James Connell, III v. CIA

• Court: Court of Appeals for the D.C. Circuit
• Decided: August 6, 2024
• Docket: 23-5118
• Status: Published

Opinion

United States Court of Appeals FOR THE DISTRICT OF COLUMBIA CIRCUIT

Argued April 9, 2024 Decided August 6, 2024

No. 23-5118

JAMES G. CONNELL, III, APPELLANT

v.

CENTRAL INTELLIGENCE AGENCY, APPELLEE

Appeal from the United States District Court for the District of Columbia (No. 1:21-cv-00627)

Brett Max Kaufman argued the cause for appellant. With him on the briefs was Arthur B. Spitzer.

Thomas G. Pulham, Attorney, U.S. Department of Justice, argued the cause for appellee. With him on the brief were Brian M. Boynton, Principal Deputy Assistant Attorney General, and Sharon Swingle, Attorney.

Before: CHILDS and GARCIA, Circuit Judges, and GINSBURG, Senior Circuit Judge.

Opinion for the Court filed by Circuit Judge GARCIA. 2 Concurring opinion filed by Senior Circuit Judge GINSBURG. GARCIA, Circuit Judge: In 2014, the Senate Select Committee on Intelligence released a report that referred to the CIA’s “operational control” over fourteen CIA detainees transferred in September 2006 to the U.S. military base at Guantanamo Bay, Cuba. Based on that reference, a lawyer representing one of the detainees requested records from the CIA under the Freedom of Information Act about the CIA’s “operational control” at Guantanamo from September 2006 through January 2007. After searching a database of records cleared for public release or previously released, the CIA identified three documents. As to any classified or otherwise unacknowledged connection between the CIA and the topic of the request, however, the agency declared that it could neither confirm nor deny the existence of such records without revealing classified intelligence sources and methods information. The sole issue in this appeal is whether the CIA can rely on such a response to the records request here. We conclude that it can. I The Freedom of Information Act (“FOIA”) provides for disclosure of agency records to the public subject to nine exemptions. 5 U.S.C. § 552(b); Wolf v. CIA, 473 F.3d 370, 374 (D.C. Cir. 2007). As in this case, agencies sometimes respond to FOIA requests by declaring that they can neither confirm nor deny the existence of records responsive to the request. This kind of response is known as a Glomar response based on a case permitting the CIA to refuse to confirm or deny whether it had records about a ship named the Glomar Explorer. See Phillippi v. CIA, 546 F.2d 1009 (D.C. Cir. 1976). In 2009, the Senate Select Committee on Intelligence (“SSCI”) began to investigate the CIA’s post-9/11 detention 3 and interrogation program. The SSCI investigation included reviewing CIA documents. In 2012, the Committee sent drafts of the resulting report and executive summary to the Executive Branch for comment, which the CIA submitted. The Committee then requested that the executive summary be declassified, a process involving a review by the Director of National Intelligence and the CIA. The executive summary was released in redacted form in 2014. The full, unredacted report remains classified. The SSCI executive summary states that fourteen CIA detainees were transferred “to Department of Defense custody at Guantanamo Bay” in September 2006. J.A. 114.1 According to the executive summary, the detainees “remained under the operational control of the CIA.” Id. Footnote 977 cited a document titled “CIA Background Memo for CIA Director visit to Guantanamo, December [], 2006, entitled Guantanamo Bay High-Value Detainee Detention Facility.” J.A. 114 n.977. And a footnote on an earlier page cited a “September 1, 2006, Memorandum of Agreement Between the Department of Defense (DOD) and the Central Intelligence Agency (CIA) Concerning the Detention by DOD of Certain Terrorists at a Facility at Guantanamo Bay Naval Station.” J.A. 112 n.848. These unredacted references formed the basis for the records request at issue in this case. Appellant James G. Connell III is a lawyer who represents one of the fourteen detainees transferred to Guantanamo in September 2006. In May 2017, citing the SSCI executive summary’s reference to “operational control,” Connell submitted a FOIA request to the CIA for “any and all information that relates to such ‘operational control’ of the CIA over Guantanamo Bay detainees including but not limited to the document cited in the

1 Cites reflect the J.A.’s pagination, though some pages are not marked with the page number. 4 footnote 977.” J.A. 58. The CIA asked Connell to clarify the scope of his request. Connell’s response specified an interest in records that shed light on the meaning and extent of the CIA’s “operational control” over a specific part of Guantanamo called Camp 7 from September 1, 2006 to January 31, 2007. J.A. 63. Connell also listed “[b]y way of example and not limitation,” seven “possible topics,” including whether any “operational control” included facilities other than Camp 7, what organization had decisionmaking authority over Camp 7, whether CIA “operational control” ended before or after January 31, 2007, whether “operational control” involved CIA personnel, any detainee records maintained by the CIA during such a period, how other agencies could access detainees during such a period, and how the facilities transitioned from CIA to DOD “operational control.” Id. The CIA deemed this an amended FOIA request and responded in September 2020. It produced in partially redacted form the itinerary and background memo cited in footnote 977 of the SSCI executive summary, which had been previously released. The CIA stated that it could neither confirm nor deny the existence of any other responsive records. Connell filed an administrative appeal. The CIA failed to timely respond, and Connell filed his complaint in this suit in district court on March 8, 2021. In July 2021, the CIA provided a final response to Connell’s FOIA request. As CIA Information Review Officer Vanna Blaine later explained in a declaration in this case, see Blaine Decl. (J.A. 33–57), the CIA searched for records “that would reveal an unclassified or openly acknowledged association between the Agency and the subject of [Connell]’s Amended FOIA request,” id. ¶ 16 (J.A. 38–39); see also J.A. 73, in a database of “all Agency records that have been reviewed and/or compiled for potential release, or that have 5 been previously disclosed to the public,” Blaine Decl. ¶ 20 (J.A. 40–41). That search located three documents. Two were released with redactions: another version of the itinerary and background memo in footnote 977 that the CIA had previously produced, and the Memorandum of Agreement (“MOA”) between the DOD and CIA cited in footnote 848. The CIA identified a third document but withheld it in full.2 As to any other records, the CIA stated that “it could neither confirm nor deny the existence of records that may reveal a classified connection between the Agency and the subject of [Connell]’s Amended FOIA request because confirming or denying the existence or nonexistence of such records would reveal classified intelligence sources and methods information that is protected from disclosure” under FOIA Exemptions 1 and 3. Id. ¶ 26 (J.A. 43); see also J.A. 74. According to the agency, responding otherwise could “reveal sensitive details about CIA’s intelligence sources and methods and jeopardize the safety of . . . CIA employees and the employees of other agencies” or “provide adversaries with insight into the CIA’s priorities, resources, capabilities, and relationships with other agencies.” Blaine Decl. ¶ 34 (J.A. 47). The CIA moved for summary judgment, relying on Blaine’s declaration. Connell opposed, arguing that the CIA could not refuse to confirm or deny the existence or nonexistence of further responsive records in light of the

2 Connell does not challenge that withholding, nor does he attempt to use this third document in any way to support his other arguments in this case. Connell’s counsel attempted to do so for the first time at oral argument, but that came far too late. U.S. ex rel. Davis v. District of Columbia, 793 F.3d 120, 127 (D.C. Cir. 2015) (“Generally, arguments raised for the first time at oral argument are forfeited.”). 6 documents it had produced, the SSCI executive summary, and other non-CIA documents which, according to Connell, indicated that the CIA had records about its “operational control” of Camp 7 during the specified time period. The district court granted summary judgment in favor of the CIA, concluding that the CIA adequately justified its Glomar response to show entitlement to summary judgment and had not otherwise waived such a response. Connell timely appealed. II We review de novo a district court’s grant of summary judgment in favor of an agency that invokes a FOIA exemption, including when the agency has issued a Glomar response. See Montgomery v. IRS, 40 F.4th 702, 709 (D.C. Cir. 2022). Whether the CIA is entitled to summary judgment here depends on two inquiries—whether the CIA waived its ability to assert a Glomar response through official acknowledgment and, if not, whether the CIA’s justification for its Glomar response was sufficient to show it was entitled to summary judgment. We address each inquiry in turn. A “[A]n agency can waive a Glomar response through official acknowledgment,” Mobley v. CIA, 806 F.3d 568, 584 (D.C. Cir. 2015), because “[o]nce an agency has officially acknowledged that records exist, there is no value in a Glomar response. The secret is out.” Leopold v. CIA, 987 F.3d 163, 167 n.5 (D.C. Cir. 2021). To show such a waiver, a plaintiff must “identify information in the public domain that (1) matches the information requested, (2) is as specific, and (3) has ‘been made public through an official and documented disclosure.’” Knight First Amend. Inst. v. CIA, 11 F.4th 810, 815 (D.C. Cir. 7 2021) (quoting Fitzgibbon v. CIA, 911 F.2d 755, 765 (D.C. Cir. 1990)). To satisfy the first two requirements in “the Glomar context, the prior disclosure must confirm the existence or nonexistence of records responsive to the FOIA request.” Id. at 813. These requirements are exacting: “Prior disclosure of similar information does not suffice; instead, the specific information sought by the plaintiff must already be in the public domain by official disclosure.” Wolf, 473 F.3d at 378. In cases like this one, this “insistence on exactitude recognizes ‘the Government’s vital interest in information relating to national security and foreign affairs.’” Id. (quoting Pub. Citizen v. Dep’t of State, 11 F.3d 198, 203 (D.C. Cir. 1993)). Crucially for this case, the third requirement is also strict: A disclosure is “‘official’” only if made by “the agency from which the information is being sought.” Knight First Amend. Inst., 11 F.4th at 816 (quoting Frugone v. CIA, 169 F.3d 772, 774 (D.C. Cir. 1999)). Our cases have repeatedly affirmed the rationale for such a narrow approach: “While information from outside an agency may be viewed as ‘possibly erroneous,’ confirmation by the agency itself ‘would remove any lingering doubts.’” Id. at 816 (quoting Frugone, 169 F.3d at 774–75). We have also explained that “the rationale for not imputing statements by one agency to another applies with greater force, not lesser, in the intelligence context.” Id. at 818. We have applied the rule that an official acknowledgment must come from the agency whose records are sought “in various cases and contexts.” Id. at 816. For example, the FBI cannot make an official acknowledgment on behalf of the CIA. Moore v. CIA, 666 F.3d 1330, 1333–34 (D.C. Cir. 2011). Neither can the State Department, Knight First Amend. Inst., 11 F.4th at 816–18, the Office of Personnel Management, Frugone, 169 F.3d at 774–75, nor Congress, Fitzgibbon, 911 F.2d at 765–66. We have recognized one “limited exception” to this agency-specific rule: An agency is bound by 8 a disclosure “made by an authorized representative of the agency’s parent,” Knight First Amend. Inst., 11 F.4th at 816 (quoting ACLU v. CIA, 710 F.3d 422, 429 n.7 (D.C. Cir. 2013))—that is, a disclosure by another component within the same executive department or by the President as the head of the entire Executive Branch, id. at 816–17. Connell argues that the CIA waived its ability to assert a Glomar response here based on the SSCI executive summary that gave rise to his request and the documents the CIA produced in this litigation—the itinerary and background memo and the CIA-DOD MOA. Connell argues that these documents officially confirm the existence of responsive records showing a classified or otherwise unacknowledged connection between the CIA and the subject of his FOIA request. We are not persuaded. 1 Start with the SSCI executive summary and its reference to CIA “operational control.” The SSCI executive summary’s reference to CIA “operational control” is not an “official” acknowledgment: It was made by a congressional committee, not by the CIA or an authorized representative of the agency’s parent, and thus cannot be attributed to the CIA for purposes of waiver under our case law. Knight First Amend. Inst., 11 F.4th at 816–18 (noting that the CIA does not have a parent agency, but acknowledging the President or their authorized representative could qualify). In so holding, we follow a well- trodden path—indeed, as just explained, we have specifically rejected imputing disclosures by Congress to the CIA before. See, e.g., Fitzgibbon, 911 F.2d at 766; see also Knight First Amend. Inst., 11 F.4th at 816 (noting that this Court has “rejected attempts to establish an agency’s official acknowledgment based on disclosures by Congress”). 9 Connell argues that we can nonetheless consider the SSCI executive summary an “official” acknowledgement by the CIA because the summary would be seen as “similarly credible” in the eyes of “the public and U.S. adversaries,” Reply Brief 23– 24, in part because the CIA “submitted . . . comments” and participated in the report’s declassification review, J.A. 248. That approach would create a new exception to our well- established and “‘strict’” insistence that an “official” statement must be made by the agency itself; that rule has never turned on the perceived credibility of the other speaker. Leopold, 987 F.3d at 170 (quoting Moore, 666 F.3d at 1333). Nor does the CIA’s submission of comments and participation in the declassification review transmute the congressional report into a CIA one. We have rejected similar arguments that disclosures by former employees are official acknowledgments where the CIA participated in some advance review or failed to prevent the disclosure. See, e.g., Afshar v. Dep’t of State, 702 F.2d 1125, 1133–34 (D.C. Cir. 1983); Phillippi v. CIA, 655 F.2d 1325, 1330–31 (D.C. Cir. 1981). Those cases are instructive here. The CIA’s review does not make the Committee’s choice to use the phrase “operational control” an “official” disclosure attributable to the CIA. That is true at least where, as here, Connell has not pointed to anything in the record that describes the scope or content of the CIA’s comments or the extent to which the Committee implemented them, much less anything that would support attributing the particular phrase “operational control” to the CIA. Lacking support in our FOIA case law, Connell turns to two non-FOIA cases. But both are inapposite. In United States v. Zubaydah, 595 U.S. 195 (2022), a Guantanamo detainee sought to depose two former CIA contractors in ways that would reveal the existence (or not) of a CIA detention site in Poland. Id. at 199. The government moved to quash the subpoenas based on the state secrets privilege. Id. at 208. The Court concluded that the privilege applied, reasoning that even 10 though there was already public speculation that such a site existed, disclosures by the contractors could reasonably be expected to significantly harm national security interests. Id. at 207. Because the contractors played a “central role in the relevant events,” their disclosure would be “tantamount to a disclosure from the CIA itself.” Id. at 211. In a portion of the opinion joined by only two other Justices, Justice Breyer drew “some support” for this conclusion from FOIA cases, including ours, id. at 210–11, for the proposition that disclosure from an agency “insider,” id. at 208, like the contractors or the agency itself, would carry greater weight, and thus inflict more potential harm to national security interests, than mere public speculation, id. at 207–09. Connell argues that Zubaydah undermines our official acknowledgement case law, and that now statements from sufficiently credible non-agency actors (like, he says, the SSCI here) waive an agency’s rights under FOIA. This argument fails for at least two reasons. First, it is implausible to read the Court in Zubaydah as casting doubt on our FOIA case law—to the contrary, only three Justices joined the portion of the opinion discussing the FOIA cases, and even those Justices treated those cases as settled law and drew a “rough[] analog[y]” from them to support their conclusion in the different context presented in that case. Id. at 210. Second, and in any event, those Justices found the analogy helpful only because the contractors there were agency “insider[s],” id. at 208, who played a “central role in the relevant events,” id. at 211; neither characterization applies to the Committee here. Connell’s other case, Ameziane v. Obama, 699 F.3d 488 (D.C. Cir. 2012), is also not a FOIA case. Ameziane considered whether the government could adequately justify protecting certain information under a protective order governing all Guantanamo habeas litigation. Id. at 490. In holding that the case was not mooted by certain unofficial disclosures of the 11 information at issue, the court reasoned that if, as the plaintiff requested, his attorney—a government official and officer of the court—could disclose the information, that would be treated as tantamount to a similar statement by the government itself. Id. at 493. As with Zubaydah, however, Ameziane nowhere casts doubt on our FOIA precedent, and (as our description of the case shows) is both legally and factually inapposite. In short, our precedent squarely prohibits treating the Committee’s statement that the detainees remained under the CIA’s “operational control” as an official acknowledgment of the same by the CIA, and the non-FOIA cases Connell points to cast no doubt on that conclusion. 2 We turn next to the CIA-produced documents. As an initial matter, the CIA’s production of some documents in response to Connell’s FOIA request does not foreclose its ability to assert a Glomar response as to others. See Wolf, 473 F.3d at 379; see also Mobley, 806 F.3d at 583–84 (affirming CIA’s reliance on partial Glomar response). Here, the CIA explained that it identified three documents, two of which it produced, from a database of records “that have been previously disclosed to the public.” Blaine Decl. ¶ 20 (J.A. 41). That limited disclosure does not categorically prevent the CIA from invoking a Glomar response as to records showing a classified or otherwise unacknowledged connection between the CIA and the subject of Connell’s FOIA request. See Wolf, 475 F.3d at 379. And we are not persuaded that either of the two CIA-produced documents specifically matches the information protected by the CIA’s Glomar response. Neither document reveals the existence or nonexistence of records about a classified or otherwise unacknowledged connection between the CIA and the subject of Connell’s FOIA request, 12 namely, the CIA’s “operational control” over Camp 7 from September 1, 2006 to January 31, 2007. The itinerary and background memo refer to a December 21, 2006 visit by the CIA Director to Guantanamo and to the CIA transferring detainees to Guantanamo. The only reference to the CIA’s role is a description of the “CIA’s end game” as “assist[ing] DoD in any way possible in the Military Commission process, while at the same time protecting CIA equities.” J.A. 322. The MOA between DOD and the CIA “concerning the detention by DOD of certain terrorists at a facility at Guantanamo Bay Naval Station” indicates DOD, not CIA, control over detainees at Guantanamo. J.A. 307. It refers to “DoD’s detention of certain individuals,” who were “transferred to DoD and whose detention by DoD is the subject of this MOA” and states that these “DoD detainees [are] under the exclusive responsibility and control of the Secretary of Defense,” who “is solely responsible for the[ir] continued detention, release, transfer, or movement.” J.A. 307. The only reference to the CIA’s role is with respect to “coordinat[ion] with [DOD] with regard to all communications with Congress,” J.A. 313, and “on all public affairs matters and, as necessary, other US agencies,” J.A. 314. These documents do not suggest one way or the other whether the CIA has still-undisclosed records about CIA operational control over Camp 7 in the specified time period. The documents indicate only that detainees had been in CIA custody elsewhere before being transferred to DOD control at Guantanamo, and that thereafter the CIA communicated with DOD about issues relating to the detainees. Neither fact reveals the existence or nonexistence of records concerning CIA “operational control.” Indeed, Connell concedes that at least the itinerary and background memo “on its face . . . 13 doesn’t necessarily point to operational control.” Oral Argument Tr. 12:18–19; see also id. at 23:8–16. Our precedent “insist[s] on exactitude” in matching the prior disclosure with the information protected by the Glomar response. Moore, 666 F.3d at 1333 (quoting Wolf, 473 F.3d at 378). There is no such specific match here. Perhaps recognizing the problem, Connell seeks to reshape his FOIA request to fit what the CIA-disclosed documents show. Specifically, Connell argues that his FOIA request sought records showing any CIA “connection to, relationship with, and authority (or partial authority) over” Camp 7 in the specified time period. Appellant’s Brief 30. Because the CIA- produced records do show some connection between the CIA and Camp 7 in the specified time period, Connell argues, they officially acknowledged the existence of such records. But Connell’s request did not seek records of “any connection” between the CIA and Camp 7 in the specified time period. It sought records about, in the SSCI’s words, the CIA’s “operational control” of Camp 7 during that period. As explained, nothing in the documents the CIA produced discloses that the CIA had such control, much less discloses whether the CIA has other, previously undisclosed documents related to that request. Finally, Connell argues that because the CIA identified the itinerary and background memo and CIA-DOD MOA as responsive, the CIA did, in fact, confirm that the documents show “operational control.” Oral Argument Tr. 9:20–23; see also id. at 9:5–8. But Connell’s request specifically referenced the SSCI executive summary and its footnote citations. That the CIA produced these as responsive documents indicates only that the SSCI report cited them, not that the CIA was confirming that they showed “operational control” on any independent understanding of the term by the CIA. 14 Ultimately, as we have explained, what Connell needed to show was a CIA disclosure that addresses whether other CIA records exist that are responsive to the request. See Wolf, 473 F.3d at 379 (even where CIA had officially acknowledged the existence of some records pertaining to a specific person, it was required to disclose the “existence of CIA records about [him] that have been previously disclosed (but not any others)” (emphasis added)). He has not done so. B Even though the CIA has not waived its Glomar response, it must still show that it properly issued that response to be entitled to summary judgment. “An agency properly issues a Glomar response when its affidavits plausibly describe the justifications for issuing such a response, and these justifications are not substantially called into question by contrary record evidence.” Schaerr v. DOJ, 69 F.4th 924, 926 (D.C. Cir. 2023); see ACLU, 710 F.3d at 427 (“Ultimately, an agency’s justification for invoking a FOIA exemption, whether directly or in the form of a Glomar response, is sufficient if it appears logical or plausible.” (cleaned up)). 1 Recall that the CIA’s Glomar response asserted that the existence or nonexistence of records reflecting a classified or otherwise unacknowledged connection between the CIA and the subject of Connell’s FOIA request was protected from disclosure by Exemptions 1 and 3. [J.A. 43.] Because our analysis of Exemption 3 is dispositive on the issue, we do not discuss or reach Exemption 1. See Wolf, 473 F.3d at 375 (“Proper invocation of, and affidavit support for, either Exemption, standing alone, may justify the CIA’s Glomar response.”); Larson v. Dep’t of State, 565 F.3d 857, 862–63 (D.C. Cir. 2009) (similar). 15 Exemption 3 applies to “matters” that are “specifically exempted from disclosure by statute,” 5 U.S.C. § 552(b)(3), recognizing that Congress can protect particular matters from FOIA’s broad disclosure requirements. To show Exemption 3 applies, an agency must establish only “the existence of a relevant statute and the inclusion of withheld material within the statute’s coverage.” Morley v. CIA, 508 F.3d 1108, 1126 (D.C. Cir. 2007). In invoking Exemption 3 here, the CIA relied on the National Security Act, which commands the Director of National Intelligence to “protect . . . intelligence sources and methods from unauthorized disclosure.” 50 U.S.C. § 3024(i)(1). “By delegation,” the CIA Director “must do the same.” Leopold, 987 F.3d at 167. As Connell does not dispute, the National Security Act is a qualifying “withholding statute under Exemption 3.” CIA v. Sims, 471 U.S. 159, 167 (1985). The CIA’s burden was therefore to establish that disclosing whether it has other records responsive to Connell’s FOIA request would itself reveal intelligence sources and methods protected by the National Security Act. To meet that burden, the CIA relied on Blaine’s declaration. We accord “substantial weight” in the national security context to an agency’s determinations as to whether particular information is related to intelligence sources and methods or is otherwise classified. Knight First Amend. Inst., 11 F.4th at 818 (quoting Wolf, 473 F.3d at 374) (emphasis omitted); see also Sims, 471 U.S. at 179 (determinations of intelligence officials “familiar with ‘the whole picture,’ as judges are not,” as to whether information relates to intelligence sources and methods “are worthy of great deference given the magnitude of the national security interests and potential risks at stake”). We “do not require a degree of specificity that would itself possibly ‘compromise intelligence methods and sources.’” Knight First Amend. Inst., 11 F.4th at 821 (quoting Mil. Audit Project v. Casey, 656 F.2d 724, 751 (D.C. Cir. 1981)). 16 Here, the CIA’s declaration explains that a “defining characteristic of the CIA’s intelligence activities is that they are carried out through clandestine means, and therefore they must remain secret in order to be effective.” Blaine Decl. ¶ 23 (J.A. 41–42). Accordingly, “the CIA generally does not confirm or deny the existence, or disclose the target, of specific intelligence collection activities of the operations it conducts or supports.” Id. ¶ 44 (J.A. 52). Turning to the specific request here, the declaration states that “acknowledging the existence or nonexistence of records reflecting a classified or otherwise unacknowledged connection to the CIA in this matter would reveal information that concerns intelligence sources and methods, which the National Security Act is designed to protect.” Id. ¶ 39 (J.A. 49); see also id. ¶ 16 (J.A. 39) (defining scope of Glomar response as to “any records that may reveal a classified connection between the Agency and the subject of Plaintiff’s Amended FOIA Request”). The declaration also states that “confirmation or denial of the existence or nonexistence of such records would reveal sensitive information about the CIA’s intelligence interests, personnel, capabilities, authorities, and resources.” Id. ¶ 34 (J.A. 47). A Glomar response was further needed to avoid “reveal[ing] sensitive details about CIA’s intelligence sources and methods and jeopardiz[ing] the safety of the CIA employees and the employees of other agencies” and to avoid “provid[ing] adversaries with insight into the CIA’s priorities, resources, capabilities, and relationships with other agencies.” Id. Though the CIA could arguably have provided additional detail as to what intelligence sources and methods would be revealed here, the CIA met its burden of justifying its Glomar response. It is plausible that revealing the existence or nonexistence of records of a classified or otherwise unacknowledged connection between the CIA and the subject of Connell’s FOIA request could reveal intelligence sources and methods information. It is also plausible that stating 17 whether the CIA has records about its operational control (or partial control or utter lack thereof) over Camp 7 would reveal information about the CIA’s “relationships with other agencies,” including DOD, or information about the CIA’s “priorities,” “capabilities,” and “resources.” Id. ¶ 34 (J.A. 47). Furthermore, as we have recognized, protecting intelligence sources and methods information under the National Security Act allows the CIA to withhold even “superficially innocuous information on the ground that it might enable an observer to discover” an intelligence source or method. Sims, 471 U.S. at 178. Because “bits and pieces of data may aid in piecing together bits of other information even when the individual piece is not of obvious importance itself,” the CIA’s protection of intelligence sources and methods can cover “what may seem trivial to the uninformed,” but “may appear of great moment to one who has a broad view of the scene” and can “put the questioned item of information in its proper context.” Id. (cleaned up). The CIA’s declaration here makes precisely this point. See Blaine Decl. ¶ 32 (J.A. 46) (“Terrorist organizations, foreign intelligence services, and other hostile groups . . . search continually for information regarding the activities of the CIA and are able to gather information from a myriad of sources, analyze this information, and devise ways to defeat CIA activities from seemingly disparate pieces of information.”). Connell does not dispute any of those points. He does not argue that the declaration lacks sufficient specificity about which intelligence sources and methods would be revealed or how, nor does he dispute that the CIA’s explanation for its Glomar response was otherwise sufficiently logical or plausible on its own terms. 18 2 Connell instead argues that the CIA cannot plausibly claim that it has no further documents in light of “contrary record evidence,” Schaerr, 69 F.4th at 926—the documents the CIA produced and disclosures from other government entities. See Appellant’s Brief 31 (“If the record establishes that it is not logical or plausible that the agency has no such records, the CIA must acknowledge that it does, in fact, have them . . . .”). In other words, he argues that there is nothing for the CIA’s Glomar response to protect because based on already-public information it is obvious, at least to him, that the CIA does have other documents responsive to his FOIA request. Connell bases this argument not only on the two CIA documents the agency produced, but also on an array of non- CIA materials, such as statements from various parties and a judge in military commission proceedings. See also infra at note 4. Because Connell’s argument turns primarily on the non-CIA documents, we address those first. Connell’s key legal argument in asking us to focus on these materials is that even if statements that are not from the CIA or an authorized representative of its parent cannot qualify as official acknowledgments under our waiver cases, they are still relevant evidence to consider when assessing whether it is plausible for the CIA to state that confirming or denying the existence of responsive records would reveal something that is not already public. See Appellant’s Brief 32–35. We reject that argument, as agreeing with Connell would amount to an end-run around our official acknowledgment cases and contravene both their logic and results. As detailed above, the rationale underlying our official acknowledgment cases, as applied to Glomar responses, is that confirmation that an agency has responsive records (or not) by the agency itself is different from statements to that effect by 19 other sources—even trusted government sources—because confirmation by the agency itself removes “any lingering doubts” on the issue. Knight First Amend. Inst., 11 F.4th at 816; see Frugone, 169 F.3d at 774–75. For that reason, “other agencies of the Executive Branch” cannot “obligate agencies with responsibility in [the national security] sphere,” like the CIA here, to reveal protected intelligence information. Frugone, 169 F.3d at 775. The upshot for present purposes is that when an agency has not officially acknowledged whether it has records responsive to a FOIA request, we cannot assume the answer to that question based on “public speculation, no matter how widespread,” Wolf, 473 F.3d at 378; see Casey, 656 F.2d at 745 (“We cannot assume, as the appellants would have us, that the CIA has nothing left to hide.”). Yet that is exactly what Connell’s theory would have us do: assume the CIA has responsive documents based on non-CIA statements. To take just one concrete example from our case law, we held in Frugone that the CIA could plausibly maintain a Glomar response to a request for an individual’s personnel records even where the Office of Personnel Management had stated in no uncertain terms that such records were “maintained by the CIA.” 169 F.3d at 773. Because the statement was not made by the CIA, and the CIA explained why Exemptions 1 and 3 justified a Glomar response, we upheld that response. Id. On Connell’s theory, however, the plaintiff there could have sidestepped that holding by arguing that even if that non-CIA statement could not amount to an official acknowledgement, that statement (from an undoubtedly trustworthy speaker) nonetheless rendered it implausible for the CIA to assert that it might not have such records and that protected information would be revealed if the CIA itself confirmed or denied the records’ existence. Connell’s approach would undermine not 20 just Frugone but decades of settled precedent, and we decline to endorse it.3 Accordingly, the non-CIA statements on which Connell seeks to rely could not render illogical or implausible the CIA’s assertion that it would reveal protected intelligence information to confirm or deny the existence or nonexistence of records

3 Connell identifies one case that arguably relied on nonofficial statements in the way he urges: Florez v. CIA, 829 F.3d 178 (2d Cir. 2016). We do not find that out-of-circuit case persuasive. In Florez, a divided Second Circuit panel addressed whether FBI disclosures that post-dated the district court’s summary judgment opinion required remand for the district court to reconsider whether the CIA was entitled to summary judgment on its Glomar response. Id. at 180–81. The majority did not find that FBI disclosures rendered the CIA’s Glomar response implausible, but it concluded that the disclosures were “relevant” and remanded for the district court to consider in the first instance. Id. at 186–87. The dissent, however, reasoned that FBI disclosures that did not mention the CIA at all, let alone the existence of CIA records responsive to the FOIA request at issue, could not affect the adequacy of the CIA’s justification that its Glomar response was necessary to avoid unauthorized disclosures of intelligence sources and methods information under Exemptions 1 and 3. Id. at 191–95 (Livingston, C.J., dissenting). Further, the dissent pointed out—correctly, in our view—that “[t]he majority’s error in deeming these irrelevant documents germane thus appears to invite by the back door what the official acknowledgment doctrine prohibits at the front.” Id. at 196. To the extent the Florez majority characterized the FBI disclosures as “relevant” to the CIA’s justification for its Glomar response, we find the dissent’s explanation of how this improperly circumvents the official acknowledgment doctrine persuasive and in accord with this court’s case law, at least as applied to our analysis of Connell’s argument here. 21 showing a classified or unacknowledged connection between the CIA and the subject of Connell’s request.4 Connell also relies heavily on our 2013 decision in ACLU v. CIA, but that case only confirms our conclusion. The FOIA request there sought CIA records regarding the United States’ use of drone strikes, and the CIA issued a Glomar response “on the ground that it was necessary to keep secret whether the CIA itself was involved in, or interested in, such strikes.” 710 F.3d at 428 (emphasis omitted). The question was therefore whether it was logical or plausible “for the CIA to contend that it would reveal something not already officially acknowledged to say that the Agency ‘at least has an intelligence interest’ in [drone] strikes.” Id. at 429. The problem for the CIA there was that repeated official statements—from the President, his counterterrorism advisor, and the CIA Director—revealed that the United States used drone strikes. Id. at 429–30. As a result of those official statements, we held that it “strains credulity” for the CIA—“an agency charged with gathering intelligence

4 These materials include the SSCI executive summary’s footnote reference to a site daily report and cable (which, we note, does not correspond to the dates of Connell’s FOIA request and is thus not responsive); November 2006 interagency meeting materials produced by the Office of the Director of National Intelligence in response to a separate FOIA request, which show, at most, inter- agency communication related to Camp 7; testimony from Camp 7’s commander, which never identifies the CIA; a military judge’s decision and factfinding in a case concerning a Guantanamo detainee, which does not correspond to the dates of Connell’s FOIA request; the protective order in Connell’s client’s case before the military commission; and a government response to motions to compel discovery related to the CIA’s role at Camp 7. Although we do not resolve the question, we note that it is far from clear that these materials are properly read to undermine the CIA’s justification for its Glomar response even if they were accorded the same status as statements from the CIA itself. 22 affecting the national security”—to maintain that it did not at least have an “intelligence interest” in that subject. Id. at 430. ACLU indicates that even when official statements do not precisely match the secret protected by the Glomar response as required for waiver through official acknowledgment, such statements can render a Glomar response insufficiently logical or plausible if they directly undermine the justification given for that response. But the statements in ACLU were, crucially, official. Everything our cases have said about the special import of official statements (those from the agency or an authorized representative of the agency’s parent) was therefore not in tension with our rationale there. ACLU did not turn in any respect on the type of nonofficial statements Connell asks us to consider here. And, unlike in ACLU, the official statements Connell identifies here do not undermine the CIA’s justification for its Glomar response. As discussed above, the two CIA-produced documents indicate that detainees had been in CIA custody elsewhere before being transferred to DOD control at Guantanamo, and that the CIA communicated with DOD about issues relating to the detainees. But records revealing prior custody and ongoing inter-agency communication do not make it implausible that the CIA’s confirmation of the existence or nonexistence of records showing a classified or unacknowledged connection between the CIA and “operational control” over Camp 7 in the specified time period would reveal intelligence sources and methods protected by the National Security Act or information about the CIA’s relationships with other agencies, priorities, or resources. * * * In sum, the CIA did not waive its ability to assert a Glomar response through official acknowledgment. On Connell’s articulation of the topic of his FOIA request, neither the SSCI 23 executive summary nor the CIA-produced documents support waiver. Further, though its declaration could have provided more detail, the CIA’s justification for its Glomar response was logical and plausible. Connell’s “contrary record evidence” does not indicate otherwise. III For the foregoing reasons, the judgment of the district court is affirmed. So ordered. GINSBURG, Senior Circuit Judge, concurring:

I concur fully in the opinion of the Court. I write sepa- rately to make two additional points.

First, Connell’s reliance on the Second Circuit’s decision in Florez v. CIA, 829 F.3d 178 (2016), is misplaced. In that case, the Second Circuit deemed Glomar responses “justified only in ‘unusual circumstances, and only by a particularly per- suasive affidavit.’” Id. at 182 (quoting N.Y. Times v. Dep’t of Just., 756 F.3d 100, 122 (2d Cir. 2014)). The Second Circuit borrowed that wording from our opinion in ACLU v. CIA, 710 F.3d 422 (2013), but it misread that opinion. There we ex- plained that when an agency must disclose the existence of a document requested under the FOIA, but believes the content of the document is exempt from disclosure, it may issue either a “no number, no list” response or a “Vaughn index.” * See id. at 432–35. Observing that “there is a material difference be- tween a ‘no number, no list’ response and a Glomar response,” we held that a “no number, no list” response, unlike a Glomar response, is justified under the FOIA only “in unusual circum- stances, and only by a particularly persuasive affidavit.” Id. at 433. We made clear that a Glomar response, unlike a “no num- ber, no list” response, is to be judged under “the same general exemption review standards established in non-Glomar cases.” Id. at 426 (quoting Wolf v. CIA, 473 F.3d 370, 374 (D.C. Cir. 2007)). That is the standard the FOIA requires for a Glomar

* As we have previously explained, a “Vaughn index” is a filing that lists the documents an agency has withheld and explains why each is subject to a particular FOIA exemption. See, e.g., DiBacco v. U.S. Army, 795 F.3d 178, 186 n.2 (D.C. Cir. 2015), ACLU, 710 F.3d at 432–33; Judicial Watch, Inc. v. FDA, 449 F.3d 141, 145–46 (D.C. Cir. 2006). A “no number, no list” response is a filing in which an agency admits it has responsive documents but declines to enumerate or describe them at all. See, e.g., ACLU, 710 F.3d at 432–33; N.Y. Times, 756 F.3d at 105. 2 response, as we reiterated three terms ago in Knight First Amendment Institute at Columbia University v. CIA, 11 F.4th 810, 819 (2021).

Second, a litigant that challenges an agency’s justification for a Glomar response by pointing to publicly available infor- mation related to the subject of the documents it seeks would do well to remember that the touchstone of FOIA Exemption 1 is whether the document in question “‘pertains to’ either ‘intel- ligence activities’ or ‘intelligence sources or methods’” and “‘could reasonably be expected to cause identifiable or describ- able damage to the national security’ if disclosed.” Knight Inst., 11 F.4th at 813 (quoting Exec. Order No. 13,526, § 4(c), 75 Fed. Reg. 707, 709 (2009)). It is for this very reason that our past decisions “have unequivocally recognized that the fact that information resides in the public domain does not elimi- nate the possibility that further disclosures can cause harm to intelligence sources, methods[,] and operations.” Fitzgibbon v. CIA, 911 F.2d 755, 766 (D.C. Cir. 1990). We give substantial weight to the CIA’s judgment regarding that possibility, for as we have often repeated, “[t]he assessment of harm to intelli- gence sources, methods[,] and operations is entrusted to the Director of Central Intelligence, not to the courts.” ACLU v. U.S. Dep’t of Def., 628 F.3d 612, 624 (2011) (first alteration in original); Assassination Archives & Rsch. Ctr. v. CIA, 334 F.3d 55, 58 (2003); Students Against Genocide v. Dep’t of State, 257 F.3d 828, 835 (2001); Fitzgibbon, 911 F.2d at 766.