, but both because the House Report prohibited such things in 2015 and because DOJ IG has raised questions about targeting everyone who uses a Friends and Family plan, I doubt that’s what the IC really does.

Rather, I suspect this is about VPNs and other servers that facilitate operational security. The government could hypothetically obtain four orders a year getting “VPNs,” requiring providers of each of the 10 major VPNs in the country to provide the IP addresses of all the incoming traffic, which would show the IP addresses of everyone who was using their location obscuring traffic.

In such a case, the targeted VPN IP addresses wouldn’t be communicating information at all. The users would get no information back. Therefore, the IC would only report the number of targets of such orders. If the “target” were defined as VPN, the number would be reported as 4 (for each of the 4 orders); if the “target” were defined as the specific VPN providers, the number of targets would be reported as 10.

The IC would entirely hide the number of individual Americans affected.

An order collecting all the people who visit a particular website

This application would seek to learn who visited a particular website. The classic case would be Inspire magazine, the AQAP propaganda. But I could also see how the IC might want to collect people who visit WikiLeaks’ submission page, or any number of sites that would offer information of interest to foreign spies (even DNI’s report on surveillance collection!). In such a use case, the government might ask not for the information provided to the user, but instead the incoming IP addresses of every request to the website. Again, this would not reflect a communication of information (and certainly not to the end user), so would not be reported under 5B.

If the targets were defined as “AQAP propaganda sites,” Inspire and all its affiliates might be reported as just one target (or might even be counted on a more generalized 215 order targeting AQAP or WikiLeaks, and so not as a unique 215 order at all).

The end users here would, again, not be counted if the collection request deliberately asked for something that did not “communicate information,” though I’m not sure precisely what technical language the government would use to accomplish this.

An order collecting all the web browsing and internet searches of a single user

This use case would ask how a 215 order targeting an individualized target (like Carter Page) shows up in transparency reports. If this were an order served on Google targeting a single account identifier for Google (say, Page’s Gmail account), the government might treat that Gmail identifier as the unique identifier, even though the government was getting information on every time this unique identifier obtained information.

Even in the criminal context, prosecutors don’t always target Google histories (for example, they did not with Joshua Schulte, and so got Google searches going back to before he joined the CIA). In the intelligence context, the FBI is given even more leeway to obtain everything, based off the logic that it’s harder to find clandestine activity.

In other words, Wyden has pointed to three use cases, all of which the IC is surely using, which existing transparency reporting requirements would entirely obscure the impact of.

The Public Record Claims that Flynn Had No Permission from Trump to Undermine US Policy in Calling Kislyak

In the last several days, part time Director of National Intelligence and full time Twitter troll Ric Grennell declassified the names of people who unmasked Mike Flynn’s name in call transcripts with Sergey Kislyak. The public record already shows the FBI did so after they discovered his calls explained why Russia had not responded as expected after Barack Obama imposed sanctions on Russia on December 28, 2016.

The press has, predictably, chased this issue as a matter of partisan game, demonstrating utter disinterest in how obviously they are being chumps in a political ploy.

Release of the list, which would be an unprecedented move, is likely to resurrect a partisan debate over an episode that had roiled the early days of Mr. Trump’s presidency and has taken on renewed urgency after the Justice Department moved to drop a criminal case against Mr. Flynn last week.

It takes enormous leaps of willful ignorance of the facts to treat this as the partisan spat that Trump wants it to be.

That’s true, for two reasons:

  • The public record shows that the Obama Administration did need to know Flynn’s identity to understand the Kislyak intercept and accorded Flynn deference as a result until such time that it appeared Flynn had acted without official sanction
  • The public record, over three years after the call, remains consistent with Mike Flynn making that call to Sergey Kislyak without permission from Trump himself, meaning the public record is consistent with Flynn acting on his own

Under FISA, the Executive Branch may not disseminate an American’s identity obtained from a FISA intercept, “unless such person’s identity is necessary to understand foreign intelligence information or assess its importance.” But if the Executive Branch needs that person’s identity to understand foreign intelligence, they can unmask the identity.

It matters that this call was made by the incoming National Security Advisor. At first, Flynn’s identity made the call look less suspicious. But within days of its discovery, Flynn’s own actions had created reason for far greater concern that the incoming NSA had made this call.

At first, the Flynn unmasking led to deference to him, albeit with concerns about sharing intelligence with (just) him

When Russia did not respond to the December 2016 sanctions, per Jim Comey’s testimony, the Intelligence Community tasked its members to learn why not.

And so the last couple days of December and the first couple days of January, all the Intelligence Community was trying to figure out, so what is going on here? Why is this — why have the Russians reacted the way they did, which confused us? And so we were all tasked to find out, do you have anything [redacted] that might reflect on this? That turned up these calls at the end of December, beginning of January.

Some days later, the FBI provided an answer: because someone had called up Russia and asked them not to escalate, and days later Russia had called up and told the same person that Vladimir Putin had not responded because of his call. Imagine the possible implications of this call without the identity. The call could reflect an amazingly powerful private individual who for some reason had the ability to make Vladimir Putin to take action against his stated interests. Or it could reflect something fairly routine. You had to know who made the call to figure out which it was.

In his testimony, Comey made it clear that, 1) they did unmask Flynn’s name but 2) the FBI issued no finalized report on this, meaning they were protecting the discovery from wider dissemination.

We did not disseminate this [redacted] in any finished intelligence, although our people judged was appropriate, for reasons that I hope are obvious, to have Mr. Flynn’s name unmasked. We kept this very close hold, and it was shared just as I described.

Sally Yates’ 302 describes how Obama responded. He stated specifically that he wanted no more follow-up information, but he did want advice on whether his White House should treat Flynn differently as a result.

After the briefing, Obama dismissed the group but asked Yates and Comey to stay behind. Obama started by saying he had “learned of the information about Flynn” and his conversation with Kislyak about sanctions. Obama specified he did not want any additional information on the matter, but was seeking information on whether the White House should be treating Flynn any differently, given the information.


Yates recalled Comey mentioning the Logan Act, but can’t recall if he specified there was an “investigation.” Comey did not talk about prosecution in the meeting. It was not clear to Yates from where the President first received the information. Yates did not recall Comey’s response to the President’s question about how to treat Flynn.

A letter Congress sent to Susan Rice quoting from her own letter to the file makes it clear that Obama explicitly stated he wanted no involvement in any law enforcement matters. He just wanted to know whether the Administration should limit how they would share classified information with Flynn during the transition.

On January 5, following a briefing by IC leadership on Russian hacking during the 2016 Presidential election, President Obama had a brief follow-on conversation with FBI Director Jim Corney and Deputy Attorney General Sally Yates in the Oval Office. Vice President Biden and I were also present.

President Obama began the conversation by stressing his continued commitment to ensuring that every aspect of this issue is handled by the Intelligence and law enforcement communities “by the book”. The President stressed that he is not asking about, initiating or instructing anything from a law enforcement perspective. He reiterated that our law enforcement team needs to proceed as it normally would by the book.

From a national security perspective, however, President Obama said he wants to be sure that, as we engage with the incoming team, we are mindful to ascertain if there is any reason that we cannot share information fully as it relates to Russia.


The President asked Comey to inform him if anything changes in the next few weeks that should affect how we share classified information with the incoming team. Comey said he would.

As to DOJ, at first Mary McCord treated this just as Republicans would want: by assuming this was just the normal pre-inauguration outreach one would expect from an incoming National Security Advisor.

It seemed logical to her that there may be some communications between an incoming administration and their foreign partners.

There are several takeaways from this record. We don’t know exactly what the transcripts say (and neither did some of the people involved), but this reaction is entirely inconsistent with Flynn saying anything to Kislyak to indicate he was operating on Trump’s orders. If he had, then Obama would not have had a concern about sharing information with Flynn and only Flynn. If it was clear Trump was involved, Obama’s concerns would be mitigated because Trump constitutionally would be entitled to this anyway. There’s no evidence Flynn made it clear he had Trump’s sanction to make these calls.

These actions also make it clear that, while the FBI responded to this as they would any counterintelligence investigation, both Obama and Rice were very careful about respecting the transition of power. The redacted passage in Rice’s letter is consistent with Obama adopting some caution, but deferring any more drastic measures unless, “anything changes in the next few weeks.”

From January 15, 2017 to the present, the public record has always been consistent with Flynn deciding to make the call on his own — and possibly acting rogue

Ten days after the Obama Administration adopted a cautious response to learning of Flynn’s calls, something did change.

The Vice President went on Face the Nation and told a journalist that he had asked Mike Flynn and Flynn denied speaking about sanctions at all.

MIKE PENCE: I talked to General Flynn about that conversation and actually was initiated on Christmas Day he had sent a text to the Russian ambassador to express not only Christmas wishes but sympathy for the loss of life in the airplane crash that took place. It was strictly coincidental that they had a conversation. They did not discuss anything having to do with the United States’ decision to expel diplomats or impose censure against Russia.

JOHN DICKERSON: So did they ever have a conversation about sanctions ever on those days or any other day?

MIKE PENCE: They did not have a discussion contemporaneous with U.S. actions on–

JOHN DICKERSON: But what about after–

MIKE PENCE: –my conversation with General Flynn. Well, look. General Flynn has been in touch with diplomatic leaders, security leaders in some 30 countries. That’s exactly what the incoming national security advisor–


MIKE PENCE: –should do. But what I can confirm, having spoken to him about it, is that those conversations that happened to occur around the time that the United States took action to expel diplomats had nothing whatsoever to do with those sanctions.

From that moment to this day, the record is consistent with Mike Flynn working without the knowledge of or prior sanction from Trump and Pence. I believe Flynn did have prior sanction from Trump, but I believe that only because I think Trump and Flynn have hidden that detail for years. But because Flynn and KT McFarland, at least, told Mueller’s prosecutors that they had no memory of consulting with Trump about what to say to Kislyak ahead of time and Trump has categorically denied it, the public record says that Flynn made the decision both to undermine the official policy of the United States and decide what policy to pursue after consulting with a few Transition aides, but not Trump himself, which was a key conclusion of this part of the Mueller Report.

Although transition officials at Mara-Lago had some concern about possible Russian reactions to the sanctions, the investigation did not identify evidence that the President-Elect asked Flynn to make any request to Kislyak.

To be clear, starting in November 2017 — ten months after Obama’s people got Flynn’s name unmasked — Flynn and KT McFarland for the first time started admitting that Flynn had consulted with Trump’s staff at Mar-a-Lago before calling Kislyak, after denying it for that time. (This passage is largely sourced to a November 17, 2017 Flynn interview and a December 22, 2017 McFarland interview.)

Flynn recalled that he chose not to communicate with Kislyak about the sanctions until he had heard from the team at Mar-a-Lago.1241 He first spoke with Michael Ledeen, 1242 a Transition Team member who advised on foreign policy and national security matters, for 20 minutes. 1243 Flynn then spoke with McFarland for almost 20 minutes to discuss what, if anything, to communicate to Kislyak about the sanctions. 1244 On that call, McFarland and Flynn discussed the sanctions, including their potential impact on the incoming Trump Administration’s foreign policy goals. 1245 McFarland and Flynn also discussed that Transition Team members in Mar-a-Lago did not want Russia to escalate the situation. 1246 They both understood that Flynn would relay a message to Kislyak in hopes of making sure the situation would not get out of hand.1247

Immediately after speaking with McFarland, Flynn called and spoke with Kislyak. 1248 Flynn discussed multiple topics with Kislyak, including the sanctions, scheduling a video teleconference between President-Elect Trump and Putin, an upcoming terrorism conference, and Russia’s views about the Middle East. 1249 With respect to the sanctions, Flynn requested that Russia not escalate the situation, not get into a “tit for tat,” and only respond to the sanctions in a reciprocal manner.1250

Multiple Transition Team members were aware that Flynn was speaking with Kislyak that day. In addition to her conversations with Bannon and Reince Priebus, at 4:43 p.m., McFarland sent an email to Transition Team members about the sanctions, informing the group that “Gen [F]lynn is talking to russian ambassador this evening.” 1251 Less than an hour later, McFarland briefed President-Elect Trump. Bannon, Priebus, Sean Spicer, and other Transition Team members were present. 1252 During the briefing, President-Elect Trump asked McFarland if the Russians did “it,” meaning the intrusions intended to influence the presidential election. 1253 McFarland said yes, and President-Elect Trump expressed doubt that it was the Russians.1254 McFarland also discussed potential Russian responses to the sanctions, and said Russia’s response would be an indicator of what the Russians wanted going forward. 1255 President-Elect Trump opined that the sanctions provided him with leverage to use with the Russians. 1256 McFarland recalled that at the end of the meeting, someone may have mentioned to President-Elect Trump that Flynn was speaking to the Russian ambassador that evening. 1257

So Flynn had the input of Michael Ledeen, McFarland, and through McFarland, the input of Transition Team members at Mar-a-Lago.

But — as I lay out in this post — the timeline laid out in Mueller’s deliberately unclear account shows no consultation between Flynn and Trump, or even McFarland and Trump, before the call. Someone may have mentioned that Flynn was making the call in a briefing Trump attended, but there’s no evidence Trump provided input on what he should say. Moreover, by the time of that briefing, Flynn appears to have already made the first call. McFarland reported to Flynn on the briefing in the same call where he told her what had transpired on his call.

1:53PM: McFarland and other Transition Team members and advisors (including Flynn, via email) discuss sanctions.

2:07PM: [Transition Team Member] Flaherty, an aide to McFarland, texts Flynn a link to a NYT article about the sanctions.

2:29PM: McFarland calls Flynn, but they don’t talk.

Shortly after 2:29PM: McFarland and Bannon discuss sanctions; according to McFarland’s clean-up interview, she may have told Bannon that Flynn would speak to Kislyak that night.

3:14PM: Flynn texts Flaherty and asks “time for a call??,” meaning McFarland. Flaherty responds that McFarland was on the phone with Tom Bossert. Flynn informs Flaherty in writing that he had a call with Kislyak coming up, using the language, “tit for tat,” that McFarland used on emails with others and that Flynn himself would use with Kislyak later that day.

Tit for tat w Russia not good. Russian AMBO reaching out to me today.

Sometime in here but the Report doesn’t tell us precisely when: Flynn talks to Michael Ledeen, KT McFarland, and then Kislyak. [my emphasis]

4:43PM: McFarland emails other transition team members saying that,  “Gen [F]lynn is talking to russian ambassador this evening.”

Before 5:45PM: McFarland briefed President-Elect Trump, Steve Bannon, Reince Priebus, Sean Spicer, and others on the sanctions. McFarland remembers that someone at the briefing may have mentioned the upcoming Kislyak call.

After the briefing: McFarland and Flynn speak by phone. Flynn tells McFarland, “that the Russian response to the sanctions was not going to be escalatory because they wanted a good relationship with the incoming Administration,” and McFarland tells Flynn about the briefing with Trump.

Moreover, the record shows that, after Flynn reported back to McFarland after Kislyak told him Russia would not respond because of the call Flynn made, he sent an email specifically designed to cover up that Kislyak had said so.

Shortly thereafter, Flynn sent a text message to McFarland summarizing his call with Kislyak from the day before, which she emailed to Kushner, Bannon, Priebus, and other Transition Team members. 1265 The text message and email did not include sanctions as one of the topics discussed with Kislyak. 1266 Flynn told the Office that he did not document his discussion of sanctions because it could be perceived as getting in the way of the Obama Administration’s foreign policy.126

Not only did Trump say, shortly after he fired Flynn, that he did not direct Flynn to discuss sanctions with Kislyak (though he said he would have directed him to do so if he wasn’t already doing it), but according to the public record, Flynn claims to have first told Trump he may have spoken about sanctions on February 6.

The week of February 6, Flynn had a one-on-one conversation with the President in the Oval Office about the negative media coverage of his contacts with Kislyak. 193 Flynn recalled that the President was upset and asked him for information on the conversations. 194 Flynn listed the specific dates on which he remembered speaking with Kislyak, but the President corrected one of the dates he listed. 195 The President asked Flynn what he and Kislyak discussed and Flynn responded that he might have talked about sanctions.196

The record also shows that, after Trump asked Reince Priebus to get KT McFarland to write a statement asserting that Trump had not spoken with Flynn before the call, she declined to do so because she didn’t know whether it had or not and John Eisenberg advised she not do so because it would make her Ambassadorial appointment look like a quid pro quo (which recently released 302s makes it look like).

Priebus called McFarland into his office to convey the President’s request that she memorialize in writing that the President did not direct Flynn to talk to Kislyak.255 McFarland told Priebus she did not know whether the President had directed Flynn to talk to Kislyak about sanctions, and she declined to say yes or no to the request.256 P

255 KTMF _ 00000048 (McFarland 2/26/ 17 Memorandum for the Record); McFarland 12/22/ 17 302, at 17.

256 KTMF _00000047 (McFarland 2/26/ 17 Memorandum_ for the Record) (“I said I did not know whether he did or didn’t, but was in Maralago the week between Christmas and New Year’s (while Flynn was on vacation in Carribean) and I was not aware of any Flynn-Trump, or Trump-Russian phone calls”); McFarland 12/22/ 17 302, at 17.

In short, even today, there is no evidence that Flynn had any permission from Trump to make this call. For over three years, Flynn and Trump have insisted he did not, which makes the significance of the intercept very different.

The public record, over three years later, is that Mike Flynn called up the country that just attacked us and — with no permission from Trump to do so — undermined the foreign policy of the United States.

So two things happened with this intercept.

At first, the fact that it was made by the incoming National Security Advisor led top DOJ officials to treat it with deferral. That is, they decided the meaning and the context was that of an incoming NSA calling foreign countries, and therefore fairly routine.

But ten days later, the transcript would look like something entirely different, the incoming NSA — who had received direct payments from Russia in the years leading up to this action — acting on his own with the Russian Ambassador. The President specifically denied having any role in the calls and fired Flynn (though said he didn’t mind the call). He went to some lengths to create a record to substantiate that he had not spoken to Flynn about it.

It would take ten months before prosecutors would have testimony (they had call records reflecting calls by March and probably had emails by August 2017) reflecting any consultation on Flynn’s part with any of his colleagues. Until they got that testimony, Flynn would have looked like had gone rogue, and decided to not only undermine Obama’s policy, but to set Trump’s policy, all on his own.

Either of those situations would justify unmasking someone’s identity. In either one of those situations, the FBI and other national security officials would have an obligation to track who was undermining the punishment for an attack by a hostile government, whether they deferred to it (in the case for the period when it seemed routine outreach) or investigated it (once it became clear the official was lying about it).

To suggest or even parrot, as Trump’s lackeys are, that this was a partisan decision suggests the United States should ignore when top national security officials appear to go rogue, undermining the current Administration without any evidence of sanction from the incoming one.