Page 2PROFESSOR DEBORAH A. DEMOTT; SCHOLARS OF FEDERAL COURTS; PROFESSORS OF LEGAL HISTORY; FORMER MILITARY LEADERS AND LAWYERS

Amicus Supporting Appellee.

Appeal from the United States District Court for the Eastern District of Virginia, at Alexandria. Leonie M. Brinkema, District Judge. (1:08-cv-00827-LMB-JFA)

Argued: September 9, 2025 Decided: March 12, 2026

Before THACKER and QUATTLEBAUM, Circuit Judges, and FLOYD, Senior Circuit Judge.

Affirmed in part, vacated in part , and remanded with instructions by published opinion. Senior Judge Floyd wrote the opinion in which Judge Thacker joined. Judge Quattlebaum wrote a dissenting opinion.

ARGUED: John Frederick O'Connor, Jr., STEPTOE LLP, Washington, D.C., for Appellant. Baher Azmy, CENTER FOR CONSTITUTIONAL RIGHTS, New York, New York; Michael Francis Buchanan, PATTERSON BELKNAP WEBB & TYLER LLP, New York, New York; Michael Shih, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C., for Appellees. ON BRIEF: Linda C. Bailey, Joseph T. McClure, STEPTOE LLP, Washington, D.C.; Nina J. Ginsberg, GREENSPUN SHAPIRO GINSBERG & YANG, P.C., Fairfax, Virginia, for Appellant. Yaakov M. Roth, Acting Assistant Attorney General, Sharon Swingle, Civil Division, UNITED STATES DEPARTMENT OF JUSTICE, Washington, D.C.; Erik S. Siebert, United States Attorney, OFFICE OF THE UNITED STATES ATTORNEY, Richmond, Virginia, for Appellee United States. Katherine Gallagher, CENTER FOR CONSTITUTIONAL RIGHTS, New York, New York; Andrew Haddad, W. Scott Kim, Alexandra Mahler-Haug, James Mayer, PATTERSON BELKNAP WEBB & TYLER LLP, New York, New York, for PlaintiffAppellees. Agnieszka M. Fryszman, Nicholas Jacques, Washington, D.C., Benjamin F. Jackson, New York, New York, Adnan Toric, COHEN MILSTEIN SELLERS & TOLL PLLC, Philadelphia, Pennsylvania, for Amicus Professor Deborah A. DeMott. Lawrence S. Lustberg, Madhulika Murali, GIBBONS P.C., Newark, New Jersey, for Amici Scholars of Federal Courts. Tyler R. Giannini, Emily A. Ray, Jonathan B. Tucker, Human Rights Entrepreneurs Clinic, HARVARD LAW SCHOOL, Cambridge, Massachusetts, for Amici Professors of Legal History Nikolas Bowie, William R. Casto, Martin S. Flaherty, Eliga H.

Page 3Gould, Stanley N. Katz, Samuel Moyn, and Anne-Marie Slaughter. Avidan Y. Cover, CASE WESTERN RESERVE UNIVERSITY SCHOOL OF LAW, Cleveland, Ohio; Jennifer B. Condon, Center for Social Justice, SETON HALL LAW SCHOOL, Newark, New Jersey, for Amici Former Military Leaders and Lawyers.

Page 4FLOYD, Senior Circuit Judge:

Over twenty years ago, members of the United States military and military contractors committed horrific abuses on detainees at Abu Ghraib Prison during the Iraq War. Plaintiffs-Appellees are Iraqi citizens who allege they were tortured while detained and filed suit for civil damages in 2008. Now, after seventeen-and-a-half years of litigation, a jury reached a verdict finding Appellant CACI Premier Technology, Inc. (CACI) liable for conspiracy to commit torture and conspiracy to commit cruel, inhuman, and degrading treatment (CIDT) under the Alien Tort Statute (ATS). After reviewing the record and hearing oral argument, we affirm the jury’s verdict.

This is the sixth appeal in this case, and we have previously described the factual background of this case as well as the extensive procedural history. Al Shimari v. CACI Premier Tech., Inc. ( Al Shimari III ), 758 F.3d 516 (4th Cir. 2014); Al Shimari v. CACI Premier Tech., Inc. ( Al Shimari IV ), 840 F.3d 147 (4th Cir. 2016). Accordingly, we now provide a relatively brief overview of the relevant factual and procedural background.

The United States invaded Iraq in early 2003, overthrowing the existing Iraqi government. In May 2003, the United States established a temporary governing body, the Coalition Provisional Authority (CPA), led by State Department official L. Paul Bremer.

Page 5Bremer and the CPA had plenary authority until June 2004, when the CPA began to transition power to the Iraqi Interim Government.

Under the CPA regime, the U.S. military controlled the now-infamous detention center Abu Ghraib, in which the United States detained thousands of Iraqis, including individuals suspected of participating in the rising insurgency. Due to a shortage of military interrogators, the U.S. government entered into a contract with CACI, a Virginia corporation, to provide interrogation services at Abu Ghraib, which contract was issued by a contracting officer with the Department of the Interior in Arizona. The contract required CACI interrogators to obtain security clearances from the Department of Defense.

CACI hired interrogators from its headquarters. Communications reveal that CACI had concerns that many of the personnel it hired and sent to Iraq were not adequately qualified to perform interrogations. Nevertheless, CACI interrogators arrived at Abu Ghraib in September 2003. The agreements between CACI and the military, organization charts, and Army memoranda, among other documents, all called for CACI’s interrogators to be integrated into U.S. military chain of command, reporting to and under the direction and control of military personnel. But later-gathered evidence reflects a command vacuum existed, wherein the military did not adequately supervise CACI personnel and CACI employees directed members of the military police (MPs) regarding interrogation tactics, in contravention of the formal structures in place.

Between October and December 2003, CACI personnel and members of the military abused detainees at Abu Ghraib. CACI interrogators instructed MPs to abuse detainees to elicit useful information from the detainees during formal interrogations. CACI received Page 6reports of detainee abuse at its Virginia headquarters in October 2003 and took no affirmative action to intervene to prevent further misconduct. After allegations of this abuse surfaced, the U.S. Department of Defense commissioned an official investigation into Abu Ghraib. Major Generals George Fay and Antonio Taguba were appointed to lead the investigation, resulting in each publishing a report with factual findings and conclusions.

Plaintiffs-Appellees Suhail Najim Abdullah Al Shimari, Asa’ad Hamza Hanfoosh Al-Zuba'e, and Salah Hasan Nusaif Jasim Al-Ejaili (collectively, 'Plaintiffs') were detained at Abu Ghraib and were abused by MPs while in detention. The abuse Plaintiffs experienced included sexual assault, forced nudity, dog threats and attacks, prolonged stress positions, and threats to Plaintiffs and their family members.

Plaintiffs’ theory of liability to CACI, despite not alleging that CACI interrogators physically abused Plaintiffs, is that CACI employees conspired with MPs to abuse detainees to 'soften [them] up' so that they would be more responsive during later interrogations. Resp. Br. at 1.

B.

In 2008, Suhail Najim Abdullah Al Shimari, Taha Yaseen Arraq Rashid, Asa’ad Hamza Hanfoosh Al-Zuba’e, and Salah Hasan Nusaif Jasim Al-Ejaili brought this case alleging twenty causes of action under both the ATS and state tort law. Along the way, Plaintiffs voluntarily dismissed their tort claims, leaving only their claims under the ATS. In their ATS causes of actions, Plaintiffs allege that CACI’s actions constituted torture, Page 7CIDT, and war crimes in violation of the law of nations. Plaintiffs allege these violations under theories of direct liability, aiding and abetting, and conspiracy. In 2018, CACI filed a third-party complaint against John Does 1–60 and the United States. The district court dismissed Plaintiffs’ direct liability claims, dismissed Taha Yaseen Arraq Rashid from the case, dismissed CACI’s third-party complaint against the United States, and severed and stayed CACI's third-party claims against John Does 1-60 pending resolution of the primary case between Plaintiffs and CACI.

In April 2024, Plaintiffs proceeded to trial on their claims for conspiracy and aiding and abetting torture, CIDT, and war crimes. During that trial, Plaintiffs withdrew their war-crimes-related claims. The jury could not reach a verdict, and the district court declared a mistrial. Plaintiffs moved for a new trial, which the district court granted.

In November 2024, the second trial commenced on Plaintiffs’ claims for torture and CIDT on the secondary liability theories of conspiracy and aiding and abetting. During this trial, the district court granted in part CACI’s motion under Rule 50 of the Federal Rules of Civil Procedure on the aiding and abetting claims. The case went to the jury only on the surviving two claims: conspiracy to commit torture and conspiracy to commit CIDT. The jury returned a verdict for Plaintiffs on both claims and awarded each Plaintiff $3 million in compensatory damages and $11 million in punitive damages.

CACI timely appeals. It questions the propriety of proceeding under the ATS and raises threshold issues of justiciability, immunity, preemption, and state secrets. It further argues it was entitled to judgment as a matter of law or a new trial on Plaintiffs’ conspiracy claims and its borrowed servant defense. Finally, CACI submits that it was entitled to Page 8remittitur or a new trial on damages. CACI also appeals the judgment granted to the United States on CACI’s third-party complaint.

We first address the issues surrounding the application of the Alien Tort Statute. We will then consider CACI’s threshold issues and its third-party claims against the United States, before considering the merits of whether CACI was entitled to judgment as a matter of law and whether the jury’s verdict was supported by the evidence. Lastly, we will review the district court’s decision on damages.

CACI appeals the district court’s denial of its motion for judgment as a matter of law or alternatively, for a new trial under Rules 50 and 59 of the Federal Rules of Civil Procedure. We review a district court’s denial of a motion for judgment as a matter of law de novo , “view[ing] the evidence in the light most favorable to the prevailing party[ and] assessing whether there was a legally sufficient evidentiary basis for a reasonable jury to find for that party.” FDIC v. Bakkebo , 506 F.3d 286, 294 (4th Cir. 2007); Fed. R. Civ. P. 50(b). We review denial of a motion for a new trial “for clear abuse of discretion and will not reverse absent exceptional circumstances.” Dennis v. Columbia Colleton Med. Ctr., Inc. , 290 F.3d 639, 650 (4th Cir. 2002).

CACI first raises two issues attacking the applicability of the ATS: (1) whether this case presents an improper application of the statute to conduct occurring outside the Page 9territorial jurisdiction of the United States, and (2) whether the district court erred in recognizing Plaintiffs' causes of action under the ATS. Before turning to CACI's arguments, we revisit the development of ATS jurisprudence and determine what showing is needed for us to maintain jurisdiction under this statute.

The Alien Tort Statute provides that '[t]he district courts shall have original jurisdiction of any civil action by an alien for a tort only, committed in violation of the law of nations or a treaty of the United States.'²⁸ U.S.C. § 1350. The statute was first introduced in the Judiciary Act of 1789; at the time, the “three specific offenses against the law of nations” likely contemplated by the First Congress were “violation of safe conducts, infringement of the rights of ambassadors, and piracy.' Sosa v. Alvarez-Machain , 542 U.S. 692, 715 (2004). Accordingly, the “jurisdiction” conferred by the inclusion of the ATS in the Judiciary Act “was originally understood to be available to enforce a small number of international norms that a federal court could properly recognize as within the common law enforceable without further statutory authority.” Id. at 729. Indeed, a 1795 letter written by Attorney General William Bradford, “as well as the two early federal precedents discussing the ATS, point to a prevalent assumption that Congress did not intend the ATS to sit on the shelf until some future time when it might enact further legislation.” Id. at 724.

Our contemporary understanding of the ATS is that it “is a jurisdictional statute creating no new causes of action,” but “that the statute was intended to have practical effect Page 10the moment it became law.” Id. To harmonize this incongruity, the Sosa Court directed lower courts to “require any claim based on the present-day law of nations to rest on a norm of international character accepted by the civilized world and defined with a specificity comparable to the features of the 18th-century paradigms we have recognized.' Id. at 725. In other words, the ATS “allows federal courts to recognize certain causes of action based on sufficiently definite norms of international law.” Kiobel v. Royal Dutch Petrol. Co. , 569 U.S. 108, 116 (2013).

To determine whether a cause of action is cognizable under the ATS, courts assess claims “against the current state of international law, looking to those sources we have long, albeit cautiously, recognized.' Sosa , 542 U.S. at 733. Sosa suggests that these sources include treaties, “controlling executive or legislative act[s,] or judicial decision[s],” and, in their absence, Sosa allows courts to refer “to the customs and usages of civilized nations; and, as evidence of these, to the works of jurists and commentators, who by years of labor, research and experience, have made themselves particularly aware of the subjects of which they treat.' Id. at 734 (quoting The Paquete Habana , 175 U.S. 677, 700 (1900)).

After Sosa , the Supreme Court has provided two further directives to courts considering ATS cases relevant here. First, “the presumption against extraterritoriality applies to claims under the ATS.' Kiobel , 569 U.S. at 124. Second, and relatedly, “plaintiffs must establish that ‘the conduct relevant to the statute’s focus occurred in the United States'' for courts to conclude that 'the case involves a permissible domestic application even if other conduct occurred abroad.' Nestlé USA, Inc. v. Doe , 593 U.S. 628, 633 (2021) (quoting RJR Nabisco, Inc. v. European Cmty. , 579 U.S. 325, 337 (2016)).

With this background, we now consider CACI’s first challenge to our jurisdiction under the ATS in this case. CACI argues that the district court lacked subject matter jurisdiction because the claims are impermissibly extraterritorial. According to CACI, all conduct relevant to the focus of the ATS occurred in Iraq. Even if we accept this premise, we find that applying the ATS to conduct in Iraq does not offend the presumption against extraterritoriality in this case under two alternative theories. First, conduct in Iraq's detention centers at the end of 2003 may properly be considered within the “territorial jurisdiction” of the United States. See Kiobel , 569 U.S. at 121. Second and alternatively, conduct in Iraq at the end of 2003 occurred “beyond the territorial jurisdiction of the United States or any other country,” and the nature of the conduct at issue allows for jurisdiction. See id. We also find that sufficient conduct occurred within the geographic borders of the United States to displace the presumption. We examine each theory in turn.

Although it is well established that the ATS does not reach purely extraterritorial conduct, the presumption against extraterritoriality “has no application to the operation of [a] statute . . . within 'the territorial jurisdiction' of the United States.' Rasul v. Bush , 542 U.S. 466, 480 (2004) (quoting Foley Bros., Inc. v. Filardo , 336 U.S. 281, 285 (1949)) (finding that for the habeas corpus statute, the presumption against extraterritoriality does not apply to Guantanamo Bay Naval Base). Plaintiffs and amici contend that Abu Ghraib, like Guantanamo Bay, may be considered within the territorial jurisdiction of the United Page 12States because the United States “exercised plenary legal and political control” over the site. Resp. Br. at 31; Br. of Amici Curiae Scholars of Fed. Cts. in Supp. of Pls.-Appellees and Affirmance at 6-18 (hereinafter Fed. Cts. Br.). We revisit the circumstances surrounding Plaintiffs' detention and conclude that the United States had 'complete jurisdiction and control” over Abu Ghraib in late 2003.¹ See Rasul , 542 U.S. at 480 .

As previously discussed, the United States’ occupation of Iraq began in May 2003, when the United States deposed the existing Iraqi government, created the CPA to temporarily replace it, and installed Bremer to head the CPA. In May 2003, Bremer promulgated CPA Order 2, which dissolved certain Iraqi government offices and military and paramilitary organizations and made the CPA the custodian of any assets, records, and data previously maintained by those offices and organizations. CPA Order No. 2 (May 23, 2003), reprinted by Univ. of N. Tex. Libr., https://govinfo.library.unt.edu/cpairaq/regulations/20030823_CPAORD_2_Dissolution_of_Entities_with_Annex_A.pdf [https://perma.cc/7TH5-TXD2]. In June, Bremer promulgated CPA Order 10, which vested “[f]ull authority and control over all detention and prison facilities” in the Ministry Page 13of Justice, and directed the Ministry of Justice, “under the authority, direction and control of the Administrator of the CPA,” to “prescribe any administrative procedures necessary to ensure a properly coordinated transfer of the detention and prison facilities.” CPA Order No. 10 (June 8, 2003), reprinted by Univ. of N. Tex. Libr., https://govinfo.library.unt.edu/cpairaq/regulations/20030605_CPAORD10_Management_of_Detention_and_Prison_Faciliti es.pdf [https://perma.cc/X3MP-CL4E]. Accordingly, the detention center at Abu Ghraib was specifically under the authority and control of the Ministry of Justice and Bremer beginning in June 2003. The CPA maintained authority until June 28, 2004. CPA Order No. 100 (June 28, 2004), reprinted by Univ. of N. Tex. Libr., https://govinfo.library.unt.edu/cpairaq/regulations/20040628_CPAORD_100_Transition_of_Laws__Regulations__Orders_ _and_Directives.pdf [https://perma.cc/8LL4-VK82].

These facts demonstrate that the United States, through the CPA it established and the administrator it appointed, had 'complete jurisdiction and control' over Iraq's detention facilities from June 2003 to June 2004, which includes the period during which Plaintiffs were abused. So under whose 'territorial jurisdiction' do these facilities fall during that time if not the United States? There was no sovereign, no Iraqi government, Page 14no armed forces. Although other countries participated in the CPA, the United States had ultimate authority and control.²

In rejecting the applicability of the presumption against extraterritoriality, Rasul considered only that the “express terms of [the United States’s] agreements with Cuba” provide that 'the United States exercises complete jurisdiction and control over the Guantanamo Bay Naval Base, and may continue to [exercise such control] permanently if it so chooses.” 542 U.S. at 480. The decision mentions only two factors bearing on the Court’s conclusion: the existence of “complete jurisdiction and control” and the possibility that such control could be permanent. Id.

Applying Rasul here leads us to conclude that the presumption against extraterritoriality has no application in this case. Through the CPA, the United States exercised equal or greater 'complete jurisdiction and control' over Iraq's detention facilities, including Abu Ghraib, at the relevant time as it does at Guantanamo Bay. The

Fed. Cts. Br. 17–18 (footnotes and citation omitted).

Page 15temporary nature of the occupation is of no moment when, at the time of the conduct at issue, there was truly no alternative authority in place over the prison. Further, the duration of the occupation was entirely at the discretion of the United States.³

The dissent says we cannot apply Rasul to ATS claims because it specifically involved the habeas statute. Dissent at 944. But we do not read Rasul so narrowly. As the dissent explains, applying the presumption against extraterritoriality is a two-step inquiry where the first step is directed at the statute, and the second step is directed to the facts of the case. If the Rasul Court intended to hold that the presumption against extraterritoriality is rebutted entirely as to the habeas statute, its discussion of whether Guantanamo Bay is within the territorial jurisdiction of the United States is superfluous: once the presumption is rebutted, a statute can properly reach conduct outside of the United States. In our view, Rasul 's discussion of what it means for conduct to be extraterritorial with citation to Foley Brothers reveals that, whatever the relationship between the habeas statute and the presumption in the first place, a court can also consider whether conduct occurring outside of the fifty states is nonetheless “within ‘the territorial jurisdiction’ of the United States.” Rasul , 542 U.S. at 480 (quoting Foley Bros., Inc. , 336 U.S. at 285).

CACI states that “[t]he U.N. Security Council specifically acknowledged that the temporary Coalition presence in Iraq did not affect ‘the sovereignty and territorial integrity of Iraq.’” Reply Br. at 4–5 (quoting S.C. Res. 1483 (May 22, 2003)). But the “sovereignty”Page 16 mentioned in the resolution (“Reaffirming the sovereignty and territorial integrity of Iraq,” S.C. Res. 1483) means something different than the contemplations of de facto or de jure sovereignty in Rasul . The Security Council references Iraq’s sovereignty— notwithstanding the absence of a sovereign government—to distinguish the occupation by the United States from a military invasion intended to annex or conquer the invaded nation. The language of the resolution merely reaffirms that the occupation would be temporary and eventually, power would transition back to an Iraqi government. This statement by the Security Council does not undermine our conclusion.

CACI points to the Fifth Circuit’s consideration of Rasul in Adhikari v. Kellogg Brown & Root, Inc. , 845 F.3d 184 (5th Cir. 2017). This case holds that for purposes of the ATS, conduct in Iraq in 2004 'cannot constitute domestic conduct relevant to [the plaintiffs’] ATS claims.” Id. at 197. The court relied on the impermanent nature of the United States’s occupation, which it said distinguished Iraq from Guantanamo Bay, “over which the United States had ‘unchallenged and indefinite control.’” Id. (quoting Rasul , 542 U.S. at 487 (Kennedy, J., concurring)). Because we do not find the temporary nature of the occupation dispositive, we decline to follow the Fifth Circuit’s conclusion.⁴

Page 17The D.C. Circuit also considered whether Rasul could apply somewhere other than Guantanamo Bay, this time applying the case to Bagram Airfield in Afghanistan. Al Maqaleh v. Gates , 605 F.3d 84 (D.C. Cir. 2010). There, the court held that Rasul 's reasoning did not render Bagram within the United States' territorial jurisdiction, and expressed concern that the opposite conclusion “would seem to create the potential for the extraterritorial extension of the [provision at issue] to noncitizens held in any United States military facility in the world, and perhaps to an undeterminable number of other UnitedStates-leased facilities as well.” 5 Id. at 95. But we are not concerned with any floodgates opening as a result of our decision here. Distinct from the situation the Al Maqaleh court confronted, our opinion does not apply federal law to all military facilities. We specifically limit our holding to the prisons and detention centers in Iraq over which the United States exercised plenary authority from June 2003 to June 2004 by virtue of CPA Order 10. A logical extension of this conclusion is that federal law applied to any conduct occurring in Iraq’s prisons between June 2003 and June 2004, but we have faith in the jurisdictional, prudential, and practical limits of litigation in our federal courts such that this decision will Page 18not allow plaintiffs to bring cases here that do not truly belong.⁶ Accordingly, we decline to follow the D.C. Circuit. We conclude Rasul 's logic applies to conduct occurring in prison and detention centers in Iraq between June 2003 and June 2004, such that this conduct was not extraterritorial. See Rasul , 542 U.S. 480.

Even if Iraq could not be properly considered within the territorial jurisdiction of the United States during the relevant period, the Kiobel Court offered one alternative avenue to apply the ATS to conduct occurring in Iraq without offending the presumption against extraterritoriality. Kiobel petitioners sought a holding from the Court that Congress did intend for the statute to apply extraterritorially. 569 U.S. at 118. One of their arguments stemmed from the original understanding at the time of the Founding that piracy was a paradigmatic violation of the law of nations: if Congress intended the ATS to apply to piracy, which “typically occurs on the high seas, beyond the territorial jurisdiction of the United States or any other country,' then Congress 'necessarily anticipated the statute Page 19would apply to conduct occurring abroad.” Id. at 121. In response, the Court commented that “pirates may well be a category unto themselves,” id. , implying that applying the ATS to piracy did not implicate the presumption against extraterritoriality. “Applying U.S. law to pirates[] . . . does not typically impose the sovereign will of the United States onto conduct occurring within the territorial jurisdiction of another sovereign, and therefore carries less direct foreign policy consequences.' Id. It further observed that '[p]irates were fair game wherever found, by any nation, because they generally did not operate within any jurisdiction.' Id. ; accord Sosa , 542 U.S. at 762 (Breyer, J., concurring) (“[I]n the 18th century, nations reached consensus not only on the substantive principle that acts of piracy were universally wrong but also on the jurisdictional principle that any nation that found a pirate could prosecute him.”).

Why is piracy relevant here? Beginning with the first modern ATS case, jurists have commented that “for purposes of civil liability, the torturer has become—like the pirate and slave trader before himhostis humani generis , an enemy of all mankind.' Filartiga v. Pena-Irala , 630 F.2d 876, 890 (2d Cir. 1980); Nestlé , 593 U.S. at 647 (Sotomayor, J., concurring); Linder v. Portocarrero , 963 F.2d 332, 336 (11th Cir. 1992); see also Tr. of Oral Argument at 26, Kiobel , 569 U.S. 108 (No. 10-1491) (Justice Breyer: '[W]ho are today's pirates? And if Hitler isn't a pirate, who is? And if, in fact, an equivalent torturer or dictator who wants to destroy an entire race in his own country is not the equivalent of today’s pirate, who is?”). And consistent with the notion expressed in Kiobel and Sosa that “[p]irates were fair game wherever found, by any nation,” Kiobel , 569 U.S. at 121; Sosa , 542 U.S. at 762 (Breyer, J., concurring), international law similarly Page 20requires that universal jurisdiction exists for torture, see Karen Parker & Lyn Beth Neylon, Jus Cogens: Compelling the Law of Human Rights , 12 Hastings Int'l & Comp. L. Rev. 411, 455–56 (1989) (discussing the requirement of universal jurisdiction for violations of jus cogens norms, including torture); U.N. Convention Against Torture and Other Cruel, Inhuman or Degrading Treatment or Punishment art. 5, ¶ 2, Apr. 18, 1988, 1465 U.N.T.S. 113 (hereinafter Convention Against Torture) (“Each State Party shall likewise take such measures as may be necessary to establish its jurisdiction over such offences in cases where the alleged offender is present in any territory under its jurisdiction . . . .'); Eugene Kontorovich, The Piracy Analogy: Modern Universal Jurisdiction’s Hollow Foundation , 45 Harv. Int’l L.J. 183, 217 (2004) (listing torture as one of the “offenses that, by analogy to piracy, have come within the ambit of [new universal jurisdiction]”). Thus, piracy and torture share two important defining characteristics: first, the torturer and the pirate are in the same small category of criminals representing enemies of all mankind, Filartiga , 630 F.2d at 890; and second, the gravity of these jus cogens violations demands that universal jurisdiction apply to allow any country to prosecute pirates and torturers wherever found, Kiobel , 569 U.S. at 121; Convention Against Torture art 5 ¶ 2. So we offer that for purposes of the Alien Tort Statute, we may consider acts of torture commensurate with acts of piracy.

In the reasoning of Kiobel , we find the idea that even absent congressional intent for a statute to apply extraterritorially, a statute may still be applied to conduct—like piracy or torture—occurring outside “the territorial jurisdiction of the United States or any other country” when it does not implicate “the territory of another sovereign.” See 569 U.S. atPage 21 121 . And those conditions are satisfied in this case. If not the United States, under whose territorial jurisdiction was Iraq at the end of 2003? The U.S. military had deposed the prior sovereign, Saddam Hussein, and had not yet ceded its occupational power to the Iraqi Interim Government. We run no risk of offending any sovereign by imposing the sovereign will of the United States to Iraq while the CPA was the only relevant authority in the country.

Thus, this case presents the unique convergence of two circumstances: a case against modern-day pirates in a jurisdiction that was either territory of the United States or was outside the territory of any other sovereign. Together, these facts lead us to conclude that even if Iraq was not within the “territorial jurisdiction” of the United States at the relevant time, its place outside the sovereignty of any country displaces the presumption against extraterritoriality with enough force to allow our courts to prosecute conduct akin to piracy. See Kiobel , 569 U.S. at 121.

Finally, even if we determine that conduct occurring in Iraq was extraterritorial for the purposes of the presumption, we are not convinced that the district court erred in concluding that sufficient conduct relevant to the focus of the ATS occurred domestically. We last considered the sufficiency of domestic conduct in Al Shimari III . We applied Kiobel 's touch-and-concern test by considering the location of the defendant-corporation's headquarters, reviewing where the relevant employees undertook the operative actions ( i.e. , hiring, contracting with the U.S. Department of the Interior, obtaining security Page 22clearances), and the receipt of interrogation reports and alleged cover up at the Virginia headquarters. 758 F.3d at 528–29. We concluded that “the plaintiffs’ ATS claims ‘touch and concern' the territory of the United States with sufficient force to displace the presumption against extraterritorial application.” Id. at 530 (quoting Kiobel , 569 U.S at 124-25). This conclusion must now be reexamined in light of the Supreme Court's intervening decision in Nestlé .

To best understand Nestlé , we review the relevant conclusions of the lower courts as well as the holding of the Supreme Court. After the district court dismissed the case on extraterritoriality grounds, the Ninth Circuit reversed, concluding that the facts of the case “paint a picture of overseas slave labor that defendants perpetuated from headquarters in the United States” such that the domestic conduct alleged was relevant to the ATS’s focus. Doe v. Nestle, S.A. , 929 F.3d 623, 642 (9th Cir. 2019), rev’d and remanded sub nom. Nestlé USA, Inc. v. Doe , 593 U.S. 628 (2021). The appellate court considered the following domestic activity: financing arrangements consisting of payments “akin to ‘kickbacks’” made to Ivory Coast farmers and cooperatives, coupled with inspections of Ivory Coast operations by employees from defendants’ U.S. headquarters who would then “report back to the United States offices where these financing decisions[] . . . originated.” Id. at 641– 42.

The Supreme Court rejected the Ninth Circuit's approach, concluding that the plaintiffs 'impermissibly seek extraterritorial application of the ATS.' Nestlé , 593 U.S. at 634. Where the Ninth Circuit considered that “every major operational decision by both companies is made in or approved in the U.S.” as evidence of relevant domestic conduct, Page 23the Supreme Court characterized this decisionmaking as “general corporate activity” which “cannot alone establish domestic application of the ATS.” 593 U.S. at 634. The Court reaffirmed that “a plaintiff does not plead facts sufficient to support domestic application of the ATS simply by alleging ‘mere corporate presence’ of a defendant.” Id. (quoting Kiobel , 569 U.S., at 125). After Nestlé , courts must also exclude “activity common to most corporations” like “making ‘operational decisions’” where “generic allegations of this sort do not draw a sufficient connection between the cause of action respondents seek . . . and domestic conduct.' Id.

Reading the decisions of the Ninth Circuit and the Supreme Court together, Nestlé does not permit us to consider conduct like the domestic location of CACI’s headquarters, the domestic contract issuance, or domestic payment processing activities that we discussed in Al Shimari III because these activities are too close to 'general corporate activity.” But the remaining conduct occurring within the United States—hiring, issuance of security clearances, and attempted cover up-is still properly considered following Nestlé .

Nestlé commands that 'plaintiffs must establish that 'the conduct relevant to the statute’s focus occurred in the United States.’” 7 593 U.S. at 633 (quoting RJR Nabisco, Page 24Inc. , 579 U.S. at 337). We agree with the Ninth Circuit that “conduct that occurs within the United States and violates customary international law is most relevant to the ATS’s aim of providing a forum to address violations of international norms that take place in U.S. territory.” Doe I v. Cisco Sys., Inc. , 73 F.4th 700, 737 (9th Cir. 2023), cert. granted , --- S. Ct. ----, 2026 WL 73088 (2026). And we find that customary international law addresses far more conduct than merely the act of torture or CIDT. The Convention Against Torture imposes affirmative responsibilities on signatory nations.⁸ It requires signatories to 'take effective legislative, administrative, judicial or other measures to prevent acts of torture in any territory under its jurisdiction”; “ensure that education and information regarding the prohibition against torture are fully included in the training of law enforcement personnel, civil or military, . . . and other persons who may be involved in the custody, interrogation or treatment of any individual subjected to any form of arrest, detention or imprisonment'; and 'keep under systematic review interrogation rules, instructions, methods and practices as well as arrangements for the custody and treatment of persons subjected to any form of arrest, detention or imprisonment in any territory under its jurisdiction, with a view to preventing any cases of torture.' Convention Against Torture art. 2, ¶ 1, art. 10, ¶ 1, art. 11.

The affirmative obligations of the Convention Against Torture add special weight to the conduct in the United States related to hiring, issuing security clearances, and Page 25mishandling or ignoring reports of possible torture because of the heightened “connection between the cause of action [Plaintiffs] seek . . . and domestic conduct.” 9 See Nestlé , 593 U.S. at 634. In this case, these activities are more than mere “‘operational decisions’ . . . common to most corporations,' see id. ; rather, they are points at which international norms would require that CACI implement training, distribute appropriate interrogation protocols, and intervene to stop inappropriate interrogation activity, all in service of preventing torture. In our view, CACI’s domestic conduct was in and of itself a violation of international treaty obligations with respect to torture. Thus, this conduct is sufficiently relevant to the focus of the ATS to conclude that the statute's application here is not impermissibly extraterritorial.

Of course, we are mindful of Nestlé 's warning that '[t]he presumption against extraterritorial application would be a craven watchdog indeed if it retreated to its kennel whenever some domestic activity is involved in the case.' 593 U.S. at 634 (quoting Morrison v. Nat’l Austl. Bank Ltd. , 561 U.S. 247, 266 (2010)). Morrison comments thatPage 26 'it is a rare case of prohibited extraterritorial application that lacks all contact with the territory of the United States,” and in application, Morrison reminds lower courts to look to the focus of the statute to determine whether the domestic contact satisfies the presumption. 561 U.S. at 266. Thus, when the focus of the statute is regulation of purchases and sales of securities, the statute attaches to securities fraud only when the security in question was listed on an American exchange or purchased or sold within the United States. Id. When the focus of the statute is to regulate the conditions of employment, it properly applies to protect employees working within the United States; employees working abroad are not protected even if hired within the United States because hiring does not speak to the focus of the statute. Id. But here, where we have determined that conduct violating customary international norms is most relevant to the focus of the ATS, and where the failure to implement certain controls to prevent torture is itself a violation of customary international law, domestic activity like hiring, issuing security clearances, and responding to alleged misconduct is directly relevant to this focus.

The dissent finds that the domestic conduct in this case is all essentially general corporate activity similar to what was rejected in Nestlé . See Dissent at 102. But the dissent fails to ground this conclusion in any articulated focus of the ATS. As Morrison explains, the focus of the statute is the lens through which conduct becomes relevant or irrelevant. 561 U.S. at 266. And in considering the ATS, the specific violation of the law of nations must also inform the inquiry. Because of this, the same conduct may be relevant with respect to one statute and irrelevant to another; and for the ATS, the same conduct may be relevant in some cases but not others by reference to which violation of the law of Page 27nations is at issue. The dissent does not grapple with our analysis of how the domestic conduct at issue here itself violated international norms and treaty obligations with respect to torture, which is clearly within the focus of the statute.

* * *

In sum, we conclude that the district court properly exercised subject matter jurisdiction in this case. We find that application of the ATS here does not implicate the presumption against extraterritoriality because conduct in Iraq is properly considered within the territorial jurisdiction of the United States. See Rasul , 542 U.S. 480. Alternatively, because international law demands universal jurisdiction for torture, and because the torture in this case occurred beyond the jurisdiction of any sovereign, application of the ATS to conduct occurring in Iraq does not offend the presumption against extraterritoriality. See Kiobel , 569 U.S. at 121. We also find that sufficient conduct relevant to the focus of the ATS occurred within the United States to maintain a domestic application of the statute. See Nestlé , 593 U.S. at 633.

Next, CACI argues that the district court erred in recognizing Plaintiffs’ causes of action, making broad and unnuanced arguments against “judge-created claims.” Opening Br. at 29; cf. Kiobel , 569 U.S. at 116 (noting that the ATS 'allows federal courts to recognize certain causes of action based on sufficiently definite norms of international law”). Rather than accept CACI’s sledgehammer approach, we address the propriety of the causes of action in this case under Sosa by considering (1) whether the ATS permits Page 28conspiracy as a cause of action, and (2) whether corporate liability is appropriate under the ATS.

Sosa prescribes a two-part test to recognize a cause of action under the ATS. First, courts must decide if the claim “rest[s] on a norm of international character accepted by the civilized world and defined with a specificity comparable to the features of the 18thcentury paradigms we have recognized.” Sosa , 542 U.S. at 725. Second, 'it must be determined further whether allowing this case to proceed under the ATS is a proper exercise of judicial discretion, or instead whether caution requires the political branches to grant specific authority before corporate liability can be imposed.” Jesner v. Arab Bank, PLC , 584 U.S. 241, 258 (2018) (citing Sosa , 542 U.S. at 732–33 & nn.20–21).

i.

We address Sosa 's first step. Although the Courts of Appeals are so far united in finding that “aiding and abetting liability claims may proceed under the ATS,” Cisco Sys., Inc. , 73 F.4th at 718 (collecting cases), 10 only the Eleventh Circuit has explicitly addressed Page 29conspiracy liability, see Doe v. Drummond Co., Inc. , 782 F.3d 576, 597 (11th Cir. 2015) ('Our precedent makes clear that claims based on aiding and abetting and conspiracy liability are cognizable under the ATS.”). Expanding on the work of our sister circuits, which universally permit theories of secondary liability in ATS cases, we agree with the Eleventh Circuit that conspiracy liability is also proper under the statute.

We find that conspiracy to commit torture or CIDT violates “norm[s] of customary international law so well defined as to support the creation of a federal remedy.' See Sosa , 542 U.S. at 738. To conclude that conspiracy to commit torture “rest[s] on a norm of international character accepted by the civilized world,' id. at 725, we survey various sources of international law 11 (borrowing from the existing body of work on this subject from other federal courts). Similar to aiding-and-abetting liability, liability for conspirators was “[r]ecognized as part of the customary law which authorized and was applied by the war crimes trials following the Second World War' and subsequently 'it has been frequently invoked in international law instruments as an accepted mode of liability.” See Khulumani v. Barclay Nat’l Bank Ltd. , 504 F.3d 254, 270 (2d Cir. 2007) (Katzmann, J., concurring) (discussing liability for aiding and abetting). Given the consensus among these Page 30authorities, we find that “the concept of criminal [conspiracy] liability is well established in international law.” See id. (citation modified).

International laws used to prosecute Nazi war criminals “created criminal liability not only for principals who committed acts of genocide or war crimes but also for those who were connected with any plans or enterprises involving the commission of such crimes.” Presbyterian Church of Sudan v. Talisman Energy, Inc. , 244 F. Supp. 2d 289, 322 (S.D.N.Y. 2003) (citing William A. Schabas, Enforcing International Humanitarian Law: Catching the Accomplices, 83 Int'l Rev. Red Cross 439, 442 (2001)). The London Charter 12 specifically assigned responsibility to '[l]eaders, organisers, instigators and accomplices participating in the formulation or execution of a common plan or conspiracy.” Agreement for the Prosecution and Punishment of Major War Criminals of the European Axis, and Establishing the Charter of the International Military Tribunal art. 6, Aug. 8, 1945, 82 U.N.T.S. 279. Thus, conspiracy liability was expressly recognized in prosecuting war crimes and crimes against humanity after World War II.

The Statutes of the International Criminal Tribunal for the former Yugoslavia and International Criminal Tribunal for Rwanda similarly establish criminal liability for those who have “planned, instigated, ordered, committed, or otherwise aided and abetted in the planning, preparation or execution of a crime.” Talisman Energy, Inc. , 244 F. Supp. 2d at 322 (quoting S.C. Res. 827, annex, Statute of the International Criminal Tribunal for the Page 31former Yugoslavia art. 7(1) (May 25, 1993); S.C. Res. 955, annex, Statute of the International Criminal Tribunal for Rwanda art. 6(1) (Nov. 8, 1994)). The broad definition of accomplice liability in these statutes encompasses the conspiracy theory present here, and its inclusion in the statutes “reflects a determination by both the Secretary-General and the Security Council . . . that such liability is firmly established in customary international law.' Khulumani , 504 F.3d at 274–75 (Katzmann, J., concurring).

In addition to these sources, the Rome Statute for the International Criminal Court provides that criminal liability attaches for torture and “inhuman treatment” when a person commits a crime individually or jointly, orders the commission of the crime, “aids, abets[,] or otherwise assists in its commission,' or '[i]n any other way contributes' to the commission of the crime “by a group of persons acting with a common purpose” as long as the contribution is intentional and either with the purpose of furthering the criminal activity or purpose of the group or with knowledge of the group’s intent to commit a crime. Rome Statute of the International Criminal Court art. 8, ¶ 2(a)(ii), art.²⁵ ¶ 3, July 17, 1998, 2187 U.N.T.S. 3. Finally, the Convention Against Torture requires that signatories extend criminal liability for torture 'to an act by any person which constitutes complicity or participation in torture.” Convention Against Torture art. 4, ¶ 1.

From reviewing these sources of international law, we conclude that conspiracy liability is an appropriate cause of action under the ATS based on Sosa 's first step. Specifically, this claim is appropriately “based on the present-day law of nations” as it 'rest[s] on a norm of international character accepted by the civilized world,' as Page 32specifically defined by the sources of international law reviewed above. See Sosa , 542 U.S. at 725.

After determining that international law has identified a specific controlling norm recognizing liability for conspiracy, Sosa ’s second step dictates that “it must be determined further whether allowing this case to proceed under the ATS is a proper exercise of judicial discretion, or instead whether caution requires the political branches to grant specific authority before . . . liability can be imposed.” Jesner , 584 U.S. at 258 (citing Sosa , 542 U.S. at 732–33 & nn.20–21).

Sosa enumerates the concerns that “argue for judicial caution” when recognizing causes of action under the ATS. 542 U.S. at 725. These concerns arise from our changing “conception of the common law” since 1789, away from the idea that “the common law [was] ‘a transcendental body of law outside of any particular State but obligatory within it and unless changed by statute’” towards the “general practice” of “look[ing] for legislative guidance before exercising innovative authority over substantive law.' Id. at 725-26 (quoting Black & White Taxicab & Transfer Co. v. Brown & Yellow Taxicab & Transfer Co. , 276 U.S. 518, 533 (1928) (Holmes, J., dissenting)). The preference for legislative guidance is particularly acute when creating private causes of action for violating international law: “the potential implications for the foreign relations of the United States of recognizing such causes should make courts particularly wary of impinging on the Page 33discretion of the Legislative and Executive Branches in managing foreign affairs.” Id. at 727.

Notwithstanding these concerns, the Court commented both in Sosa and again in Jesner that the ATS 'was not enacted to sit on a shelf awaiting further legislation.' Jesner , 584 U.S. at 254; Sosa , 542 U.S. at 714. We consider Sosa 's advice and conclude that “allowing this case to proceed under the ATS is a proper exercise of judicial discretion.” Jesner , 584 U.S. at 258.

We are convinced that this case does not offend the authority of the other branches of government. In enacting the Torture Victim Prevention Act (TVPA), Congress sought to extend “a cause of action that has been successfully maintained under [the ATS]” to provide “a clear and specific remedy, not limited to aliens, for torture and extrajudicial killing.” 13 See H.R. Rep. No. 102–367, pt. 2, at 3 (1991). The legislative history of the TVPA makes clear that Congress envisioned that the rights of aliens to secure a civil remedy for “claims based on torture” are unequivocally secured by the ATS. Id. at 4. It Page 34expressed a clear intent for the ATS to “remain intact to permit suits based on [norms including torture] that already exist or may ripen in the future into rules of customary international law.” Id. Congress further explained that the TVPA was necessary because of judicial concerns 'that separation of powers principles required an explicit-and preferably contemporary—grant by Congress of a private right of action before U.S. courts could consider cases likely to impact on U.S. foreign relations.” Id. Given Congress's approbation of causes of action related to torture under the ATS, and because allowing this case to proceed lacks any potential to negatively affect foreign relations, it poses little if any threat to separation of powers.¹⁴

Indeed, this case presents a paradigmatic application of the ATS, such that the greatest threat to foreign relations results if this case were not permitted to proceed. Sosa explained that its concern for collateral consequences impacting foreign relations stemmed from the possibility of suits “that would go so far as to claim a limit on the power of foreign governments over their own citizens, and to hold that a foreign government or its agent has transgressed those limits,” and advised “great caution” to federal courts attempting “to craft Page 35remedies for the violation of new norms of international law” that could “raise risks of adverse foreign policy consequences.” 542 U.S. at 727–28. But this case does not seek accountability from any foreign government, agents thereof, or foreign citizens. Rather, this suit furthers “[t]he principal objective of the [ATS]”: “to avoid foreign entanglements by ensuring the availability of a federal forum where the failure to provide one might cause another nation to hold the United States responsible for an injury to a foreign citizen.” Jesner , 584 U.S. at 255. Indeed, we cannot imagine a statement more offensive to the arena of foreign relations than to proclaim that courts of the United States may not provide a remedy to foreign nationals who were tortured by members of the U.S. military as part of a conspiracy to extract intelligence using universally condemned means of interrogation. Rather, our decision is in furtherance of an 'important American national interest' in “preventing the United States from becoming a safe harbor (free of civil as well as criminal liability) for a torturer or any other common enemy of mankind.” See Kiobel , 569 U.S. at 127 (Breyer, J., concurring).

The dissent contends that we have exceeded the scope of our judicial role in finding that this case is a paradigmatic ATS case as intended by the original Congress. Dissent at 113. The Supreme Court explained that “[t]he principal objective of the statute, when first enacted, was to avoid foreign entanglements by ensuring the availability of a federal forum where the failure to provide one might cause another nation to hold the United States responsible for an injury to a foreign citizen.” Jesner , 584 U.S. at 255. On the other side of the spectrum are cases that “claim a limit on the power of foreign governments over their own citizens, and to hold that a foreign government or its agent has transgressed those Page 36limits.” See Sosa , 542 U.S. at 727. If the ATS 'was not enacted to sit on a shelf,' Jesner , 584 U.S. at 254, and if it was meant to enable courts to hear those cases posing risks of “foreign entanglement[]” absent federal jurisdiction, see id. at 255, how can it be true that the courts cannot assess where on the spectrum of foreign policy encroachment a particular case falls? Perplexingly, the dissent does not suggest that this case does pose foreign policy concerns, only that we are “not qualified” to judge if it does or not. See Dissent at 113. Indeed, the dissent mischaracterizes how we effectuate our respect for the co-extensive branches of government. We may not “supplant a foreign policy decision of the political branches with [our] own unmoored determination of what United States policy . . . should be.” See Zivotofsky ex rel. Zivotofsky v. Clinton , 566 U.S. 189, 196 (2012). Here we are not determining policy—we are applying the clear policy of the law of the United States and the law of nations that prohibits torture, including, as conceded by the dissent, for “all who participate in torture” and not “just those who directly do the torturing.” See Dissent at 1066. And we maintain that there is no offense to Congress or the Executive in recognizing this cause of action where “the potential implications for the foreign relations of the United States” from such recognition are naught. See Sosa , 542 U.S. at 727.

CACI's citation to Bivens cases does nothing to disrupt our conclusion.¹⁵ We have exercised the necessary caution under the mandates of Sosa and Jesner . To the extent that Page 37Bivens cases require courts to be even more cautious, that concern arises from the fact that Bivens suits lack any statutory cause of action. This is in stark contrast to the 'firm statutory basis' for ATS suits, and thus the Bivens cases 'in no way foreclose[]' the causes of action in this case. See Doe v. Exxon Mobil Corp. , 654 F.3d 11, 55 (D.C. Cir. 2011), vacated on other grounds , 527 F. App’x 7 (D.C. Cir. 2013) (“Although the private right of action recognized in Bivens lacked any statutory basis, the First Congress enacted the ATS with the understanding that ‘the district courts would recognize private causes of action for certain torts in violation of the law of nations.’” (first citing Corr. Servs. Corp. v. Malesko , 534 U.S. 61, 66–67 (2001); and then quoting Sosa , 542 U.S. at 724)).

That this case imposes liability on CACI, a corporation, does not offend our jurisdiction under the ATS. The Courts of Appeals have taken different approaches to the question of corporate liability, 16 but in Nestlé , five justices of the Supreme Court agreed Page 38that “[n]othing in the ATS supplies corporations with special protections against suit.” 593 U.S. at 641 (Gorsuch, J., concurring); see also id. at 652 n.4 (Sotomayor, J., concurring) (commenting that Justice Sotomayor and “four other Justices” would find that domestic corporations are not 'immune from suit under the ATS'). We adopt the reasoning of Justice Gorsuch in finding that “distinguishing between individuals and corporations would seem to make little sense” in cases under the ATS. Id. at 642 (Gorsuch, J., concurring). Justice Gorsuch examined the text and intent of the ATS and early understandings of corporate liability to conclude that ATS actions may properly be brought against corporations. Id. at 640–46. The statute took care to identify which plaintiffs may bring suit, but “nowhere does it suggest that anything depends on whether the defendant happens to be a person or a corporation.” Id. at 641. Considering the “circumstances surrounding the ATS’s adoption” lends further support: Congress’s “effort to ensure judicial recourse Page 39for tortious conduct that otherwise could have provided foreign nations with just cause for reprisals or war” would not be furthered by limiting who can be named as defendant. Id. at 642 (citation modified). 'If early Americans assaulted or abducted the French Ambassador, what difference would it have made if the culprits acted individually or corporately? Either way, this Nation’s failure to ‘oblige the guilty to repair the damage’ would have provided just cause for reprisals or worse.” Id. at 642-43 (quoting 1 E. de Vattel, The Law of Nations , bk. II, § 76, at 145 (1760)). Finally, both then and now, “the law places corporations and individuals on equal footing when it comes to assigning rights and duties.' Id. at 641. Accordingly, we conclude that the ATS permits the imposition of corporate liability.

CACI raises four additional threshold issues that it says demonstrate error by the district court in allowing the case to proceed to trial. First, CACI maintains that it is entitled to derivative sovereign immunity. Second, CACI argues that the case is nonjusticiable under the political question doctrine because it requires the court to adjudicate the propriety of military decisions. Third, CACI points to the Federal Tort Claims Act and contends that Plaintiffs’ claims are preempted. Fourth, CACI claims the district court erred in failing to dismiss the case due to the existence of state secrets, which impacted CACI’s ability to defend itself. We find no reversible error on these questions.

CACI’s argument regarding its entitlement to derivative sovereign immunity is in two parts: (1) the district court erred in finding that the United States waived sovereign immunity for jus cogens violations, and (2) that because the United States enjoys sovereign immunity, CACI should benefit from derivative sovereign immunity. We find that the district court erred in the basis on which it granted judgment to the United States because the United States is entitled to sovereign immunity. We nonetheless affirm that CACI was not entitled to derivative sovereign immunity because it does not establish that the government authorized the specific conduct at issue here.

As discussed above, CACI filed a third-party complaint against the United States in 2018, “seeking recovery from Third-Party Defendants the United States of America and John Does 1-60 for contribution, indemnification, and exoneration, and for breach of contract in the case of the United States, in the event that CACI PT is held liable on Plaintiffs’ claims in this action.” J.A. 549. The government moved to dismiss on sovereign immunity grounds. While the motion to dismiss was pending, the government filed a motion for summary judgment, arguing that the United States was entitled to judgment as a matter of law because CACI entered into a settlement agreement with the government which settled all claims and disputes arising from, inter alia , the agreement for CACI to send interrogators to Iraq. The district court denied the government’s motion to dismiss, explaining that the court was “struggling with the concept that sovereign immunity should Page 41protect any government from suit for jus cogens violations.” J.A. 4062. The district court then granted the government's motion for summary judgment, finding that 'the unambiguous wording of the settlement agreement as concerning 'all claims and disputes . . . arising out of or related to the terminated Task Orders’ bars CACI’s claim against the United States.” Al Shimari v. CACI Premier Tech., Inc. , 368 F. Supp. 3d 935, 973–74 (E.D. Va. 2019).

We conclude that the district court erred in denying the government’s motion to dismiss on sovereign immunity grounds. “[T]he Alien Tort Statute has been interpreted as a jurisdictional statute only-it has not been held to imply any waiver of sovereign immunity. Thus, any party asserting jurisdiction under the Alien Tort Statute must establish, independent of that statute, that the United States has consented to suit.' Goldstar (Panama) S.A. v. United States , 967 F.2d 965, 968 (4th Cir. 1992) (citations omitted); accord D.J.C.V. v. United States , 605 F. Supp. 3d 571, 609 (S.D.N.Y. 2022) (“[A]s almost every court to consider the question has held, the United States is immune from suit under the ATS, even for alleged violations of jus cogens norms, as the Complaint here articulates.”) (collecting cases). Because CACI does not dispute that the United States is entitled to sovereign immunity—and therefore asserts no bases from which we could consider whether the United States consented to suit—we conclude that the United States is entitled to sovereign immunity in this case. We vacate the orders of the district court denying the government's motion to dismiss and granting its motion for summary judgment and remand with instructions to dismiss CACI’s claims against the United States on the basis of sovereign immunity.

Page 42CACI argues that because the United States is entitled to sovereign immunity, it is entitled to derivative sovereign immunity. Extending immunity to government contractors prevents contractors from being “left holding the bag—facing full liability for actions taken in conjunction with government employees who enjoy immunity for the same activity.” See Filarsky v. Delia , 566 U.S. 377, 391 (2012). But a crucial predicate question is whether the actions taken by the contractor forming the basis of potential liability were actually authorized by the government. We need not be concerned for contractors “left holding the bag” where, as here, the contractor materially departed from orders and liability results wholly from its deviation.

“[A] government contractor is not subject to suit if (1) the government authorized the contractor's actions and (2) the government 'validly conferred' that authorization, meaning it acted within its constitutional power.' In re KBR, Inc., Burn Pit Litig. , 744 F.3d 326, 342 (4th Cir. 2016) (quoting Yearsley v. W.A. Ross. Constr. Co. , 309 U.S. 18, 20–21 (1940)). 'When a contractor violates both federal law and the Government's explicit instructions, as alleged here, no immunity shields the contractor from suit.” CampbellEwald Co. v. Gomez , 577 U.S. 153, 154 (2016). In other words, 'the contractor must adhere to the government’s instructions to enjoy derivative sovereign immunity.” Burn Pit , 744 F.3d at 345. '[I]t is incumbent on one who relies on [an affirmative defense] to set it up and establish it.' Dixon v. United States , 548 U.S. 1, 13 (2006). Thus, CACI bears the burden of demonstrating that its actions at issue were authorized by the government’s instructions related to the services CACI provided.

Page 43In their briefs, the parties reenact the exact dispute at the core of Burn Pit : 'whether we should construe the scope of [CACI]’s authority narrowly or broadly.” 744 F.3d at 344. CACI states generally that “its activities were authorized and directed by the government under validly-conferred authority,” that it “adhered to the terms of the contracts” in its provision of interrogation personnel to the Army, and that “the United States has never alleged in any context that CACI breached the contracts.” Opening Br. at 34–35 (citations omitted). Meanwhile, Plaintiffs point out that “CACI does not argue that the government gave it 'explicit instructions' to undertake the conduct the jury found to violate international law.” Resp. Br. at 50. CACI argues in reply that Plaintiffs’ failure to seek a special verdict on the question of breach of contract or violation of federal law waives this issue.

Plaintiffs describe the proper test in this Circuit. As we concluded in Burn Pit , “staying within the thematic umbrella of the work that the government authorized is not enough to render the contractor’s activities ‘act[s] of the government’ [sufficient to trigger Yearsley protection].' 744 F.3d at 345 (quoting Yearsley , 309 U.S. at 22) (alteration in original); see also, e.g. , In re OPM Data Sec. Breach Litig. , 928 F.3d 42, 70 (D.C. Cir. 2019) (“KeyPoint does not argue that OPM ‘authorized and directed’ it to design its system with the security flaws that Arnold Plaintiffs identify. So KeyPoint cannot wrap itself in derivative immunity garb on the ground that it 'simply performed as the Government directed.’” (quoting Campbell-Ewald , 577 U.S. at 167)); Lethgo v. CP IV Waterfront, LLC , Nos. 23-15583, 23-15804, 2025 WL 1794107, at *2 (9th Cir. June 30, 2025) (“The defense [of derivative sovereign immunity] is inapplicable here because the Navy did not direct Page 44Appellants to distribute the allegedly contaminated water to residents or otherwise control how Appellants managed the residential community.'). CACI's failure to make any argument that the specific actions at issue in this case were authorized by the government is fatal to its affirmative defense of derivative sovereign immunity. CACI even seems to concede that the conduct alleged here would be expressly prohibited by the contract and governing regulations: “The contract requirements that the military control interrogation operations, including work by CACI interrogators, were consistent with regulations prohibiting contractors from performing 'inherently governmental functions,' such as directing and controlling ‘intelligence and counter-intelligence operations.’” Opening Br. at 34–35 (quoting 48 C.F.R. § 7.503(a), (c)(8)).

That Plaintiffs did not request a special verdict is irrelevant. Rather, as noted above, it was CACI’s burden to set up and establish its affirmative defense. See Dixon , 548 U.S. at 13; Lethgo , 2025 WL 1794107, at *2 ('Appellants also did not carry their burden to establish a colorable federal defense.” (emphasis added)). If CACI felt a special verdict was needed, it was CACI’s burden to request it; it cannot shift this burden to Plaintiffs after the fact. Accordingly, we conclude that CACI did not establish its entitlement to derivative sovereign immunity in this case.

B.

CACI next argues that this case presents nonjusticiable political questions that require inappropriate judicial review of military decisions. We previously considered whether the political question doctrine bars judicial review in this case in both Al Shimari Page 45III and Al Shimari IV . Before we address CACI’s renewed arguments on appeal, we review our prior discussion of this issue and briefly recount relevant procedural history.

In Al Shimari III , we summarized our Circuit’s distillation of the six-factor test from Baker v. Carr , 369 U.S. 186 (1962), into 'two critical components' when considering cases “against government contractors who perform services for the military”:

(1) whether the government contractor was under the “plenary” or “direct” control of the military; and (2) whether national defense interests were 'closely intertwined' with military decisions governing the contractor's conduct, such that a decision on the merits of the claim “would require the judiciary to question actual, sensitive judgments made by the military.”

Al Shimari III , 758 F.3d at 533-34 (quoting Taylor v. Kellogg Brown & Root Servs., Inc. , 658 F.3d 402, 411 (4th Cir. 2011)). We further explained that, in conducting its analysis of the first prong, “a court must inquire whether the military clearly ‘chose how to carry out these tasks,’ rather than give the contractor discretion to determine the manner in which the contractual duties would be performed.' Id. at 534 (quoting Burn Pit , 744 F.3d at 339). However, we were unable to determine at that time whether the Taylor factors were satisfied from “pleadings and the limited record on appeal,” and we remanded the case for further factual development and reconsideration of the issue. Id. at 536–37.

The question returned to us after the district court granted CACI’s motion to dismiss for lack of justiciability. Al Shimari IV , 840 F.3d at 151. We again concluded that the district court’s analysis was lacking, this time because it “began and ended its analysis by drawing conclusions based on the evidence of formal control,” rather than actual control, Page 46and “failed to address the full scope of review” necessary on remand. Id. at 157. We further found that “the district court erred in failing to draw a distinction between unlawful conduct and discretionary acts that were not unlawful when committed.” Id. at 158. We directed the district court to reconsider the issue with the following guidance: 'A contractor’s acts may be shielded from judicial review under the first prong of Taylor only to the extent that those acts (1) were committed under actual control of the military; and (2) were not unlawful.” Id. at 157. Under the second prong, we instructed that Plaintiffs’ claims would “fall outside the protection of the political question doctrine” if they “rest on allegations of unlawful conduct in violation of settled international law or criminal law then applicable to the CACI employees.' Id. at 158.

At the time of Al Shimari IV , Plaintiffs maintained their causes of action under state tort law, including assault and battery and intentional infliction of emotional distress (IIED). We acknowledged at the time that “certain allegations underlying the common law claims may involve conduct that, although tortious under the common law, did not constitute a violation of applicable criminal or international law.” Id. at 160 n.8. For example, “[a] nonconsensual touching that might constitute battery, or conduct that might amount to intentional infliction of emotional distress, under the common law nevertheless may have been an interrogation tactic that the military lawfully could have authorized.” Id. After our decision, Plaintiffs voluntarily dismissed these state tort claims, and as discussed above, the only causes of action remaining in the second trial were conspiracy to commit torture and conspiracy to commit CIDT.

• Page 47At trial, the district court provided the following instructions to the jury 17 :
• (1) Each plaintiff needed to prove that he “was either tortured, or subjected to cruel, inhuman, or degrading treatment.” J.A. 6381.
• (2) To establish that he was tortured, a plaintiff needed to prove that “military personnel subjected him to severe pain or suffering, whether physical or mental' and that the personnel 'inflicted this pain or suffering on him intentionally for the purpose of obtaining information or a confession or for punishment, intimidation, or coercion.” J.A. 6385.
• (3) That “mental pain and suffering” constituting torture “must be prolonged mental harm and caused by, or resulting from, the intentional infliction or threatened infliction of severe physical pain or suffering, or the threat of imminent death.' Id.
• (4) To establish that he suffered CIDT, a plaintiff needed to prove that “military personnel intentionally inflicted acts of [CIDT] on him” and that the “person who intentionally inflicted severe or serious pain or suffering” did so while the plaintiff “was within the custody or control of that person.” J.A. 6387.

1. (5) Page 48That the definition of CIDT is causing 'feelings of fear, anguish, or inferiority capable of humiliating or debasing the victim and possibly breaking his physical or moral resistance.' Id.
2. (a) That the definition of “cruel” treatment is “if it causes serious mental or physical suffering or injury that constitutes a serious attack on human dignity.' Id.
3. (b) That the definition of “inhuman” treatment is “if it deliberately causes severe suffering, mental or physical, which, in the particular situation, is unjustified.' Id.
4. (c) That the definition of “degrading” is “if its effect is to arouse feelings of fear, anguish, or inferiority capable of humiliating or debasing any one of the plaintiffs.' Id.
5. (6) That the jury “must consider the totality of the circumstances” in determining whether CIDT occurred causing “severe psychological or physical harm,” and that whether treatment constitutes CIDT “depends upon an assessment of all the particularities of the evidence before [it], including the specific conditions at issue, duration of the measures imposed, the objectives pursued by the perpetrators, and the physical or mental effects on the person(s) involved.' Id.

Based on these instructions, the jury returned a verdict for Plaintiffs on both causes of action.

Page 49Now CACI asks us to address this issue for a third time. The central question before us is clear: whether CACI’s conduct was unlawful . Although “[a]n affirmative response to either of the two Taylor factors, namely, the fact of direct control or the need to question sensitive military judgments, generally triggers the application of the political question doctrine,” we explained in Al Shimari IV that both prongs are undermined if the conduct alleged is unlawful. 840 F.3d at 155, 157–58. Accordingly, if we find that CACI’s conduct is unlawful, it is irrelevant if the conduct was undertaken under the “direct control” of the military. See id.

The crucial distinction here is whether CACI’s actions were unlawful or if they were in a “grey area”—“although the reasonableness of military conduct may not be justiciable, the lawfulness of that conduct assuredly is.” See id. at 162 (Floyd, J., concurring) (citing Boumediene v. Bush , 553 U.S. 723 (2008); Hamdan v. Rumsfeld , 548 U.S. 557 (2006)); Marbury v. Madison , 5 U.S. (1 Cranch) 137, 177 (1803) (“It is emphatically the province and duty of the judicial department to say what the law is.”). If Plaintiffs’ treatment were not unlawful, any contemplation of its propriety would veer into non-justiciable territory assessing the reasonableness of sensitive military judgments.

CACI does not argue that activities satisfying the definitions of torture and CIDT used in the jury instructions could represent lawful conduct. Indeed, the jury instructions are consistent with the legal definitions of torture and CIDT surveyed by the district court, including under federal criminal code. See Al Shimari v. CACI Premier Technology, Inc. , 263 F. Supp. 3d 595, 600–04 (E.D. Va. 2017). Accordingly, we see no dispute that the Page 50instructions describe unlawful conduct. The jury’s verdict found that each Plaintiff proved CACI’s liability “for conspiring with military personnel to inflict torture or cruel, inhuman, or degrading treatment on detainees in the Abu Ghraib hard site that resulted in plaintiff being tortured or subjected to cruel, inhuman, degrading treatment.” J.A. 6399–6404. This finding necessarily includes the conclusion that Plaintiffs were subjected to unlawful treatment, which forecloses CACI’s position on justiciability.

CACI first argues under the first Taylor prong that '[t]he military's direct operational control over CACI’s interrogators requires dismissal.” Opening Br. at 38. But Al Shimari IV made clear that a necessary element of Taylor 's first prong is that the conduct directed by the military was lawful. 840 F.3d at 157. Thus, CACI’s arguments around control are insufficient because of the jury’s finding that Plaintiffs experienced torture or CIDT.

CACI's only attack on the nature of the conduct alleged is its statement that “treatment alleged by Plaintiffs as constituting torture or CIDT . . . was approvable under the [rules of engagement].” Opening Br. at 38. Even if the jury heard testimony about some treatment that on its own does not rise to the level of torture or CIDT, it is the role of the jury to sort the evidence and consider whether it was sufficient to find that the Plaintiffs were subjected to torture or CIDT as defined in the court’s instructions. That the jury might have heard about some treatment other than torture or CIDT does not convert the case from justiciable to nonjusticiable.

CACI further argues that the inquiry before us has changed from Al Shimari IV because Plaintiffs’ common-law claims and direct claims for torture, war crimes, and CIDTPage 51 were dismissed. But we think the dismissal of state tort claims reinforces our conclusion. We no longer need to be concerned that some contact during an interrogation could technically meet the requirements of battery under state tort law while not rising to the level of “unlawful conduct in violation of settled international law or criminal law then applicable to CACI employees.' See Al Shimari IV , 840 F.3d at 158; id. at 159 ('In instances in which the lawfulness of such conduct was not settled at the time the conduct occurred, and the conduct occurred under the actual control of the military or involved sensitive military judgments, that conduct will not be subject to judicial review.'). In contrast, conduct amounting to torture or CIDT is unlawful under any rubric.

We also disagree that the absence of direct claims for torture and CIDT impacts our conclusion here. CACI claims that “as this case went to the jury, there was no evidence of 'intentional,' 'unlawful' acts directed at Plaintiffs by CACI that would preclude application of the political question doctrine.” Reply Br. at 20. This argument is not persuasive. The thrust of CACI’s argument here is not that there was no evidence of unlawful conduct; rather, that the evidence did not reflect that CACI participated in unlawful conduct. But that miscomprehends the interplay between justiciability and the theory of this case, the theory being that members of the military subjected Plaintiffs to torture or CIDT pursuant to a conspiracy with CACI employees. The question of justiciability arises from whether the conduct of the military amounting to torture or CIDT was unlawful, rendering it reviewable. That CACI employees did not directly torture Plaintiffs is irrelevant to this issue. Accordingly, CACI presents no persuasive argument that Plaintiffs’ claims in this case are nonjusticiable.

CACI’s third threshold challenge in this case regards federal preemption based on the Federal Tort Claims Act (FTCA). The FTCA provides statutory authorization for plaintiffs to recover damages against the United States, providing a waiver of sovereign immunity for claims of tortious conduct by government employees. See 28 U.S.C. § 1346 (b)(1). “It exempted from this consent to suit, however, ‘[a]ny claim arising out of the combatant activities of the military or naval forces, or the Coast Guard, during time of war’”; this exception is known as the combatant activities exception. Hencely v. Fluor Corp. , 120 F.4th 412, 425–26 (4th Cir. 2024) (quoting 28 U.S.C. § 2680(j)), cert. granted , 145 S. Ct. 2748 (2025). CACI asks us to find that this exception preempts Plaintiffs’ claims under the ATS. It argues that this conclusion is compelled by Hencely , which applied the preemption test articulated by the D.C. Circuit in Saleh v. Titan Corp. , 580 F.3d 1 (D.C. Cir. 2009) (considering ATS claims against military contractors for torture at Abu Ghraib). We disagree that Hencely applies to claims under the ATS and accordingly decline to follow Saleh .

Both Hencely and Saleh derive this preemption test from Boyle v. United Technologies Corp. , 487 U.S. 500 (1988). Hencely summarized Boyle preemption as follows: 'in areas involving 'uniquely federal interests,' an FTCA exception can demonstrate ‘the potential for, and suggest[] the outlines of, significant conflict between federal interests and state law' sufficient to warrant federal preemption even absent statutory directive or direct conflict.” 120 F.4th at 426 (quoting Boyle , 487 U.S. at 504, 507, 511). Hencely , like Burn Pit and Saleh before it, 'extended Boyle 's logic to the Page 53FTCA’s combatant activities exception,” such that preemption ensues “when state tort laws would clash with the federal interest underlying the combatant activities exception.” Id. (citing Burn Pit , 744 F.3d at 350-51; Saleh , 580 F.3d at 9).

Saleh dealt with a case virtually identical to our own: Iraqi nationals brought claims under state tort law and the ATS against CACI and Titan Corp., another military contractor, “alleging that they or their relatives had been abused by employees of the two contractors during their detention and interrogation by the U.S. military at the Abu Ghraib prison complex.” 580 F.3d at 2. The court conducted a detailed analysis of Boyle preemption in relation to the plaintiffs’ D.C. tort law claims and articulated the following test: “During wartime, where a private service contractor is integrated into combatant activities over which the military retains command authority, a tort claim arising out of the contractor’s engagement in such activities shall be preempted.” Id. at 9. It concluded that application of this test preempted the D.C. tort claims at issue.

Turning to the ATS claims, the Saleh court devoted substantial discussion to the question of whether the D.C. Circuit’s precedent—“that the ATS provides a cause of action against states but not private persons'¹⁸ -remained good law after Sosa . Id. at 13. After more than three pages of analysis (wherein the court reaffirmed its precedent), the court spent a single paragraph putting forward an alternate basis from which to affirm the dismissal of the plaintiffs' ATS claims: Boyle preemption. In what is arguably dicta, given the court’s holding that the case is barred by its precedent forbidding ATS cases against Page 54private persons, the court states, 'If we are correct in concluding that state tort law is preempted on the battlefield because it runs counter to federal interests, the application of international law to support a tort action on the battlefield must be equally barred.” Id. at 16. CACI would have us use Hencely in this case to achieve the same result.¹⁹

We are not persuaded by Saleh ’s cursory contemplation of preemption as applied to claims under the ATS. Indeed, the next sentence of the court’s decision explains why:

As our Court has explained, however, the conflict between federal and state interests in this context 'is much broader' than the discrete inconsistency between federal and state duties in Boyle . 'Instead, when state tort law touches the military's battlefield conduct and decisions, it inevitably conflicts with the combatant activity exception’s goal of eliminating such regulation of the military during wartime.” In other words, when it comes to warfare, 'the federal government occupies the field' and 'its interest in combat is always precisely contrary to the imposition of a non-federal tort duty.”Page 55 'ATS [claims] would be drawing on federal common law that, in turn, depends on international law, so the normal state preemption terms do not apply.' Id. In fact, we need not force claims under the ATS into the same test used for claims under state tort law because there already exists a test for conflicts between federal statute and federal common law.

The correct preemption inquiry for ATS claims, which sound in federal common law, “is whether [a] statute ‘[speaks] directly to [the] question' otherwise answered by federal common law.” County of Oneida v. Oneida Indian Nation , 470 U.S. 226, 236–37, (1985) (quoting City of Milwaukee v. Illinois , 451 U.S. 304, 315 (1981)). “[S]tatutes which invade the common law . . . are to be read with a presumption favoring the retention of long-established and familiar principles, except when a statutory purpose to the contrary is evident.” United States v. Texas , 507 U.S. 529, 534 (1993) (quoting Isbrandtsen Co. v. Johnson , 343 U.S. 779, 783 (1952)); see also Pond v. United States , 69 F.4th 155, 164 (4th Cir. 2023) (“When a federal statute invades an area occupied by federal common law, we generally presume the statute does not change the established common law. . . . But the presumption . . . may be overcome-and the common law supplanted-when 'the language of a statute [is] clear and explicit for this purpose.'' (footnote and citation omitted) (quoting Fairfax's Devisee v. Hunter's Lessee , 11 U.S. (7 Cranch) 603, 623 (1812))).

Thus, while “an FTCA exception can demonstrate ‘the potential for, and suggest[] the outlines of, significant conflict between federal interests and state law’ sufficient to warrant federal preemption even absent a statutory directive or direct conflict,” Hencely ,Page 56 120 F.4th at 426, here we must consider only whether the FTCA itself 'clear[ly] and explicit[ly]” intends to preempt the common law, see Fairfax’s Devisee , 11 U.S. at 623. We find that it does not. Hencely crafted preemption from a statutory directive that would not, by a strict application of the statutory text, apply to the situation before it.¹²⁰ F.4th at 426 (acknowledging that '[b]y their terms, these provisions [exempting combatant activities from the FTCA’s consent to suit] do not apply to government contractors”). It had the authority for this preemption because Boyle allows it where 'uniquely federal interests' collide with state law. See id. But this permissive approach is not equivalent to the more searching inquiry in County of Oneida and United States v. Texas for circumstances of overlap between federal statutes and federal common law. Accordingly, because claims under ATS cannot be made against the federal government, see supra Part IV.A.1., there is in fact no overlap whatsoever with the FTCA, which statute’s purview is governing claims for 'damages to be recovered against the United States,' Boyle , 487 U.S. at 511. Even more to the point, the express exclusion of government contractors from the FTCA's combatant activities exception forecloses any possibility that this exception “speaks directly to” the claims in this case. Because there is no conflict between the ATS and the FTCA, the FTCA does not preempt Plaintiffs’ claims.

CACI’s final argument that the district court should not have reached the merits of this case is that the matter is “so pervaded by state secrets as to be incapable of judicial resolution.' Opening Br. at 43 (quoting El-Masri v. United States , 479 F.3d 296, 306 (4th Page 57Cir. 2007)). Our consideration of state secrets here is somewhat unusual because the trial has already occurred. Thus, a key concern of state secrets doctrine—“that public policy forbids the maintenance of any suit in a court of justice, the trial of which would inevitably lead to the disclosure of matters which the law itself regards as confidential,” Totten v. United States , 92 U.S. (2 Otto) 105, 107 (1875)—is of no moment where trial has already occurred without the revelation of any secrets. We are also not asked to weigh in on whether the invocation of privilege was procedurally or substantively proper. The only remaining inquiry is whether the impact of state secrets privilege improperly prejudiced CACI in its ability to mount a defense.

In El-Masri —a case where plaintiffs alleged they were tortured by the CIA—we specifically enumerated how defendants like CACI could be so prejudiced by state secrets that dismissal must result. In that case, we identified '[t]he main avenues of defense available”: (1) the plaintiff was not subject to the treatment alleged, (2) the defendants were not involved in the treatment alleged, or (3) defendants’ involvement does not give rise to liability. 479 F.3d at 309. We held that because “[a]ny of those three showings would require disclosure of information regarding the means and methods by which the CIA gathers intelligence,' dismissal was proper. Id. at 309, 311.

CACI invokes these 'main avenues of defense' and says that because of state secrets, it “never had a fair chance to try these issues.” Opening Br. at 43–44. We disagree.

First, we review the types of privileged information that CACI says were necessary to its defense:

Page 58(1) the identities of the Army and CACI interrogators who interacted with Plaintiffs, (2) evidence of the training and background of those interrogators . . . , (3) detailed interrogation plans showing exactly what the U.S. Army chain of command approved for each interrogation of Plaintiffs, and (4) contemporaneous reports summarizing Plaintiffs’ interrogations.

Reply Br. at 24 (citation and emphasis omitted). CACI also complains that it was prejudiced by the mechanisms used to protect the identities of interrogators—“CACI was forced to present these witnesses through nearly-useless pseudonymous telephone depositions”—and that “[s]tate secrets prevented CACI from ‘humanizing’ the interrogators who contradicted Plaintiffs’ testimony.” Opening Br. at 45–46. With this background on the relevant privileged information in this case, we consider its interaction with the El-Masri “main avenues of defense.” See 479 F.3d at 309.

State secrets did not prevent CACI from challenging whether Plaintiffs experienced the treatment alleged. In fact, CACI admits as much: 'The nine interrogators and interpreters participating in Plaintiffs’ interrogations denied such abuse occurred.” Reply Br. at 45. Two Plaintiffs testified live at trial and the testimony of the third Plaintiff was presented using his videotaped deposition. CACI had ample opportunity to question Plaintiffs and interrogators about Plaintiffs’ allegations, and it did so. Its only complaint as to this 'avenue of defense' is that it was prejudiced because its witnesses testified pseudonymously and via telephone. But this does not go to the relevant inquiry, which is whether CACI could not challenge the alleged treatment without resort to privileged information. See El-Masri , 476 F.3d at 309. Although the format of the testimony Page 59impacted CACI's defense qualitatively, 20 the existence of state secrets did not prevent CACI from arguing that Plaintiffs did not experience the treatment alleged.

Neither did state secrets prejudice CACI in the second avenue of defense: that it was not involved in the treatment alleged. CACI’s attempt to apply this avenue is somewhat attenuated given that Plaintiffs do not claim any CACI employee directly participated in the abuse they experienced. CACI argues that notwithstanding Plaintiffs’ theory of the case, “[t]he only way for CACI to fairly defend was to show that the CACI interrogators having any connection to Plaintiffs were not the few CACI interrogators implicated in discrete acts of misconduct—thus rebutting any inference of conspiracy.” Reply Br. at 25.

Page 60But this argument misapprehends the nature of a conspiracy. Indeed, following a conspiracy theory, Plaintiffs could prevail without alleging that even a single CACI employee had any contact with them whatsoever. Because Plaintiffs claim that CACI personnel were involved in the conspiratorial agreement to torture detainees for the purpose of extracting information, once a conspiracy is established, all members of the conspiracy are responsible for the acts of any member. See Pinkerton v. United States , 328 U.S. 640, 647 (1946) (“A scheme to use the mails to defraud, which is joined by more than one person, is a conspiracy. Yet all members are responsible, though only one did the mailing.” (citation omitted)). Because Plaintiffs did not need to establish 'that the CACI interrogators having any connection to Plaintiffs” were the same personnel as “the few CACI interrogators implicated in discrete acts of misconduct,' see Reply Br. at 25, CACI’s inability to disprove that notion is not prejudicial.

Finally, CACI argues that it could not “show the extent of the Army’s direction and control specific to Plaintiffs' for its borrowed servant defense because interrogation records were withheld. Opening Br. at 46. But this argument fails as well. Perplexingly, CACI argues forcefully that it was able to present a complete borrowed servant defense— so complete that based on the evidence it was able to present at trial, it claims “[t]here was no legally-sufficient basis for a properly-instructed jury to reject CACI’s borrowed servant defense.” 21 Id. at 50.

Page 61CACI points to significant trial evidence establishing its affirmative defense. It introduced as a defense exhibit one of the “Government Delivery Orders,” essentially a contract for CACI to supply civilian interrogators, which “required that CACI interrogators 'perform under the direction and control of the unit's [military intelligence] chain of command or Brigade S2, as determined by the supported command.” Opening Br. at 11 (quoting J.A. 8567). Testimony from Dan Porvaznik, CACI’s site lead at Abu Ghraib, read into the record 'a memorandum provided by the Army to CACI personnel when they arrived at Abu Ghraib' that described the chain of command, including the separation between “the operational and the administrative,” and the designation that the “interrogation section leader” reports to military officers in charge. J.A. 7250–51. CACI also introduced a redacted organizational chart showing “CACI interrogators integrated with soldiers into 'Tiger Teams' that reported to their military section leader,' then to military leadership, 'and then to Colonel Pappas, Commander of the 205th Military Intelligence Brigade.” Opening Br. at 12 (citing J.A. 8588). Another trial exhibit, the United States's interrogatory response, stated that for a particular interrogation, CACI interrogators “were subject to the direction of the military chain of command, beginning with their military section leader, an Army non-commissioned officer.” J.A. 8502. CACI further elicited testimony from multiple witnesses about the chain of command and the division of responsibilities between administrative and operational tasks, including the military’s exclusive role in developing interrogation protocols and plans. This evidence is Page 62more than enough to conclude that CACI was able to present its borrowed servant defense at trial notwithstanding the state secrets privilege.

In light of the foregoing, we conclude that CACI failed to establish that state secrets meaningfully prevented it from pursuing any of El-Masri ’s “main avenues of defense.” See 479 F.3d at 309. Where “a proceeding involving state secrets can be fairly litigated without resort to the privileged information, it may continue.” Id. at 306. This case was fairly litigated, and we see no grounds for reversal due to state secrets.²²

CACI argues that this case warrants dismissal under Wikimedia Foundation v. National Security Agency , 14 F.4th 276 (2021). It claims that the district court’s refusal to dismiss forced CACI into “a one-sided trial,” as in Wikimedia . Opening Br. at 45 (quoting Wikimedia , 14 F.4th at 304). In that case, we found that “the government’s hands are so clearly tied by state secrets[ that] it would be a mockery of justice for the court to permit Wikimedia to substantiate its claims by presenting half of the evidence to the factfinder as if it were the whole.” 14 F.4th at 304 (citation modified). But this is not a case where “[there is] simply no conceivable defense to [Plaintiffs’ claims] that wouldn’t also reveal the very information that the government is trying to protect,” see id. , and the evidence Page 63recounted above rebuts CACI’s position that the trial here was one-sided. It clearly was not.

* * *

We acknowledge a lack of cases within our Circuit defining the precise limit of this rule. El-Masri counsels dismissal where “virtually any conceivable response to El-Masri’s allegations would disclose privileged information.” 479 F.3d at 310. Wikimedia does the same when allowing the case to proceed would be so one-sided as to represent a “mockery of justice.” 14 F.4th at 304 (citation modified). In another case, we affirmed dismissal because “based on the nature of [the plaintiff’s] claims, virtually any reason the CIA could offer for its actions would require the disclosure of [privileged information],' such as “information about [plaintiff’s] performance as a covert operative, the nature of the jobs he sought, the requirements of those jobs, the job performance of his colleagues, and/or the criteria used by the CIA to make assignments.” Abilt v. CIA , 848 F.3d 305, 316 (4th Cir. 2017) (citation omitted). These cases show one end of the spectrum—where a defendant has essentially no ability to present a defense—but they do not explain at which point a case can survive the existence of privileged secrets and be “fairly litigated.” Cf. Fitzgerald v. Penthouse Int’l, Ltd. , 776 F.2d 1236, 1238 n.7 (4th Cir. 1985) (suggesting that our Circuit so far 'had no occasion to consider the question . . . [of] the manner in which a less sensitive case may proceed once certain evidence has become unavailable due to a valid assertion of the state secrets privilege”). CACI suggests that “[i]f facts important to the prosecution or defense of a claim are subject to the privilege, dismissal is required.” Reply Br. at 26. But this is clearly too low a bar and not supported by our precedent. The exact Page 64answer may be elusive, but as demonstrated by our analysis above, determining whether a party can establish the elements of a claim or defense without privileged information may guide the inquiry of future courts.

Having found each of the threshold issues CACI raised to be meritless, we now turn to the merits; specifically, whether CACI is entitled to judgment as a matter of law or a new trial on Plaintiffs’ claims and on its affirmative defense. We conclude that CACI does not make the necessary showing to reverse the jury’s verdict as to either of these issues.

CACI contends that it was entitled to judgment as a matter of law on the two causes of action, conspiracy to torture and conspiracy to commit CIDT. CACI makes two arguments: (1) that the Taguba and Fay reports (together, the “Reports”) were the primary evidence of conspiracy and should not have been admitted, and (2) without the Reports, the “remaining, admissible evidence was insufficient to support the verdict.” Opening Br. at 54.

We first consider whether the district court properly admitted the Reports, authored by Major Generals Taguba and Fay (the “Generals”). “We review a trial court’s rulings on the admissibility of evidence for abuse of discretion, and we will only overturn an Page 65evidentiary ruling that is arbitrary and irrational.” Benjamin v. Sparks , 986 F.3d 332, 346 (4th Cir. 2021). In so doing, “we look at the evidence in a light most favorable to its proponent, maximizing its probative value and minimizing its prejudicial effect.' Id.

Rule 803(8) of the Federal Rules of Evidence provides an exception to the hearsay rule for records or statements of a public office which include “factual findings from a legally authorized investigation” so long as “the opponent does not show that the source of the information or other circumstances indicate a lack of trustworthiness.” Fed. R. Evid. 803(8)(A)(iii), (B); see also Zeus Enters., Inc. v. Alphin Aircraft, Inc. , 190 F.3d 238 (4th Cir. 1999) (“The admissibility of a public record specified in the rule is assumed as a matter of course unless there are sufficient negative factors to indicate a lack of trustworthiness, in which case it should not be admitted.” (citations omitted)). Crucially, this exception provides for admission of 'the entirety of the report[,] . . . not merely its factual components,' when the Rule's requirements are met. 30B C. Wright & A. Miller, Federal Practice and Procedure § 6886 (2025 ed.) (citing Beech Aircraft Corp. v. Rainey , 488 U.S. 153, 162 (1988)). The D.C. Circuit held that in light of Beech , a public record was properly admitted under the exception notwithstanding that it was “based on double, triple hearsay and stale hearsay at that.” English v. District of Columbia , 651 F.3d 1, 7-8 (D.C. Cir. 2011).

CACI's arguments that the Reports are not admissible include complaints that admission of the Reports violates CACI's 'right to confront declarants upon whom investigative reports rely' and that '[a] report's repetition of and reliance on second-, third-, and fourth-hand hearsay renders it unreliable.' Opening Br. at 54. CACI's Page 66argument regarding the right to confront declarants is without merit, 23 so we turn to its concerns with reliability.

CACI contends that the Reports are unreliable because of their “repetition of and reliance on' hearsay, and their reliance 'on statements previously taken by the Army Criminal Investigation Division' which were made by 'people who were, themselves, under investigation” and whose credibility the Generals could not evaluate. Opening Br. at 54–55. We first reject CACI’s contention that the role of the Criminal Investigation Decision undermines reliability. Wright and Miller mention that '[c]ourts are fairly receptive to admitting outsider statements” under this exception, citing a First Circuit case that reversed an excluded report. Wright & Miller, supra , § 6888 (citing Robbins v. Whelan , 653 F.2d 47 (1st Cir. 1981)). In that case, the court considered a report by the Department of Transportation that compiled data provided by auto manufacturers. Robbins , 653 F.2d at 50. The court deemed the report admissible and rejected the defendant's argument that the report was untrustworthy because 'manufacturers may overstate the performance of their products to induce sales.' Id. at 51. Distinct from Robbins , here the Generals used data collected by others within their own organization.

Page 67CACI cites no convincing authority that Rule 803(8) does not permit different components of a large organization to contribute to a public record in an investigation, and we decline to announce this rule.

CACI's remaining complaints both speak to trustworthiness: it asserts that the Reports are unreliable because they rely on hearsay and due to the fact that many witnesses were under investigation. We are not convinced. Our Circuit’s factors to guide courts in the assessment of trustworthiness are “(1) the timeliness of the investigation; (2) the special skill or experience of the official; and (3) possible motivational problems.' Distaff, Inc. v. Springfield Contracting Corp. , 984 F.2d 108, 111 (4th Cir. 1993) (quoting Ellis v. Int’l Playtex, Inc. , 745 F.2d 292, 300-01 (4th Cir. 1984)). Courts may also consider indicia such as 'unreliability, inadequate investigation, inadequate foundation for conclusions, [and] invasion of the jury's province.' Id.

CACI's concerns could properly be factors in determining the trustworthiness of the report, see Wright & Miller, supra , § 6888, but many if not most investigative reports are created based on witness interviews, which alone is not enough to exclude a report under this hearsay exception, see Combs v. Wilkinson , 315 F.3d 548, 555-56 (6th Cir. 2002) (rejecting argument that report based on over 100 interviews was inadmissible where report drafter “lacked personal knowledge” of the events because “[u]nder [that] argument, an investigative report would never be admissible as such reports typically are not prepared by persons directly involved in the matter under investigation”); Chavez v. Carranza , 559 F.3d 486, 498 (6th Cir. 2009) (finding sufficient indicia of trustworthiness to affirm admitting report that was based on numerous witness interviews and 'thousands of Page 68complaints of acts of violence'). CACI has not established that the Reports 'merely collect[] otherwise inadmissible hearsay . . . with little value added.” Wright & Miller, supra , § 6889. To the contrary, the record shows that the Generals interviewed over 200 individuals, collected over 9,000 documents, and utilized analytical tools to collect and analyze relevant data in conducting the investigations and compiling the Reports.

The trial court is obliged to exclude all or part of a report under this exception that it finds untrustworthy. Beech Aircraft , 488 U.S. at 168. The district court in this case repeatedly found that the Reports had indicia of reliability and trustworthiness and refused to exclude the Reports in their entirety. See J.A. 1610 (indicating the court would not admit portions of the Reports without “sufficient indicia of reliability”); J.A. 4448 (finding that the Reports contain “indicia of reliability”); J.A. 5745 (“I find again that there are enough indicia of reliability.”). The district court considered multiple motions in limine regarding the Reports and specifically considered Plaintiffs’ designations of sections they intended to introduce. CACI’s counsel even commented that the district court “went through and painstakingly . . . approved or disapproved which parts of the [Reports] were coming in.” J.A. 7604. We agree that the trial court’s effort here was significant, and we find that it fulfilled its obligation to parse through the voluminous Reports and exclude inadmissible portions.

Page 69We conclude that CACI does not come close to making the requisite showing to overturn an evidentiary ruling, which we may do only if it is “arbitrary and irrational.” 24 See Benjamin , 986 F.3d at 346. There is a presumption of admissibility of records under this exception and the burden of establishing unreliability belongs to CACI. See Fed. R. Evid. 803 advisory committee notes (noting specifically with respect to reports of public entities under Rule 803(8), “the rule . . . assumes admissibility in the first instance but with ample provision for escape if sufficient negative factors are present”); Zeus Enters., Inc. , 190 F.3d at 241 (“The party opposing admission has the burden to establish unreliability.”). CACI must do more than simply ask us to “substitute [our] judgment for that of the district court” and rule in their favor. United States v. Vidacak , 553 F.3d 344, 348 (4th Cir. 2009). CACI fails to show that the district court, for example, applied the incorrect legal standard Page 70or refused to consider necessary factors. CACI cannot even say that the court ignored its arguments. See J.A. 1601–10 (expressing concern that the existence of “an inappropriate hearsay problem” could impact the consideration of whether “there’s sufficient indicia of reliability that it would not be unfair to CACI' to admit the Reports); J.A. 4448 (acknowledging “potential hearsay issues”); J.A. 5745 (noting that the Reports “are reports of findings, and the actual witnesses, to some degree, are not present to be heard from directly”). Accordingly, CACI has made no showing from which we could find that “the [district] court’s exercise of discretion, considering the law and the facts, was arbitrary or capricious,” and we conclude there is no reversible error as to the admitted Reports. See Vidacak , 553 F.3d at 348 (quoting United States v. Mason , 52 F.3d 1286, 1289 (4th Cir. 1995)) (alteration in original).

2.

CACI then asks that we reverse because Plaintiffs’ conspiracy claims were legally insufficient. We may do so only if we agree that there was no 'legally sufficient evidentiary basis for a reasonable jury” to find conspiracy “view[ing] the evidence in the light most favorable to [Plaintiffs].' FDIC v. Bakkebo , 506 F.3d 286, 294 (4th Cir. 2007).

We distill CACI’s arguments as threefold: first, that Plaintiffs put forth insufficient evidence “from which [a] jury could reasonably infer an agreement between CACI and soldiers to abuse detainees”; second, that “Plaintiffs offered no evidence” that the three CACI interrogators accused of abusing detainees were specifically part of an agreement to abuse Plaintiffs or other detainees; and third, that Plaintiffs had “no evidence that CACI, Page 71the corporation, joined any conspiracy, or explaining why joining a conspiracy would serve a corporate purpose.” Opening Br. at 52, 53, 55. None of these grounds supports reversal.

There is ample evidence in the record from which the jury could determine that an agreement existed between CACI and the military personnel who carried out the abuse.²⁵ General Taguba testified that in his investigation of Abu Ghraib, he heard from the MPs that CACI’s Steve Stefanowicz, assistant site lead, “was giving them instructions on setting the conditions, which in our turn would be negative conditions to make their interrogation successful the following day or the following night.” J.A. 6943. Taguba’s report, read into the trial record in relevant part, made a specific finding that Stefanowicz “[a]llowed and/or instructed MPs . . . to facilitate interrogations by ‘setting conditions’ which were neither authorized and in accordance with applicable regulations/policy. He clearly knew his instructions equated to physical abuse.” J.A. 8215; J.A. 6955–56. Counsel asked Sabrina Harmon, an MP deployed to Abu Ghraib in late 2003 to early 2004, whether she saw or heard Stefanowicz give instructions to another MP; she agreed that she had seen or heard Stefanowicz “give other MPs instructions.” J.A. 7043. Ivan Frederick, an Army MP in 2003, testified that his recollection was that interrogators instructed the MPs on interrogation conditions, including to “soften up detainees” by stripping them of clothing, placing them in female clothing, forced nudity, in order “to get them to talk.” J.A. 5887– 89. General Fay testified that Stefanowicz had a “very close relationship” with MPs and that Stefanowicz directed the MPs to use dogs to intimidate detainees. J.A. 5980. General Page 72Fay testified that for a specific incident, an Army Sergeant was using a dog to intimidate a detainee while Stefanowicz was upstairs giving directions to the Sergeant, likely for the purpose of “softening-up” the detainee for future interrogations. J.A. 5985–86. These facts in the record provide a sufficient evidentiary basis for the jury to conclude that there was an agreement between CACI interrogators and MPs to abuse detainees. See United States v. Dennis , 19 F.4th 656, 669 (4th Cir. 2021) (“[W]e have recognized that the jury need not rely on direct evidence but may infer conspiracy from the facts and circumstances of the case.”).

CACI next argues that it is entitled to judgment as a matter of law because the evidence did not show that the three CACI interrogators who allegedly abused detainees did so as part of an agreement to abuse Plaintiffs or detainees. But this partially developed thought lacks any citation to authority indicating the elements of conspiracy and which element Plaintiffs failed to establish. Assuming arguendo that CACI's representations accurately reflect the trial record, CACI does not explain why Plaintiffs needed to make this showing at all. Accordingly, we are not convinced this is grounds for reversal.

CACI’s final argument fails for the same reason. CACI complains about the lack of “evidence that CACI, the corporation, joined any conspiracy or explaining why joining a conspiracy would serve a corporate purpose.” 26 Opening Br. at 55. CACI again fails to Page 73cite any authority whatsoever for the proposition that Plaintiffs needed to make this showing. Indeed, its argument runs afoul of the general consensus that “[a]n employer is subject to liability for torts committed by employees while acting within the scope of their employment.' See, e.g. , Restatement (Third) of Agency § 2.04 (2006); Martin v. Cavalier Hotel Corp. , 48 F.3d 1343, 1351 (4th Cir. 1995) (“An employer is liable at common law for the wrongful acts of an employee that take place within the scope of the employee’s employment.”). CACI’s attempt to evade corporate liability is unpersuasive. Accordingly, we conclude that CACI had not put forth a proper basis from which we may reverse the jury’s verdict on the conspiracy counts.

CACI’s final attack on the merits of this case is that it is entitled either to judgment as a matter of law or a new trial on its affirmative defense under the borrowed servant doctrine. First, it argues that the district court erred in issuing a supplemental instruction following a question from the jury during deliberations. Second, it argues that the district court should have entered judgment for CACI as a matter of law on this issue. We address each argument in turn.

We perform a de novo review as to “whether the district court’s instructions to the jury were correct statements of law.” Gentry v. E.W. Partners Club Mgmt. Co. , 816 F.3d 228, 233 (4th Cir. 2016) (quoting Emergency One, Inc. v. Am. FireEagle, Ltd. , 228 F.3dPage 74 531, 538 (4th Cir. 2000)). CACI’s challenge to the district court’s supplemental instruction arose from a question by the jury during deliberations. The initial instructions told the jury to determine “under whose direction and control were the employees when they engaged in the alleged misconduct.” J.A. 7721. The jury asked, “Does control mean full control or some control?” J.A. 7746. The court provided the following supplemental instruction:

It is a question of fact that the jury must decide whether CACI had the power to control the interrogation work being performed by CACI employees — I’m sorry, CACI employees at Abu Ghraib when the alleged torture or cruel, inhuman or degrading treatment occurred. Whether the Army alone, or both the Army and CACI had this power to control, is a factual question that you must decide.

J.A. 7763. CACI takes issue with the final sentence, which it argues “incorrectly suggested that any ability to control defeats the borrowed servant defense.” Opening Br. at 48. CACI claims that the jury could have improperly inferred that CACI’s ability to hire, fire, or discipline employees was sufficient to defeat the borrowed servant defense.

CACI’s interpretation of the instruction fails linguistically, and its argument fails legally. The district court’s use of the adjective “this” to modify “power to control” refers to the previous sentence, which clarifies that the relevant “power to control” is “the power to control the interrogation work being performed by CACI employees . . . when the alleged torture or [CIDT] occurred.'²⁷ J.A. 7763. '[W]e do not evaluate a judge's instructions in isolated segments, but we look at the instructions given as a whole.” United Page 75States v. Cropp , 127 F.3d 354, 360 (4th Cir. 1997). It is unreasonable to conclude that the jury would interpret the final sentence as being divorced from the preceding sentence such that any power by CACI to control CACI employees would be sufficient for the affirmative defense to fail.

We also conclude that the supplemental instruction was a correct statement of law. Our cases and the Restatement (Third) of Agency state that the entity which “has the power to control and direct the servants in the performance of their work” is usually the relevant principal for purposes of liability. Est. of Alvarez v. Rockefeller Found. , 96 F.4th 696, 694 (4th Cir. 2024) (quoting Standard Oil Co. v. Anderson , 212 U.S. 215, 221-22 (1909)); Restatement (Third) of Agency § 7.03 cmt. d(2) ('Liability should be allocated to the employer in the better position to take measures to prevent the injury suffered by the third party. An employer is in that position if the employer has the right to control an employee’s conduct.”). The Restatement also acknowledges that courts sometimes “allocate liability to both general and special employer on the basis that both exercised control over the employee and both benefited to some degree from the employee’s work.” Restatement (Third) of Agency § 7.03 cmt. d(2). The supplemental instruction does not contradict these principles. Accordingly, we find no reversible error regarding the supplemental instruction.

CACI argues there was “no legally-sufficient basis for a properly-instructed jury to reject CACI’s borrowed servant defense.” Opening Br. at 50. Although this question is Page 76similar to our above consideration of judgment as a matter of law against Plaintiffs on their conspiracy claim, when considering entering judgment for a party with the burden of proof, the “standard is in critical respects different from and more demanding than that applicable to the grant of [judgment] against the proponent.' Allen v. Zurich Ins. Co. , 667 F.2d 1162, 1164 (4th Cir. 1982). Indeed, it is “so stringent that its exercise is but rarely appropriate.” Id. The 'body of evidence' must be considered 'not for its insufficiency to support a finding, but rather for its overwhelming effect.” Id. (quoting Mihalchak v. Am. Dredging Co. , 266 F.2d 875, 877 (3d Cir. 1959)). We must determine 'that not only was there sufficient evidence, so manifestly credible that it must be believed,” to support CACI’s contention, 'but also that there was insufficient evidence from which the jury could rationally have made any other finding.' Id. at 1165. Judgment is therefore inappropriate when “the evidence was in substantial conflict, particularly with respect to the essentially evaluative element.' Id. Because this is a conjunctive test, if we find that either there was not sufficient evidence in favor of CACI’s position or that there was enough evidence in favor of Plaintiffs’, we must find for Plaintiffs.

We first consider whether Plaintiffs put forward enough evidence for a reasonable jury to reject CACI’s affirmative defense. In accordance with Alvarez , we specifically consider whether there was sufficient evidence for the jury to find that CACI “ha[d] the power to control and direct [its personnel] in the performance of their work.” 96 F.4th at 694. This includes evaluating whether CACI or the military was “in the better position to exercise control in a manner that reduces the risk of injury to third parties.” Restatement (Third) of Agency § 7.03 cmt. d(2).

Page 77The Restatement identifies helpful factors from caselaw that show whether the court should assign liability to the general employer (in this case, CACI) or the special employer (the United States). Cases look to whether ties between the employee and the general employer “remain strong despite the employee’s emplacement in the special employer’s workplace” versus “retain[ing] only formal ties to a general employer.” Id. Liability can result for the general employer when they remain in control of screening prospective employees “to determine their general aptitude and fitness” and training “those it selects for employment.” Id. In contrast, liability can result to the special employer when the nature and duration of the employee’s relationship with the special employer “weakens the likelihood that the general employer retains any practical capacity to control the borrowed employee's conduct.' Id. Another basis to assign liability to the special employer is if “a borrowed employee’s work requires coordinated effort as part of a skilled team and close direction or supervision by the team's leader.' Id.

Substantial evidence supports the jury’s rejection of the borrowed servant defense.²⁸ Plaintiffs point to trial exhibits including CACI’s contract with the military and CACI’s code of conduct, which discuss CACI’s responsibility for supervising its personnel and affirming CACI’s accountability “for what we do.” J.A. 8107 (code of conduct); J.A. 8027Page 78 (contract). Furthermore, there is ample testimony that CACI site lead Dan Porvaznik “ha[d] the power to control and direct the [CACI interrogators] in the performance of their work.” See Alvarez , 96 F.4th at 694. At trial, Porvaznik described his role and how he carried out his duties.²⁹ He 'reviewed the qualification of CACI screeners and interrogators” and spoke to them about their qualifications “to get a feel for . . . whether or not they were good to go or what I considered to be good to go.” J.A. 7283–84. He spoke with the officer in charge at Abu Ghraib, Carolyn Wood, approximately daily about the CACI interrogators’ qualifications, including providing his opinion about whether certain CACI employees should be promoted to the interrogator role. J.A. 7286–87. He observed “quite a few” interrogations and assisted “with a few others.” J.A. 7297. He would have 'stopped interrogations in which [he] saw any type of physical abuse' because of his responsibility “as the site manager, . . . [a]s part of quality control.” J.A. 7298–99. He would sometimes review interrogation plans and reports “just to see how the guys were doing” for the primary purpose of making sure the interrogators were “performing right.” J.A. 7300–01.

Testimony from CACI interrogator Torin Nelson reinforces Porvaznik’s testimony. Nelson was on site at Abu Ghraib from November 2003 to early 2004. J.A. 6759, 6762. Nelson testified that in his understanding, Porvaznik reported “primarily to Amy Jenson,” a CACI employee based in Virginia. J.A. 6760. Nelson stated that Porvaznik and his Page 79assistant site manager, Steve Stefanowicz, were responsible for ensuring compliance with CACI’s code of conduct “on the ground at Abu Ghraib.” J.A. 6761. Nelson described some of Porvaznik’s duties as “liais[ing] on almost a daily basis with the client to make sure that their personnel are performing well to expectations, that there’s no issues that arise that the contracting company needs to be made aware of.” J.A. 6855. Nelson at one point told Porvaznik that he was contemplating reassignment or resignation, and he says that Porvaznik “asked [him] to stick it out for a couple of weeks or a few weeks to see if maybe things got better' because Porvaznik considered Nelson 'one of the most experienced guys” and said, “we really need you.” J.A. 6764. After Nelson reported concerning conduct to Army CID investigators, he spoke with Porvaznik about possible retaliatory behavior “to see if Mr. Porvaznik was in any position or condition or state of mind to be able to be of assistance to me as a CACI employee.” J.A. 6856.

This evidence provides a legally sufficient basis from which the jury could reject CACI's affirmative defense. See Allen , 667 F.2d at 1165. Considering the factors enumerated by the Restatement, the evidence and testimony support continued strong ties between CACI personnel and CACI leadership ( i.e. Porvaznik), including through Porvaznik’s ongoing role in retaining high quality employees and his monitoring of CACI personnel performance in interrogations. See Restatement (Third) of Agency § 7.03 cmt. d(2). The short timeframe between the deployment of CACI personnel to Abu Ghraib in September 2003 and the conduct at issue, occurring between October and December 2003, is not enough to 'weaken the likelihood that [CACI] retain[ed a] practical capacity to control' its employees, especially in light of Porvaznik's on-site management. See id.

Page 80Finally, even though CACI presented evidence that the interrogators were integrated into 'Tiger Teams' reporting up through the military chain of command, testimony from Porvaznik and Nelson support Plaintiffs’ theory that CACI exerted more control over both its employees and the MPs than the formal organization chart intended. See id. (suggesting that participation in a 'skilled team' may support liability for a special employer). Accordingly, CACI does not establish that “there was insufficient evidence from which the jury could rationally have made any other finding” than accepting its borrowed servant theory. See Allen , 667 F.2d at 1165. Because this alone is enough to hold that CACI does not meet the high bar for judgment as a matter of law on its affirmative defense, we need not also assess whether CACI presented sufficient evidence in support of its borrowed servant theory. See Allen , 667 F.2d at 1165.

CACI’s final challenge to the judgment awarded in this case is that it is entitled to remittitur or a new trial on damages. The jury awarded each Plaintiff $3 million in compensatory damages and $11 million in punitive damages, for a total damages award of $42 million, which the district court sustained. CACI argues that the court erred as to both compensatory and punitive damages.³⁰ We consider each form of damages in turn.

We review the compensatory damages award “for abuse of discretion, giving ‘the benefit of every doubt to the judgment of the trial judge.’” Konkel v. Bob Evans Farms, Inc. , 165 F.3d 275, 280 (4th Cir. 1999) (quoting Gasperini v. Ctr. For Humanities, Inc. , 518 U.S. 415, 438–39 (1996)). “A jury’s award of compensatory damages will be set aside on the grounds of excessiveness only if the verdict is against the clear weight of the evidence, or is based upon evidence which is false, or will result in a miscarriage of justice.” Hetzel v. County of Prince William , 89 F.3d 169, 171 (4th Cir. 1996) (citation modified).

CACI argues that “[a] plaintiff’s testimony, without supporting expert diagnosis, is 'insufficient to support a sizeable award for emotional distress.'' Opening Br. at 58 (quoting Hetzel , 89 F.3d at 171). But Hetzel does not mention any requirement of expert diagnosis; at most it points out that the plaintiff’s evidence of emotional distress resulting from employment discrimination was “based almost entirely on [the plaintiff’s] own selfserving testimony concerning stress and headaches,” which it considered “unsupported by the evidence.' 89 F.3d at 171. The Hetzel court concluded that the “thin evidence of rather limited damages' did not support the $500,000 compensatory damages award. Id.

Page 82The compensatory damages awards in this case are not grounded only in selfserving testimony describing minimal effects from emotional distress. Here, the jury found that Plaintiffs experienced torture or CIDT, the harm from which cannot reasonably be compared to harm resulting from employment discrimination. CACI does not—and could not—argue that the compensatory damages awards are against the weight of the evidence in this case. 'Pain and suffering, past and future, are obviously a reasonably certain consequence of torture.” Hekmati v. Islamic Republic of Iran , 278 F. Supp. 3d 145, 163 (D.D.C. 2017). The awards are easily within the range of other awards for similar conduct. E.g. , Xuncax v. Gramajo , 886 F. Supp. 162, 198-99 (D. Mass. 1995) (awarding compensatory damages of $1–3 million for torture and CIDT claims); Chiminya Tachiona v. Mugabe , 216 F. Supp. 2d 262, 268 (S.D.N.Y. 2002) (awarding $1 million compensatory damages for torture); Mehinovic v. Vuckovic , 198 F. Supp. 2d 1322, 1358–60 (N.D. Ga. 2002) (awarding $10 million in compensatory damages per plaintiff for, inter alia , torture, cruel and inhumane treatment, arbitrary detention, violations of the law of war, and crimes against humanity). We see no basis to find that the trial court abused its discretion in maintaining the compensatory damages awards.

We review punitive damages for abuse of discretion, Adkins v. Crown Auto, Inc. , 488 F.3d 225, 233 (4th Cir. 2007), except to the extent CACI claims they violate due process, which is reviewed de novo , Saunders v. Branch Banking & Tr. Co. of Va. , 526 F.3d 142, 152 (4th Cir. 2008).

Page 83CACI argues for reversal of the punitive damages awards on several grounds: (1) the award does not meet the standard for extending punitive damages to a corporate employer; (2) the district court erred in allowing revenue-based, as opposed to profits-based, punitive damages; (3) the judgment awards punitive damages for injuries to others, violating due process; and (4) the district court erred in failing to apply Virginia’s punitive damages cap. We address each in turn.

Plaintiffs correctly point out that CACI does not cite controlling case law on the question of corporate punitive damages. CACI invokes Virginia state law, which has no application to this case invoking a federal cause of action. Its only other authority is an employment discrimination case. But punitive damages for employment discrimination claims are subject to statutory requirements with respect to the employer’s conduct or state of mind. See Ward v. AutoZoners, LLC , 958 F.3d 254, 263 (4th Cir. 2020) (discussing Title VII’s requirements before punitive damages may be awarded). There are no similar requirements under the ATS.

CACI next argues that the district court provided an inappropriate benchmark for the punitive damages determination. Again, CACI fails to identify relevant authority, citing only cases considering state law to support the contention that it was Plaintiffs' burden to quantify CACI’s profits. Plaintiffs correctly report that in ruling on CACI’s motion in limine on damages, the district court limited punitive damages to $35 million based on the estimated revenues from CACI’s delivery orders and informed CACI that CACI could introduce evidence of actual profits to reduce that cap. J.A. 4444–45 (“And you can certainly come in . . . and show[] . . . what the actual profit was if that’s how you Page 84want to put it to the jury; all right?”). CACI does not respond to Plaintiffs’ argument or identify any place in the record where it provided evidence of its profits from the Iraq contracts. Accordingly, we find that CACI was on notice that it had the burden of proving its profits if it wanted to lower the cap on punitive damages, which it did not do.³¹ Therefore, we see no reversible error.

CACI’s third argument stems from a jury question raised during deliberations. The jury asked the following:

If the jury decides to make a punitive award, can we ask that part of the award be directed to a credible non-profit organization that provides emotional support and workforce placement for similar victims of . . . PTSD, . . . emotional trauma[,] and displacement from Abu Ghraib human rights violations.

J.A. 7775. The district court responded that it could not. Id. The district court did not disclose this question to the parties until the jury announced its verdict. CACI argues that the question indicates that the jury improperly awarded punitive damages for injuries to others in violation of due process.³² CACI contends that, if the district court had disclosed the jury's question before the jury reached a verdict, it would have requested a supplemental instruction that “due process prohibits punitive damages for injuries suffered by others.” Opening Br. at 60. But CACI does not say that the district court made an Page 85erroneous instruction; indeed, the court told the jury, “I can tell you the answer to that would be no. Punitive damages don’t work that way.” J.A. 7775. We presume that a jury follows instructions. Weeks v. Angelone , 528 U.S. 225, 234 (2000). If the jury does not “inform the court that after reading the relevant [portion] of the instruction, it still did not understand its role,” we presume it “understand[s] a judge’s answer to its question. To presume otherwise would require reversal every time a jury inquires about a matter of constitutional significance, regardless of the judge's answer.' Id. (citation omitted). Because CACI does not argue that the court made an erroneous instruction or that it ignored an express concern from the jury that it misunderstood the instruction, we do not agree that the punitive damages reflect an improper award to others.

Finally, CACI is mistaken that a state cap on punitive damages applies in this case. CACI cites Sines v. Hill , 106 F.4th 341 (4th Cir. 2024), which imposed Virginia's cap to the “historic $24 million in punitive damages” awarded to victims of the Charlottesville “Unite the Right” attack on counterprotestors. Id. at 344. But the jury in that case only reached a verdict for the plaintiffs on claims under state law. Id. at 346 & n.4. CACI did not identify, and neither could we find, any case involving only federal causes of action imposing a state damages cap. Cf. Worldwide Network Servs., LLC v. Dyncorp Int’l , LLC , 365 F. App’x 432, 457 n.1 (4th Cir. 2010) (Jones, J., concurring) (commenting that on remand, punitive damages on state tort claim “will be capped by state law,” but “there is no statutory cap on § 1981 punitives'); Alford v. Appalachian Power Co. , 951 F.2d 30, 30 (4th Cir. 1991) (“[I]n this case if Virginia civil law were to apply, . . . any punitive damage award might be subject to a statutory cap, . . . which is [not] the case under admiralty law.”Page 86 (citations omitted)). Thus, it was not error to disregard the cap here. Because we find CACI’s challenges to punitive damages on appeal are without merit, the Court concludes that the jury did not err in awarding $11 million in punitive damages to each Plaintiff.

We affirm the jury’s verdict in full. Because we must find that CACI’s third-party complaint against the United States is barred by sovereign immunity, we vacate the district court’s judgment on the third-party claims and remand with instructions to dismiss the complaint.

AFFIRMED IN PART, VACATED IN PART, AND REMANDED WITH INSTRUCTIONS.

The issue in this case is not whether interrogators committed atrocities on detainees held at the Abu Ghraib prison during the aftermath of the Iraq War. The issue is not whether CACI Premier Technology, Inc. should have known about and prevented the atrocities. The issue is not even whether there is evidence that CACI officials in the United States knew about the Abu Ghraib atrocities and failed to stop them.

No. The question we must decide is whether federal courts have subject matter jurisdiction over three alleged victims’ claims under the Alien Tort Statute for conspiracy to commit torture and conspiracy to commit cruel, inhuman and degrading treatment (CIDT) based on those atrocities at Abu Ghraib. We do not. The Supreme Court has told us repeatedly that federal courts lack jurisdiction over claims brought under the ATS where the conduct that is the focus of the statute occurred outside the United States. Here, such conduct unquestionably occurred in Iraq. Also, the Supreme Court has repeatedly cautioned federal courts against recognizing tort causes of actions under the ATS if there are good reasons to defer to the other branches of our government. Here, there unquestionably are. So, no matter how upset we might be with CACI for allowing this to happen, we lack the power to act.

This conclusion should be nothing new. The jurisdictional flaws of this case have been apparent from day one. Despite that, the district court has denied CACI’s motions to dismiss the ATS claims for lack of jurisdiction on multiple occasions, and our court has refused to consider interlocutory appeals on these issues. As a result, the case has ping- Page 88ponged back and forth between federal trial and appellate courts for almost two decades.¹ During that time, the district court recognized claims under the ATS for conspiracy to commit torture and conspiracy to commit CIDT-something that neither we nor the Supreme Court have done—and permitted those claims to go to a jury. And after a verdict for the plaintiffs, the district court rejected CACI’s subject matter jurisdiction arguments again, this time in response to its motions for judgment as a matter of law and for a new trial under Rules 50 and 59 of the Federal Rules of Civil Procedure. Under Supreme Court precedent, we should reverse the district court’s recognition of the detainees’ claims under the ATS for conspiracy to commit torture and conspiracy to commit CIDT and its decisions that federal courts have subject matter jurisdiction under the ATS of those claims.

The majority, however, goes in a different direction. Disregarding that precedent, it affirms the judgment by expanding the jurisdiction of courts in our circuit to hear and recognize claims under the ATS beyond what any court has ever allowed. To accomplish this, the majority makes four unprecedented and, in my view, dubious, moves.

Page 89First, it tries to avoid the requirement that ATS claims involve conduct that occurred domestically by declaring that the Abu Ghraib prison was effectively U.S. territory, something no court has ever done. In reaching this decision, the majority effectively holds that temporary military control of foreign soil in the middle of wartime activities renders it U.S. territory such that conduct occurring there is subject to tort liability under the ATS.

Second, in another attempt to avoid the prohibition on applying the ATS to foreign conduct, the majority declares that the presumption against extraterritoriality does not apply to claims of torture-again, something no court has ever done. In reaching that decision, it reasons that if the presumption did not apply to pirates on the high seas in the 1800s, it does not apply to torture now. But recent Supreme Court decisions make clear modern atrocities are subject to the presumption against extraterritoriality. Only by ignoring these decisions can the majority reach its desired outcome.

Third, as a backup to its first two moves, the majority concludes that even if the plaintiffs must show conduct in the United States that violated the law of nations, they have done so. The majority seems to acknowledge that the alleged torture and the alleged conspiracy to commit it all took place in Iraq. Despite that, it declares that the ATS still can be applied because CACI's domestic failures to hire and train employees and to otherwise implement protocols to prevent torture is itself a violation of the norms of international law. But the Supreme Court has rejected an almost identical argument, holding that the domestic conduct needed to apply the ATS cannot be attenuated corporate activity; it must be conduct that involves the focus of the ATS. Here, that conduct all occurred in Iraq.

Page 90Fourth, although the Supreme Court has repeatedly cautioned courts against using the ATS to create new causes of action, the majority affirms the district court’s recognition of conspiracy to commit torture and conspiracy to commit CIDT as torts that violate international law. In doing so, the majority ignores the fact that neither torture, nor CIDT, nor conspiracy to commit these acts has ever been recognized as a tort. And relying on legislative history and its own views about what would be the best foreign policy for our country, it declares the creation of this new cause of action does not create separation of powers concerns. This is the exact opposite of how the Supreme Court has told us to address this issue. Our job is not to wade into what is or is not good foreign policy— something we are uniquely unqualified to do. Our job is to consider whether creating a new cause of action might create separation of powers concerns. Applying the right test reveals several apparent separation of powers concerns. Under Supreme Court precedent, therefore, we cannot judicially create this new theory of liability.

It’s natural to share the majority’s concern over what took place at Abu Ghraib. But concerns, no matter how earnest or justified, do not permit us to act when the ATS, and Supreme Court decisions interpreting it, restrict our power. That is the case here. Because the majority disregards that binding precedent, I respectfully dissent.

Page 91Though CACI's appeal involves a myriad of issues, 2 my dissent focuses on the threshold question of federal court’s power—did the district court err in applying the ATS extraterritorially and in recognizing causes of action under it when it should have deferred to the executive and legislative branches? The answer is yes.

First, the district court’s decision impermissibly permitted the plaintiffs to apply the ATS extraterritorially. Enacted in 1789, the ATS provides that “[t]he district courts shall have original jurisdiction of any civil action by an alien for a tort only, committed in violation of the law of nations or a treaty of the United States.” 28 U.S.C. § 1350. Like other federal statutes, it is subject to the presumption against extraterritorial application. RJR Nabisco, Inc. v. Eur. Cmty. , 579 U.S. 325, 335 (2016). That means “unless a contrary intent appears, [legislation] is meant to apply only within the territorial jurisdiction of the United States.” Foley Bros. v. Filardo , 336 U.S. 281, 285 (1949); see also Morrison v. Nat’l Austl. Bank Ltd. , 561 U.S. 247, 255 (2010) (“When a statute gives no clear indication Page 92of an extraterritorial application, it has none.”). And “[w]hether the ATS bars claims related to extraterritorial conduct presents an issue of subject matter jurisdiction.” Warfaa v. Ali , 811 F.3d 653, 658 (4th Cir. 2016).

Applying the presumption against extraterritoriality involves two steps. First, courts must ask “whether the presumption against extraterritoriality has been rebutted—that is, whether the statute gives a clear, affirmative indication that it applies extraterritorially.” RJR Nabisco , 579 U.S. at 337. Absent such clear indication, step two requires courts to determine “whether the case involves a domestic application of the statute, and we do this by looking to the statute's 'focus.'' Id . 'If the conduct relevant to the statute's focus occurred in the United States, then the case involves a permissible domestic application even if other conduct occurred abroad; but if the conduct relevant to the focus occurred in a foreign country, then the case involves an impermissible extraterritorial application regardless of any other conduct that occurred in U.S. territory.” Id .; see also Nestlé USA, Inc. v. Doe , 593 U.S. 628, 634 (2021) (rejecting claim under the ATS that Nestlé aided and abetted forced labor in the Ivory Coast by providing training, funds and materials because all such alleged conduct took place in the Ivory Coast and Nestlé’s corporate presence and general activity in the United States were insufficient); Abitron Austria GmbH v. Hetronic Int’l, Inc. , 600 U.S. 412, 418 (2023) (“The focus of a statute is ‘the object of its solicitude,’ which can include the conduct it ‘seeks to “regulate,”’ as well as the parties and interests it ‘seeks to “protect”’ or vindicate.” (quoting WesternGeco LLC v. ION Geophysical Corp. , 585 U.S. 407 (2018))).

Page 93Applying the first step of the presumption against extraterritorial application analysis is easy. “[N]othing in the text of the ATS evinces the requisite clear indication of extraterritoriality” and not even ATS’s historic background can “overcome the presumption against application to conduct in the territory of another sovereign.' Kiobel v. Royal Dutch Petroleum Co. , 569 U.S. 108, 119 (2013).

Applying step two is not much harder. We look to see whether the conduct on which the ATS focuses occurred in the United States or abroad. Recall that the ATS provides that district courts have jurisdiction for torts “committed in violation of the law of nations or a treaty of the United States.” 28 U.S.C. § 1350. Thus, the conduct that involves the focus of the ATS is that which violates the law of nations. And here, the conduct that violates the law of nations is the CACI interrogators’ torture and the CIDT of the plaintiffs. All of that took place at the Abu Ghraib prison in Iraq.

This should be a straightforward case. Since all the conduct that involves the focus of the ATS took place in Iraq, the plaintiffs’ claims involve an impermissible extraterritorial application of the ATS. We have no subject matter jurisdiction over such claims. The district court should never have recognized the claims and, in fact, should have granted CACI's Rule 12, Rule 56 and post-trial motions for lack of jurisdiction.

The majority offers a series of alternative reasons the plaintiffs' claims do not improperly apply the ATS extraterritorially. First, it finds that even though the torture took place in Iraq, Abu Ghraib should be considered U.S. territory. Second, it determines that Page 94the presumption against extraterritoriality does not apply to the plaintiffs’ causes of action. Third, it holds that even if the plaintiffs have not rebutted the presumption, CACI’s conduct in the United States is sufficient. In each instance, the majority boldly breaks new ground, without even acknowledging that it is doing so. In each instance, it is also wrong. And rather than insulating each erroneous conclusion, the majority's progressively less defensible alternative rulings highlight the weaknesses of each.

The majority’s lead position is that, although the Abu Ghraib prison is in Iraq, we should consider it to be U.S. territory. As a result, the majority concludes the plaintiffs’ case does not involve the extraterritorial application of the ATS. In advancing this theory, the majority primarily relies on Rasul v. Bush , 542 U.S. 466, 480 (2004). There, the Supreme Court held that claims by Guantanamo Bay detainees alleging torture under the federal habeas statute did not involve an extraterritorial application of United States law. According to the majority, we should treat claims by the Abu Ghraib detainees the same way. It reasons that, as a matter of law, there is no reason to treat the ATS differently from the federal habeas statute and, as a matter of fact, the United States’ control over Abu Ghraib is similar to its control over Guantanamo Bay. In both instances, I disagree.

First, Rasul does not apply to ATS claims. Rasul involved the application of the federal habeas statute, not the ATS. 542 U.S. at 473. And this distinction is important. Rasul emphasized the historical tradition of applying habeas jurisdiction beyond a country’s borders. Id. at 481–82 (“Application of the habeas statute to persons detained at Page 95the base is consistent with the historical reach of the writ of habeas corpus. At common law, courts exercised habeas jurisdiction over the claims of aliens detained within sovereign territory of the realm, as well as the claims of persons detained in the so-called ‘exempt jurisdictions,’ where ordinary writs did not run, and all other dominions under the sovereign’s control. As Lord Mansfield wrote in 1759, even if a territory was ‘no part of the realm,’ there was ‘no doubt’ as to the court’s power to issue writs of habeas corpus if the territory was ‘under the subjection of the Crown.’ Later cases confirmed that the reach of the writ depended not on formal notions of territorial sovereignty, but rather on the practical question of ‘the exact extent and nature of the jurisdiction or dominion exercised in fact by the Crown.'' (footnotes and citation omitted)). In fact, the Court expressly limited its holding to the habeas statute. Id . at 480 ('Whatever traction the presumption against extraterritoriality might have in other contexts, it certainly has no application to the operation of the habeas statute with respect to persons detained within 'the territorial jurisdiction' of the United States.'). The ATS has no such historical tradition of extraterritorial application. Thus, we should not extend Rasul 's reasoning, which is rooted in the features unique to the habeas statute, to the ATS, a totally different statute that provides federal jurisdiction in limited circumstances for tort claims for money damages.

Second, the factual circumstances of the Abu Ghraib prison are nothing like Guantanamo Bay. The district court addressed this issue years ago, before the case’s third trip to our court. It found Abu Ghraib was not U.S. territory. The court pointed out that while the United States occupied Guantanamo Bay, the Coalition Provisional Authority— a multinational coalition of countries including but not limited to the United States- Page 96occupied Abu Ghraib. For that reason, the district court said the plaintiffs failed to show that the United States alone had total and complete control. It also explained that unlike in Rasul , where the United States' treaty and long-term lease with Cuba provided for “complete jurisdiction and control over Guantanamo Bay,” no such agreement existed in Iraq. J.A. 386. And it relied on another district court case rejecting the claim that the United States’ military presence in Iraq demonstrated sovereignty over that country during the same time at issue in this case. J.A. 387 (citing In re Iraq & Afg. Detainees Litig. , 479 F. Supp. 2d 85, 102–03 (D.D.C. 2007)).

Without saying so, the majority reverses the district court. It does so without identifying or applying any standard of review.³ And that issue aside, the district court's conclusion was correct. In addition to the reasons cited by the district court, the Court in Rasul applied the federal habeas statute to Guantanamo Bay because the United States' occupation was both “unchallenged and indefinite.” Rasul , 542 U.S. at 487 (Kennedy, J., concurring). In contrast, whatever role the United States had in the occupation of Abu Ghraib, it was neither unchallenged nor indefinite. It was challenged not only because Iraq did not enter a lease allowing the coalition to occupy Abu Ghraib like Cuba did for Guantanamo Bay but also because Iraqi insurgents actually fought that occupation with Page 97active military resistance. Nor was the occupation indefinite. Even the majority acknowledges that whatever control that existed only lasted about a year. There is just no comparison between Abu Ghraib and Guantanamo Bay.

I’m not the only one who thinks that. The only other case to address this situation is the Fifth Circuit’s decision in Adhikari v. Kellogg Brown & Root, Inc. , 845 F.3d 184 (5th Cir. 2017). There, a Nepali citizen who worked at the Al Asad Air Base, a U.S. military base in Iraq, and several families of Nepali citizens who had been killed while traveling through Iraq to work at the base in 2004 sued a military contractor under the ATS claiming the base constituted de facto U.S. territory. The Fifth Circuit rejected this theory, explaining that other courts have found U.S. military bases do not constitute de facto U.S. territory where the United States has not demonstrated intent to exercise sovereignty over the base permanently. Id . at 197. It also contrasted the 'unchallenged and indefinite control' of Guantanamo Bay with the more limited relationship of the military base in Iraq. Id . (quoting Rasul , 542 U.S. at 487 (Kennedy, J., concurring)).⁴ The Fifth Circuit got it right and we should not break new ground to create a circuit split.

Page 98Last, consider the implications of the majority’s view that temporary control during a military conflict turns one area into another country’s territory. Before June 6, 1944, was Normandy Beach French territory or German? On D-Day, during that monumental battle, whose territory was it? And on June 7, did it suddenly become U.S. territory? Or perhaps the territory of another Allied country? Or the Allied coalition? Under the majority's reasoning, the answers would be different each day as control shifted from Germany, to no one and then to the United States. That cannot be how we answer this question.

In sum, the majority improperly extends Rasul beyond the habeas context. But even if Rasul applies, the United States' control over Abu Ghraib was neither unchallenged nor indefinite as that decision requires. The majority also establishes an unmanageably fluid test for what does and does not count as U.S. territory. The Abu Ghraib prison is not, and never has been, effectively U.S. territory.

Perhaps acknowledging the weakness of its principal argument, the majority invokes a backup plan to avoid the presumption against extraterritoriality. Under it, even if Abu Ghraib is not U.S. territory, the presumption does not apply to torture. To reach this result, the majority points out that piracy was one of the original types of conduct Page 99contemplated by the ATS. It then explains that piracy, by its very nature, was committed not on U.S. soil, but on the high seas. The majority then concludes that torture is the modern-day equivalent of piracy. Building off those determinations, the majority declares that, like piracy, ATS claims for torture do not have to involve domestic conduct.

The majority attempts to invoke originalist reasoning in this argument, seemingly looking to how the ATS applied at the time of enactment. But for two primary reasons, its efforts miss the mark.

First, it is true that piracy is one of the three types of conduct recognized as violating international norms at the statute's enactment. See Sosa v. Alvarez-Machain , 542 U.S. 692, 721 (2004). And it is also true that generally, acts of piracy took place on the high seas. See GENERAL TREATISE OF NAVAL TRADE AND COMMERCE, AS FOUNDED ON THE LAWS AND STATUTES OF THIS REALM 253 (1738-1739) (noting that pirates lacked “any fixed Place of Residence” and “support[ed] themselves by Pillage and Depredations at Sea”). But that doesn't mean there's an exception to the presumption against extraterritorial application for ATS claims based on piracy. In fact, when Congress decided to criminalize piracy, it included language that it applied to conduct that took place “on the high seas.” 5 18 U.S.C. § 1651 (“Whoever, on the high seas, commits the crime of piracy as defined by the law of nations, and is afterwards brought into or found in the United States, shall be imprisoned for life.'). In contrast, the ATS contains no such language giving 'a clear, affirmative indication that it applies extraterritorially.' RJR Nabisco , 579 U.S. at 337. Since Page 100Congress rebutted the presumption against extraterritorial application in § 1651 but didn’t with the ATS, the presumption applies to the plaintiffs’ claims.

Second, even assuming there was an exception to the presumption against territorial application for piracy, there’s no such exception for the plaintiffs’ claims. To conclude otherwise, the majority equates piracy in the late eighteenth and early nineteenth centuries to modern day torture. It appears to believe that since the presumption did not apply to piracy (which, again, is not correct), we should also ignore it with modern day human rights claims. The majority’s argument rests on law review articles and a line of questioning by Justice Breyer during the Kiobel oral argument. That's because no case supports this theory. In fact, Supreme Court decisions foreclose the majority’s conclusion. In Nestlé , which the majority does not even cite on this point, the plaintiffs brought an ATS claim that Nestlé aided and abetted child slave labor in the Ivory Coast. But the Supreme Court rejected that claim, holding that it involved the extraterritorial application of the ATS. 593 U.S. at 634. So, if the presumption against extraterritoriality applies to slave labor, it certainly applies to torture. We must follow Supreme Court precedent, not musing from law school faculty lounges.

The majority goes on to suggest that under Kiobel , the ATS can be applied to foreign conduct that does not implicate the territory of another sovereign. This backup to its backup argument is grasping for straws. As already described, under United Nations Security Council resolutions, Abu Ghraib remained part of Iraq’s sovereign territory. And putting that aside, RJR and Nestlé preclude this argument too. Both mandate that we apply the presumption against extraterritoriality and neither provides any exception of the sort the Page 101majority tries to invoke. There is no way to square this argument with Supreme Court precedent.

In its third argument, the majority finally confronts CACI's contention that the district court improperly applied the ATS extraterritorially. Frankly, I understand why the majority waits so long to address this issue. Supreme Court decisions since the plaintiffs filed this case doom their claims.

The majority correctly acknowledges that Nestlé renders untenable our Al Shimari III decision, which previously rejected CACI's arguments that the plaintiffs' claims improperly applied the ATS extraterritorially. Despite that, the majority contends that the ATS still can be applied here because CACI’s domestic failure to hire and train employees or, once they received reports of potential misconduct, to take other domestic actions that ensured detainees would not be tortured violated established norms of international law. According to the majority, Nestlé permits this result. It doesn't.

The plaintiffs in Nestlé alleged that victims of slave labor in the Ivory Coast could maintain aiding and abetting claims, even though the slave labor occurred abroad, because Nestlé initiated payments to farms in the Ivory Coast from the United States and because U.S.-based Nestlé employees inspected activities in the Ivory Coast and reported back to other U.S.-based employees who made the decisions about payments. The Ninth Circuit agreed with the plaintiffs, but the Supreme Court rejected that decision. Nestlé , 593 U.S. at 630. It characterized the domestic conduct the plaintiffs relied on as “general corporate Page 102activity.' Id. at 634. And it then held that such general corporate activity did not support a domestic application of the ATS. To properly apply the ATS, the conduct that violates the law of nations—there the slave labor—must occur domestically. And since it didn’t, the majority reversed the Ninth Circuit. Id. at 634.

Despite that, the majority reads Nestlé 's prohibition against the consideration of general corporate activity to only apply to “conduct like the domestic location of CACI’s headquarters, the domestic contract issuance, or domestic payment processing activities.” Maj. Op. at 23. It concludes “the remaining conduct occurring within the United States— hiring, issuance of security clearances, and attempted cover up—is still properly considered following Nestlé .” Maj. Op. at 23.

Nestlé forecloses that argument. Nestlé called the payments and related decisions in that case general corporate activity that cannot support an ATS claim. The majority ignores what Nestlé actually said and reimagines a definition of general corporate activity more to its liking. We cannot do that. The corporate conduct on which the majority relies here is very similar to the conduct rejected in Nestlé . Nestlé makes clear that for federal courts to have jurisdiction over ATS claims, the conduct that violates the law of nations must take place in the United States. And the Court rejects any attempt to cut corners by using attenuated corporate activity in the United States when the direct conduct occurred aboard. The Court did that in Nestlé when the slave labor occurred in the Ivory Coast, and we must do it here when the torture occurred in Iraq.

Page 103To wrap up the extraterritorial application of the ATS issue, Supreme Court precedent forecloses the plaintiffs' claims. Abu Ghraib is not U.S. territory. The presumption against extraterritorial application applies to torture. And even acknowledging that the plaintiffs have evidence of conduct that a reasonable jury might have found to be torture and CIDT, it all happened in Iraq. So, the plaintiffs’ claims impermissibly apply the ATS extraterritorially. We have no subject matter jurisdiction over such claims. The district court should have granted CACI’s motions.

Second, the district court improperly recognized the plaintiffs' causes of action under the ATS for conspiracy to commit torture and conspiracy to commit CIDT. Those claims do not satisfy the requirements the Supreme Court has established for judicially recognizing causes of action under the ATS. So, I would reverse the district court on this issue as well.

Looking at the statute’s text, it’d be natural to question whether the ATS authorizes federal courts to recognize causes of action not created by Congress at all. The text only says that “district courts shall have original jurisdiction of any civil action by an alien for a tort only, committed in violation of the law of nations or a treaty of the United States.” 28 U.S.C. § 1350. Interpreting that language, the Supreme Court has said that the ATS “does no more than vest federal courts with jurisdiction, neither creating nor authorizing the courts to recognize any particular right of action without further congressional action.”Page 104 Sosa , 542 U.S. at 712. So, to be clear, nothing in the text of the ATS permits federal courts to invent new causes of action.

And from a separation of powers standpoint, one might come to the same conclusion. Courts apply and interpret the law. We do not make it. Why then would we ever step out of that lane to meddle in the work normally reserved for Congress?

Despite those concerns, the Supreme Court has cracked the door to judicial claim making under the ATS. In Sosa , the Court held that the 'ATS was meant to underwrite litigation of a narrow set of common law actions derived from the law of nations.” Id. at 721. Then, in Jesner v. Arab Bank, PLC , 584 U.S. 241 (2018), the Court questioned whether “the foreign-policy and separation-of-powers concerns inherent in ATS litigation” suggested that 'that a proper application of Sosa would preclude courts from ever recognizing any new causes of action under the ATS.” Id. at 265. But because the Court could conclude that it would “be inappropriate for courts to extend ATS liability to foreign corporations' in that case, it did not need to go that far. Id. Rather the Court explained that, if it was to ever recognize such a private right of action under the ATS, a two-part test must be satisfied to overcome the “general reluctance to extend judicially created private rights of action' under the ATS. Id. at 264. First, a plaintiff must establish a 'norm that is specific, universal, and obligatory.' Jesner , 584 U.S. at 258 (quoting Sosa , 542 U.S. at 732). Second, the court must decide that recognizing the cause of action under the ATS is a proper exercise of judicial discretion rather than requiring Congress to grant specific authority. Id . As explained below, applying that two-part test here requires that we reverse the district court’s orders recognizing the plaintiffs’ causes of action under the ATS for conspiracy to Page 105commit torture and conspiracy to commit CIDT and denying CACI’s motion for relief from judgment for lack of jurisdiction over those claims.⁶

At step one, the district court held that torture is a specific, universal and obligatory norm. To support this conclusion, it pointed to the 1991 Torture Victim Protection Act and the Convention Against Torture, to which the United States and other nations are a party. To the district court, this statute and treaty reveal the universal recognition of torture as a violation of the law of nations and a common definition of the term. The district court then jumped to the conclusion that if torture and CIDT qualified, then conspiracy to commit that conduct did as well.

I disagree. Supreme Court precedent focuses on whether the parties’ actual cause of action violates a specific, universal and obligatory norm. Here, the plaintiffs' cause of action is conspiracy. Even though they allege conspiracy to commit torture, the plaintiffs must show that conspiracy itself—not the underlying conduct the conspiracy is allegedly Page 106about—violates a specific, universal and obligatory norm. The plaintiffs didn’t try to do this. And the district court didn’t even address it.

To its credit, the majority at least tries. Citing to the London Charter, the Statutes of the International Criminal Tribunal for the former Yugoslavia and International Criminal Tribunal for Rwanda, the Rome Statute for the International Criminal Court and the Convention Against Torture, the majority concludes that “conspiracy to commit torture or CIDT violates ‘norm[s] of customary international law so well defined as to support the creation of a federal remedy.’” Maj. Op. at 29 (citing Sosa , 542 U.S. at 738). I think it’s fair to say those sources show that torture, at least currently, violates an established international norm. And I also think it’s fair to say that those sources support the conclusion that, again currently, all who participate in torture, as opposed to just those who directly do the torturing, can be held responsible for such conduct.⁷ But that's different from the plaintiffs’ conspiracy theory. Conspiracy is more nebulous than torture itself. Conspiracies can be shown by tacit agreements that are inferable from circumstantial evidence. See generally Halberstam v. Welch , 705 F.2d 472, 477 (D.C. Cir. 1983). This seems way less specific, universal and obligatory than torture, even if the conspiracy is to commit torture. To this point, the Supreme Court has recognized that “none of the major treaties governing Page 107the law of war identifies conspiracy as a violation.' Hamdan v. Rumsfeld , 548 U.S. 557, 610 (2006).⁸

The plaintiffs’ claims fail at step one.

B.

As to step two, I first apply it as I believe is appropriate before discussing my disagreements with the majority’s application. I then point out why text and history also show that the plaintiffs’ claims fail at step two.

At step two, we must examine whether there are compelling reasons not to recognize the plaintiffs’ claims for conspiracy to commit torture and conspiracy to commit CIDT under the ATS. As the Supreme Court explained, this general reluctance to create private rights of action applies to the ATS. Jesner , 584 U.S. at 264. 'In fact, the separation-ofpowers concerns that counsel against courts creating private rights of action apply with particular force in the context of the ATS.” Id . at 264–65. For it is the “political branches, not the Judiciary, have the responsibility and institutional capacity to weigh foreign-policy concerns.” Id . at 265 Indeed, that “the ATS implicates foreign relations ‘is itself a reason Page 108for a high bar to new private causes of action for violating international law.'' Id . (quoting Sosa , 542 U.S. at 727).

How we decide whether there are compelling reasons not to recognize the plaintiffs’ claims is important and lies at the heart of my disagreement with the majority on this issue. In describing this step, Jesner commands that 'if there are sound reasons to think Congress might doubt the efficacy or necessity of a damages remedy, . . . courts must refrain from creating the remedy in order to respect the role of Congress.” 584 U.S. at 264 (quoting Ziglar v. Abbasi , 582 U.S. 120, 137 (2017)).⁹ In other words, it is not our job to decide whether we think recognizing the plaintiffs’ claims is the best approach. Our job is to decide if Congress might-not would, not should, but might-think recognizing these claims would be a bad idea. Doing that here reveals multiple reasons we must refrain from creating the cause of action the plaintiffs urge us to recognize.

Page 109First, Congress has already exercised its prerogative to create a cause of action in this area. As already mentioned, in 1991, Congress passed the Torture Victim Protection Act.²⁸ U.S.C. § 1350 note. This statute creates a private right of action for torture victims against individuals. While it would not provide recovery for these plaintiffs against CACI, that hurts, not helps, the plaintiffs’ arguments. It shows that Congress legislated in this area and decided not to provide the type of relief the plaintiffs seek. In other words, “Congress did not want a money damages remedy” for the claims the plaintiffs assert. Tun-Cos v. Perrotte , 922 F.3d 514, 527 (4th Cir. 2019). When Congress decides not to legislate in an area, courts should not expand the contours of a federal remedy Congress elected to limit. Since Congress did not include the remedy the plaintiffs seek here when it enacted the TVPA, it is legitimate to conclude Congress didn’t think it was a good idea. S ee Saleh v. Titan Corp ., 580 F.3d 1, 16 (D.C. Cir. 2009) (“In light of the Supreme Court’s recognition of Congress’ superior legitimacy in creating causes of action . . . we note that it is not as though Congress has been silent on the question of torture or war crimes. Congress has frequently legislated on this subject in such statutes as the TVPA, the Military Commissions Act, 10 U.S.C. § 948a et seq. , the federal torture statute, 18 U.S.C. §§ 2340– 2340A, the War Crimes Act, 18 U.S.C. § 2441, and the Uniform Code of Military Justice, 10 U.S.C. § 801 et seq. , but Congress has never created this [ATS] cause of action . . . and we must assume that was a deliberate decision . . . .'). So, under Jesner , we must refrain.

Second, recognizing a private cause of action under the ATS stemming from the interrogations of detainees during an armed conflict risks interfering with the legislative and executive branches' military responsibilities. The Constitution allocates matters of Page 110foreign affairs to the branches of government the people elect. See U.S. Const. art. I, § 8, cls. 1, 11–16; art II, § 2, cls. 1–2 . Thus, as Judge Wilkinson recognized in this case’s trip to our en banc court, allowing the plaintiffs’ claims to continue “inflicts significant damage on the separation of powers, allowing civil tort suits to invade theatres of armed conflict heretofore the province of those branches of government constitutionally charged with safeguarding the nation’s most vital interests.” Al Shimari v. CACI Int’l, Inc. , 679 F. 3d 205, 225 (4th Cir. 2012) (en banc) (Wilkinson J., dissenting).

Consider just one implication of recognizing this cause of action in an area closely connected to our military and foreign affairs. If military contractors are subject to new theories of tort liability for their conduct carrying out contracts with the military, they might be less willing to enter into them. They may say it’s not worth the unquantifiable financial risk. In that case, our country will lack a tool—private contractors—that our executive and legislative branches apparently think we need. Or they may instead agree to enter such contracts but only for much higher fees. In that instance, the public fisc is depleted requiring either a reduction in expenditure elsewhere or an increase in revenues. The majority and others may think that is a price worth paying. Maybe it is. Maybe it isn’t. That’s not for us to decide. Our question is whether there is a reason Congress might think creating this remedy is a bad idea. And there plainly is. So, we must refrain. Jesner , 584 U.S. at 264.

Third, another reason Congress might think it’s a bad idea to create a cause of action for conspiracy to commit torture is that subjecting actions taken by our military and its contractors to litigation risks exposing sensitive national security information. This case Page 111illustrates that problem. To protect the state secrets privilege, the district court excluded from discovery certain materials from the United States military that related to the claims and defenses in the case and required the redaction of information from other documents. And it permitted the United States to identify several interrogators by pseudonyms. These restrictions limited CACI’s ability to utilize pertinent documents and to conduct discovery into the background of those interrogators which, it contends, would have helped it develop its defense that any abusive techniques were carried out by the military. If you assume the case is going forward, perhaps the district court struck the right balance. My dissent should not be construed as critical of that specific decision. Rather, the need for these types of limits on what would otherwise be discoverable shows the basic incompatibility of permitting litigation over military matters.

On this case’s sixth trip to our court, I expressed concern over where this case was heading. There, we dismissed CACI’s appeal of the district court’s denial of its claim of derivative sovereign immunity for lack of appellate jurisdiction. Al Shirmari v. CACI Premier Tech., Inc. , 775 F. App’x. 758 (4th Cir. 2019). While I concurred in that decision because our court’s precedent required it, I stated that “[o]ur narrow interpretation of the collateral order doctrine in this case has taken us down a dangerous road.” Id. at 760. It was apparent then that the road would be plagued by questions on how to handle “discovery into sensitive military judgments and wartime activities,” and questions about “immunity for claims that our military activities violated international norms—whatever those are.” Id. at 760-61. That turned out to be true. And it shows yet another reason we should refrain from recognizing the plaintiffs’ claims.

Page 112There are many reasons the executive and legislative branches might view creation of the plaintiffs’ causes of action as an unwanted invasion of their constitutional roles. When that is the case, whether we might agree with those reasons or disagree with them, Supreme Court precedent commands courts not to expand federal court jurisdiction. Regrettably, the majority does just that.

The majority's treatment of step two is baffling. It begins by claiming that legislative history of the TVPA indicates that Congress believed the ATS “unequivocally” secured the rights of aliens to bring torture claims. This is remarkable. For starters, using legislative history to determine congressional intent is of dubious value. A. SCALIA & B. GARNER, READING LAW: THE INTERPRETATION OF LEGAL TEXTS 369–98 (2012). There are countless reasons a member of Congress might have voted for the TVPA. What some report or some legislators may have said about a statute cannot possibly represent what the bodies as a whole thought. The best way—the only way, really—to know what Congress intended in passing a statute is to look at the words they used. See Alexander v. Sandoval , 532 U.S. 275, 288 (2001) (“We therefore begin (and find that we can end) our search for Congress’s intent with the text and structure of [the statute].”).

Besides, the legislative history the majority cites undermines, rather than supports, its theory. According to the majority, 'Congress further explained that the TVPA was necessary because of judicial concerns ‘that separation of powers principles required an explicit—and preferably contemporary—grant by Congress of a private right of action Page 113before U.S. courts could consider cases likely to impact on U.S. foreign relations.’” Maj. Op. at 34 (citing H.R. Rep. No. 102–367, pt. 2, at 3 (1991)). But if Congress thought an 'explicit' 'grant by Congress of a private right of action' was needed for the TVPA because of “separation of powers principles,” why in the world does that same concern not mandate an explicit grant for the right of action the majority endorses in this case?

The majority’s next argument is even more concerning. It reasons that “the greatest threat to foreign relations results if this case were not permitted to proceed.” Maj. Op. at 34. It then adds, “we cannot imagine a statement more offensive to the arena of foreign relations than to proclaim that courts of the United States may not provide a remedy to foreign nationals who were tortured by members of the U.S. military as part of a conspiracy to extract intelligence using universally condemned means of interrogation.” Maj. Op. at 35. With respect to my friends and colleagues in the majority, making this type of value judgment about what is good or bad for our nation’s foreign policy is not our job. The Constitution allocates these responsibilities to the President and Congress. And for good reason. We’re not qualified to make these judgments. We lack foreign policy training. And we lack the information that those members of the political branches that address these issues have. As a result, we should stay out of this lane. The majority’s willingness to wade into foreign policy debates is the exact thing the Supreme Court has told federal courts not to do. Jesner , 584 U.S. at 265 ('The political branches, not the Judiciary, have the responsibility and institutional capacity to weigh foreign-policy concerns . . . . That the ATS implicates foreign relations is itself a reason for a high bar to new private causes of action for violating international law.” (citation modified)).

Page 114Last, the majority concludes that a Bivens -like analysis is not required.¹⁰ To the majority, such an analysis is not required because in contrast to judge-made Bivens claims, ATS claims have a “firm statutory basis.” Maj. Op. at 37. That’s wrong twice over. For one, as already mentioned, the ATS doesn’t say a word about permitting federal courts to invent new theories of liability. For two, even though the majority may wish it wasn’t so, the Supreme Court has told us a Bivens -like analysis is required. In describing the required analysis at step two, Jesner quoted Ziglar v. Abassi , 582 U.S. 120, 136 (2017)—a Bivens case—to explain that “the Legislature is in the better position to consider if the public interest would be served by imposing a new substantive legal liability.” Jesner , 584 U.S. at 264 (citation modified). Jesner then followed that up with a command-again quoting Ziglar -that 'if there are sound reasons to think Congress might doubt the efficacy or necessity of a damages remedy, . . . courts must refrain from creating the remedy in order to respect the role of Congress.” Id. (citation modified).

The majority cannot pretend we are not bound by that test. But it does just that, repeating several times “that the ATS ‘was not enacted to sit on a shelf awaiting further legislation.'' See Maj. Op. at 33 (quoting Jesner , 584 U.S. at 254 and citing Sosa , 542 U.S. at 714). Cherry-picking that quote and ignoring the Court’s test that explains how the quote should be understood leads to an incorrect analysis.

Finally, although not expressly parts of Jesner 's two-part test, text and history counsel against recognizing ATS causes of action for conspiracy to commit torture and conspiracy to commit CIDT.

Leading up to the enactment of the ATS, the criminal law of England identified only three offenses against the law of nations: “violation of safe conducts, infringement of the rights of ambassadors, and piracy.' Sosa , 542 U.S. at 715 (citing 4 WILLIAM BLACKSTONE, COMMENTARIES ON THE LAWS OF ENGLAND 68 (1769)). 'It was this narrow set of violations of the law of nations, admitting of a judicial remedy and at the same time threatening serious consequences in international affairs, that was probably on minds of the men who drafted the ATS with its reference to tort.' Id.

What’s more, the text of the statute gives no indication the list of torts that violate the law of nations for purposes of the ATS was intended to be expansive. To the extent the term “law of nations” is “amorphous,” it serves as a “logical reason for section 1350’s limitation to ‘torts only.’” William R. Castro, The Federal Courts’ Protective Jurisdiction Over Torts Committed in Violation of the Law of Nations , 18 CONN. L. REV. 467, 505 (1986). “Without the limitation, the federal district courts would have been vested with a complete jurisdiction to try a broad category of private international commercial claims regardless of the amount in controversy.' Id . The ATS simply does not permit all wrongful conduct to be classified as a tort in violation of the law of nations. See Viet. Ass'n for Victims of Agent Orange v. Dow Chem. Co., 517 F.3d 104, 123 (2d Cir. 2008) (concluding that plaintiffs’ claims with respect to wartime use of Agent Orange poison as a weapon Page 116failed to satisfy the standard set forth in Sosa for recognizing a tort in violation of international law and that the district court lacked jurisdiction under the ATS to consider claims including “war crimes, genocide, crimes against humanity, and torture”); see also Abiodun v. Martin Oil Serv., Inc. , 475 F.2d 142, 145 (7th Cir. 1973) (rejecting the plaintiffs’ attempt to recharacterize a breach of contract claim as a violation of the law of nations and concluding that “[a]lthough the concept of ‘law of nations’ is an elusive one, there is nothing in the authorities cited by the plaintiffs that would support such an expansive interpretation of federal court jurisdiction under 28 U.S.C. § 1350'). Said differently, conduct that was not a tort at the time Congress passed the ATS should not be able to support a claim under the statute.

Under this approach, we must reject the plaintiffs’ claims. A survey of American law around the time Congress passed the ATS shows that neither torture nor CIDT, nor conspiracy to commit such conduct, was recognized as a tort. For example, A History of American Law , which describes torts recognized in American law at the time Congress passed the ATS, does not mention conspiracy claims. See LAWRENCE FRIEDMAN, A HISTORY OF AMERICAN LAW 222 (3d ed. 2005) (“All in all, tort law was not a highly developed field in 1776, or for a good many years thereafter. . . . Negligence was the merest dot on the law.”); see also John C.P. Goldberg, Tort Law at the Founding , 39 FLA. ST. U. L. REV. 85 (2011) (concluding that while a version of the ATS has been interpreted to grant jurisdiction to federal courts to hear suits by foreign citizens alleging violations of their human rights, “[i]t is very unlikely, however, that human rights litigation is what anyone had in mind back in 1789”). Also, my independent research has not uncovered any reported Page 117decisions adopting, embracing or recognizing the sort of conspiracy claims the plaintiffs advance.

True, the underlying conduct that the plaintiffs allege constitutes torture and CIDT might satisfy the elements of battery, which is a tort under American law and was so at the time Congress passed the ATS. See FREDRICK POLLOCK, THE LAW OF TORTS 141 (1895) (citing, inter alia, Cole v. Turner , [1705] 6 Mod. Rep. 149, and recognizing that the touching of another in anger is a battery). But the plaintiffs dropped this common law claim. The text of the ATS provides federal jurisdiction in a civil action for “tort[s] only”— not the underlying conduct that might satisfy a recognized tort somewhere.²⁸ U.S.C. § 1350. We should recognize the limitation to jurisdiction inherent in the statutory text. § 3661.1 Alien Tort Claims Act-In General, 14A WRIGHT & MILLER'S FEDERAL PRACTICE AND PROCEDURE § 3661.1 (4th ed. 2025) (recognizing that the statute's limitation to actions for torts only restricted its grant of federal jurisdiction). The fact that the ATS applies to 'tort only,' and the theories the plaintiffs pursued at trial were not recognized torts, is yet another reason the plaintiffs' claims should not have been recognized.

To wrap up, the plaintiffs’ conspiracy claims fail at step two of Jesner 's test. They are also inconsistent with the text of the ATS and with the history surrounding that statute. We should reverse the district court’s recognition of the plaintiffs’ claims and its denial of CACI’s Rule 50 and 59 motions.

Page 118It's quite natural, as humans, to feel that the interrogators who engaged in the abusive conduct at Abu Ghraib should be punished. But federal courts are not roving righters of the world’s wrongs. Under the law, not all wrongs give rise to civil liability. We operate under an ordered system of justice that only permits us to act when we have subject matter jurisdiction. And binding Supreme Court precedent makes clear that we lack it here because the plaintiffs’ claims ask us to impermissibly apply the ATS extraterritorially and to impermissibility create new causes of action. Thus, we should reverse the district court’s order denying CACI’s motions for judgment as a matter of law or for a new trial. Because the majority does not do so, I respectfully dissent.