Oral Argument Preview: When Are Companies Liable for Aiding Human Rights Abuses?

On April 28, the Supreme Court will hear oral argument in Cisco Systems, Inc. v. Doe I to decide whether two federal statutes allow lawsuits against parties for aiding and abetting violations of international law. The petition was brought by a group of Chinese Falun Gong practitioners, proceeding under pseudonyms, who allege that Cisco Systems and two of its executives helped China’s government build and refine a surveillance system used to identify, detain, and torture them in violation of human rights law. They argue that Cisco’s conduct constitutes aiding and abetting under the Alien Tort Statute (ATS) and the Torture Victim Protection Act (TVPA). Neither statute mentions aiding and abetting.

That silence is at the center of the case. The ATS gives federal courts jurisdiction over claims by foreign nationals for torts that violate international law (the “law of nations,” under the terms of the statute), and the TVPA creates a cause of action against individuals who, under color of foreign law, subject someone to torture or extrajudicial killing. Cisco argues that because neither statute expressly authorizes aiding-and-abetting claims, courts have no power to create them. The respondents argue that liability for assisting violations of international law was already recognized when both statutes were enacted, and the statutory text is broad enough to reach those who knowingly help carry out abuse.

The Court’s answer will determine whether people who are tortured or detained with the help of American technology and expertise can sue the companies and executives who allegedly made it possible, or whether those claims are available only against the foreign governments and officials who carried out the abuse directly, parties against whom suit is made difficult by foreign sovereign immunity.

More generally‚ the answer relates to the question of how abuses of human rights should be balanced against U․S․ policy in favor of holding on to commercial relations with foreign governments and exporting technologies that may be used to assist repression․ It is also worthwhile considering the cost of limiting the global participation of U․S․ firms‚ which cedes markets and technology leadership to other countries and deprives the United States of control over how these technologies evolve and are implemented. This effect might run counter to U.S. policy that, for strategic or national security reasons, may favor continued commercial engagement with foreign governments.

Background

China’s “Golden Shield” is a nationwide surveillance and censorship system that aggregates data from internet activity, communications networks, and public records to monitor the population and flag individuals deemed security threats. Respondents (and journalists) allege that the Chinese government has used the system to identify and detain members of disfavored groups, including Falun Gong practitioners.

In this case, the respondents (plaintiffs in the original case), whose names are kept anonymous, are Chinese Falun Gong practitioners. Falun Gong is a spiritual movement that has been banned by the Chinese government since 1999. The government has also characterized the movement as a political threat, in part due to its perceived opposition to Communist Party authority. Practitioners of Falun Gong have faced surveillance, detention, forced labor, torture, and incarceration. The respondents allege that the petitioners (defendants in the original case), Cisco Systems, Inc., and two individual Cisco executives, contributed to the Chinese “Golden Shield.” The respondents allege that Cisco designed and customized surveillance technology to identify Falun Gong practitioners; provided networking equipment, software, and technical support; and helped build databases that enabled tracking, identification, and arrest. According to the respondents, Cisco marketed its products as effective at suppressing Falun Gong, changed its systems to improve identification and tracking, and trained Chinese security personnel, while knowing that the technology would be used to facilitate human rights abuses. Respondents further allege that, as a result of surveillance they say Cisco helped enable, they were identified, detained, imprisoned, tortured, and subjected to forced labor.

Cisco disputes these allegations, maintaining that it sold standard, off-the-shelf networking products, did not design its technology to target Falun Gong practitioners, and complied with applicable U.S. export laws. It denies that it intended to facilitate, or knowingly enabled, human rights abuses.

Procedural Background

The suit is brought under the Alien Tort Statute and the Torture Victim Protection Act. The respondents bring ATS claims against Cisco and the individual defendants, Cisco Systems, Inc., and two Cisco executives, Fredy Cheung and James Kilpatrick, under an aiding-and-abetting theory. They also bring TVPA claims against the individual defendants, arguing that the statute extends beyond direct perpetrators to include those who assist in carrying out torture or extrajudicial killing.

Respondents first filed suit in 2011 in the U.S. District Court for the Northern District of California. The district court dismissed most of the case in 2014. The court held that the plaintiffs did not plausibly allege that Cisco acted with the required intent to aid and abet human rights violations. It also ruled that the ATS claims failed because the alleged conduct did not sufficiently involve the United States. The TVPA claims were dismissed because the plaintiffs did not adequately allege that the defendants were acting on behalf of, or exercising authority granted by, the Chinese government.

In 2023, the U.S. Court of Appeals for the Ninth Circuit issued a ruling that allowed some of the claims to proceed. Applying the framework set out by the Supreme Court in Sosa v. Alvarez-Machain, the circuit court first considered whether aiding-and-abetting liability is sufficiently well established today under contemporary international law and held that it reflects a specific and widely accepted norm. It then considered whether practical concerns, such as foreign policy implications, counseled against recognizing such claims and concluded that they did not. The court further concluded that the plaintiffs had plausibly alleged that Cisco knew its conduct would assist the alleged abuses. It also held that the plaintiffs had sufficiently alleged that Cisco engaged in significant parts of the work in question from within the United States, including designing the surveillance system, making important decisions at its U.S. headquarters, developing and testing the technology, and providing ongoing technical support.

The Supreme Court has now agreed to hear the case. It will consider whether the Alien Tort Statute permits aiding-and-abetting claims in these circumstances and whether the alleged conduct in the United States is sufficient to allow the claims to proceed.

Briefs for Petitioners

Cisco frames the case as an opportunity to shut the door on judicially created ATS causes of action, advancing three independent arguments, each narrower than the last, alongside a separate TVPA argument grounded in statutory text. The broadest asks the Court to hold that federal courts may no longer recognize any new causes of action under the ATS. The second argues that, even within the existing Sosa v. Alvarez-Machain (2004) framework, aiding-and-abetting liability fails. The third contends that the respondents’ particular claims would fail on case-specific grounds.

On the ATS, Cisco’s lead argument is that the door Sosa left “ajar,” in the Court’s word, for judicial recognition of new causes of action should now be closed. Cisco traces a line from Ziglar v. Abbasi (2017) through Egbert v. Boule (2022), importing the skepticism toward judicially implied causes of action that the Court has developed in the Bivens context. Cisco argues that under Egbert, if there is “any rational reason” to think Congress is better suited to create a cause of action, courts must decline to do so. Cisco maintains that foreign policy concerns provide the reason in every ATS case.

Cisco adds a historical argument: The First Congress enacted the Crimes Act of 1790 shortly after the ATS, authorizing criminal penalties for those who “aided and assisted” piracy while declining to create any civil aiding-and-abetting liability. Cisco describes what it calls a “doctrinal dinosaur,” pointing out that, in the history of ATS litigation, only six cases have resulted in collected monetary judgments out of approximately 300 filed.

As a fallback, Cisco argues that aiding-and-abetting liability fails at Sosa’s second step. In Central Bank of Denver v. First Interstate Bank (1994), a securities fraud case, the Court held that courts should not impose liability on those who assist wrongdoing unless Congress has expressly authorized it. Cisco contends that principle applies with equal force in the ATS context. Cisco describes aiding-and-abetting as a “cause-of-action multiplier”: The respondents assert seven theories of underlying wrongdoing, none among the three historically recognized ATS torts, and each theory would generate additional claims against anyone alleged to have assisted. At minimum, Cisco argues, the respondents’ claims would require a federal court to find that the Chinese government violated international law in its treatment of its own citizens on its own soil, a determination Sosa itself cautioned against.

On the TVPA, Cisco’s argument centers on statutory text. The Torture Victim Protection Act imposes liability on individuals who “subject” a person to torture or extrajudicial killing. Cisco argues that “subjects” means the defendant must directly cause the victim’s injury. An aider and abettor who provides assistance from a distance, Cisco contends, does not meet that standard. The TVPA’s definition of torture, which requires that acts be “directed against an individual in the offender’s custody or physical control,” reinforces this reading, Cisco contends that direct perpetrators and commanders exercise custody or control, while defendants whose involvement consists of remote assistance generally do not. Cisco argues that Central Bank forecloses implied secondary liability here as well, because the statute contains no reference to aiding, abetting, or assisting.

Brief for the United States

The United States supports reversal on both questions and endorses Cisco’s broadest argument: that federal courts should no longer recognize any new causes of action under the ATS. The government goes beyond Cisco’s brief by offering a doctrinal framework for closing Sosa’s door. In Edwards v. Vannoy (2021), the Court confronted a similar situation: An earlier decision, Teague v. Lane (1989), had left open a theoretical exception that no court ever applied. The Court eventually acknowledged the exception was “moribund” and set it aside. The government argues that Sosa’s door-ajar language has followed the same path. The Court has rejected every request to recognize a new ATS cause of action since Sosa and has “repeatedly yoked” ATS claims to its case law on implied causes of action, where it has not recognized a new one in over 45 years. The government contends the Court can acknowledge that reality without overruling Sosa outright.

The government echoes Cisco’s fallback arguments under Sosa’s second step, emphasizing three reasons Congress might doubt the efficacy of aiding-and-abetting liability: Central Bank’s requirement of express congressional direction for secondary liability, the foreign policy risks inherent in claims that require courts to evaluate the lawfulness of foreign sovereign conduct, and the TVPA’s decision to eschew aiding-and-abetting as legislative guidance that courts should not exceed. The need for congressional direction, the government contends, is “magnified in the context of the ATS,” where the question is what courts may create rather than what Congress has enacted. The government has maintained this position consistently since at least 2008, across administrations, lending it weight under Sosa’s instruction that courts give “serious weight” to the executive branch’s assessment of foreign policy implications.

The government also compares the TVPA to other federal statutes where Congress expressly created aiding-and-abetting liability. When Congress wanted to impose aiding-and-abetting liability in the anti-terrorism context, it amended the Antiterrorism Act in 2016 to add express language: “aids and abets” and “knowingly providing substantial assistance.” Congress also limited that expanded liability to cases where the primary tortfeasor is an organization already designated as a foreign terrorist organization. The TVPA contains no comparable language, and the government maintains that Congress’s demonstrated ability to create such liability elsewhere makes the TVPA’s silence dispositive.

The government also distinguishes between command responsibility, where a superior is liable for ordering or failing to prevent a subordinate’s conduct, and aiding-and-abetting. The TVPA, as the Court recognized in Mohamad v. Palestinian Authority (2012), covers the “officer who gives an order to torture,” a form of liability in which a superior is held responsible for a subordinate’s actions based on the authority relationship between them. A superior who directs the conduct of an underling, the government emphasizes, is categorically different from an aider and abettor who provides remote assistance without custody or control of the victim. Chambers and Cheung, who allegedly designed and supported surveillance technology from Cisco’s California headquarters, did not have custody or control of any respondent detained in China.

Amici for Petitioners

Amicus briefs supporting Cisco focus on three themes: the practical consequences of expanded ATS liability, the risk that a narrow holding would prove insufficient, and the absence of an international consensus supporting civil aiding-and-abetting.

The U.S. Chamber of Commerce, joined by the Business Roundtable, the National Association of Manufacturers, and TechNet, argues that ATS aiding-and-abetting liability would chill foreign investment in developing countries, disadvantage U.S. firms relative to foreign competitors who are immune from suit under Jesner v. Arab Bank, PLC (2018), and impose severe financial and reputational costs through protracted litigation. The Chamber coalition points to Talisman Energy, a Canadian oil company that divested from Sudan “shortly after the litigation was filed” despite ultimately prevailing, with Chinese companies filling the gap. The coalition also cites Al Shimari v. CAC (2013), which produced the first ATS jury verdict against an American business in 2024 after 18 years of litigation. On the TVPA, the coalition warns that aiding-and-abetting liability for individual executives would create an “end run” around Jesner’s limits on corporate liability.

CACI Premier Technology, a government contractor that has faced ATS litigation since 2008 over its provision of civilian interrogators to the U.S. military, argues that the Court should foreclose all secondary-liability claims under the ATS. CACI warns that a narrow holding limited to aiding-and-abetting would produce a “whack-a-mole” effect, with plaintiffs shifting to conspiracy, respondeat superior, or other secondary-liability theories. The $42 million Al Shimari verdict rested on co-conspirator liability combined with respondeat superior, a theory that a ruling on aiding-and-abetting alone would leave untouched. CACI also contends that lower courts have resisted this Court’s prior limitations, citing the Al Shimari district court’s conclusion that Egbert was “not material to this ATS action.”

The Washington Legal Foundation, a public-interest law firm, emphasizes that no universal standard for civil aiding-and-abetting liability exists in international law. International criminal tribunals have recognized accomplice liability, but they “cannot agree on the mens rea and actus reus required.” The Washington Legal Foundation argues that recognizing civil aiding-and-abetting under the ATS would make the statute a “global anomaly,” because no nation has adopted a general civil aiding-and-abetting statute for torts. Several foreign governments have reinforced these concerns in prior proceedings: Jordan, Canada, Indonesia, and Papua New Guinea have each objected to ATS suits involving allegations against their sovereign conduct.

Briefs for Respondents

Respondents argue that both the ATS allows for aiding-and-abetting liability and that the TVPA likewise reaches those who assist in carrying out torture or extrajudicial killing, and that the Ninth Circuit correctly allowed those claims to proceed.

Regarding the ATS claim, respondents look to the statute’s history. They argue that when Congress enacted the ATS in 1789, liability for assisting violations of the law of nations was already recognized. Looking to sources going back to the Founding era, including early legal opinions and dictionary definitions, they argue that individuals who aided offenses such as piracy or attacks on ambassadors could themselves be held responsible. They contend that recognizing aiding-and-abetting liability in this case reflects the statute’s original understanding as written, not an expansion of it.

They also argue that aiding-and-abetting liability works within the Sosa v. Alvarez-Machain framework, which remains the governing ATS framework and should not be revised. Under Sosa, courts may recognize claims based on clearly established norms of international law, such as the prohibitions on torture and extrajudicial killing. Respondents argue that responsibility for assisting those violations is itself widely accepted across legal systems. On that view, if the underlying conduct is actionable under Sosa, then knowingly helping to carry it out can be actionable as well.

Respondents also discuss the Court’s recent ATS decisions, including Kiobel v. Royal Dutch Petroleum Co., Jesner v. Arab Bank, PLC, and Nestlé USA, Inc. v. Doe, which have emphasized limits on the ATS’s reach, including a limit on cases where extending ATS liability could impact foreign policy. However, respondents argue that the case at hand falls within the set limits because recognizing aiding-and-abetting liability does not require courts to intrude on foreign affairs and can be managed through existing doctrines where appropriate. To further support this argument, they note that the United States has condemned the conduct alleged in this case, thereby reducing the risk of conflict with foreign policy interests. Furthermore, they contend that policy concerns about ATS litigation are overstated because of the small number of recent ATS cases and the aforementioned limits imposed by the Court.

Looking to the alleged facts, the respondents assert that Cisco’s role was more involved than merely providing general-purpose products. They allege that Cisco helped design and refine surveillance tools used to identify Falun Gong practitioners, provided ongoing technical support, and worked in coordination with Chinese authorities. On their account, this conduct constitutes active and substantial assistance to the alleged abuses, which they argue is sufficient to support aiding-and-abetting liability.

This naturally opens the question of whether any technological company selling services to the Chinese government opens itself up to similar potential liability under this theory. Respondents emphasize that aiding-and-abetting liability requires knowing and substantial assistance that has a causal effect on the underlying violation, arguing that “[Cisco’s] assistance provided has a ‘substantial effect’ on the commission of the violation.” They argue that the alleged conduct here meets that standard because Cisco’s role was “active, substantial, and direct,” including the customization of technology and provision of support. This kind of argument suggests a distinction between ordinary commercial activity and conduct that is tailored and knowingly facilitates unlawful acts.

Meanwhile, on the TVPA, respondents look to the statutory text and structure. The statute imposes liability on any individual who “subjects” another to torture or extrajudicial killing. Respondents argue that this language can encompass those who meaningfully contribute to the harm, not only those who directly inflict it. Separately, looking to the statute’s structure, respondents point out that the TVPA defines torture in terms of an “offender” who directly commits it, but the liability provision applies more broadly to any “individual” who “subjects” someone to torture. Respondents argue that this difference in wording suggests Congress did not limit liability to the person who physically carries out the abuse, but left room to include those who help bring it about.

Finally, respondents argue that Congress enacted the TVPA in a context where aiding-and-abetting liability was already recognized in both domestic and international law. They contend that interpreting the statute to exclude such liability would narrow the remedy Congress created.

Amici for Respondents

Eighteen amici briefs have been filed on behalf of the respondents. They fall roughly into four categories of focus: doctrine, the practical realities of modern torture, corporate accountability, and policy concerns. Although they differ in focus, all the amici generally recognize aiding-and-abetting liability as a feature of both domestic and international law, and consistent with the standing interpretation of the ATS framework.

The doctrine-focused set of briefs argued that the Ninth Circuit acted within the established ATS framework in allowing the case to proceed. International Law Scholars, the Constitutional Accountability Center, the Center for Justice and Accountability, and Scholars of Federal Jurisdiction all contend that aiding-and-abetting liability fits comfortably within the framework the Court articulated in Sosa. According to these amici, the relevant question is not whether courts should recognize a new cause of action, but whether a well-established form of liability can be applied to conduct that already violates universally accepted norms such as torture. They also emphasize that the ATS is a congressional enactment, not a judge-made remedy, and that the TVPA’s text and structure likewise leave room for liability for those who facilitate abuse.

Another set of amici, including that of the former UN Special Rapporteurs on Torture, persecuted religious minorities and their supporting organizations, and a coalition including the Center for a Free Cuba and Doctors Against Forced Organ Harvesting, focus specifically on the prohibition of torture and the realities of how such violations occur. These amici argue that torture is rarely carried out by a single actor working in isolation but depends instead on systems composed of multiple participants. With this in mind, limiting liability to direct perpetrators would leave significant gaps in accountability and undermine the prohibition’s effectiveness. They suggest that the conduct described is not so much an aiding-and-abetting liability but, rather, direct liability under the TVPA.

The next group of amici, including those submitted by Human Rights First, the Uyghur Human Rights Project and affiliated organizations, and the Electronic Frontier Foundation, focuses on the role of private actors and the implications of the case for corporate conduct (although the Uyghur Human Rights Project brief also leans into a more doctrinal discussion). These briefs contend that contemporary human rights violations generally involve complicated relationships between state actors and private entities, especially in fields such as surveillance technology. The amici argue that recognizing aiding-and-abetting liability is important to address in just such situations.

This debate is not new and has recurred throughout history. For example, IBM’s punch-card technology, supplied through its German subsidiary (Dehomag), was used by the Nazi government for census-taking and tracking populations, including Jews and other targeted groups. U.S. computer firms, including IBM, supplied systems used by the South African government, which helped the police, military, and administrative units implement apartheid restrictions. In 2021, the humanity and war crimes unit of the Paris Judicial Court indicted four executives of the French firms Amesys and Nexa Technologies for complicity in torture after selling surveillance systems to Egypt, which the Egyptian government used to identify and target dissidents. Across these varying contexts, the question relates not to ordinary commercial activity in the abstract, but commercial activity that is sufficiently tailored, integrated, and knowingly supports unlawful action.

The last few briefs examine policy and foreign relations concerns. These include a brief filed by Oxfam America and a group of victims or estates of victims of the Hamas attacks of Oct. 7, 2023. The latter argues that recognizing aiding-and-abetting liability can be managed within existing doctrinal constraints and does not necessarily create the kinds of foreign policy tensions that have concerned the Court in prior cases. As evidence for their claim, they point to limiting principles, including requirements related to domestic conduct and the nature of the underlying norm. Oxfam America also looks to economic efficiency and market incentives to argue that aiding-and-abetting liability produces positive incentives, does not harm investment, and may improve the fairness of trade.

Gaming Out Possible Rulings

If the Court reverses the Ninth Circuit, the effect would likely be to further narrow the scope of claims available under the ATS, which may limit liability to direct perpetrators, making it more difficult to bring claims against actors who allegedly provided assistance (e.g., technology or services). The reversal would likely reinforce the trajectory of recent cases emphasizing limits on the ATS, including extraterritoriality and foreign policy concerns, and may lead lower courts to dismiss similar claims at earlier stages.

Reflecting on the brief by the UN Special Rapporteurs for Torture and the complicated nature of modern torture, this could significantly limit which cases can be heard before U.S. courts under the ATS. This would especially affect corporations—reducing litigation exposure for corporations where alleged conduct is indirect or several steps removed from the underlying abuse.

However, if the Court affirms the Ninth Circuit, such a ruling would preserve (or create, depending on one’s position) a pathway for plaintiffs to bring aiding-and-abetting claims under the ATS. It would allow claims to proceed against actors alleged to have substantially assisted violations of international law, even if they did not directly commit them. The Court could also explicitly clarify that aiding-and-abetting liability fits within the framework set out in Sosa v. Alvarez-Machain. Instead of creating a new framework, it would likely leave existing doctrinal limits in place, including requirements related to domestic conduct and the nature of the underlying norm. Importantly for corporations, such a decision would preserve litigation risk, particularly in cases involving close coordination with state actors or tailored support for the good or service, as is alleged by the respondents in the current case.

The Court could also land somewhere in the middle by recognizing aiding-and-abetting liability in principle while imposing stricter limits on when such claims may proceed. For example, it could adopt heightened pleading standards or require stronger connections to domestic conduct. Such a decision would leave aiding-and-abetting claims available, but more narrowly defined. It would shift significant interpretive responsibility to lower courts, allowing for more discretion and also potential variability and uncertainty in case outcomes. This would continue the Court’s trend of narrowing ATS litigation through incremental doctrinal refinement rather than categorical rules dismissing large swaths of liability.