Editor’s Note: Part I of this series is available here. Both blogposts derive from a keynote chapter in Brianna Rosen, ed., Perpetual War and International Law: Legacies of the War on Terror (Oxford University Press forthcoming 2024) and draw upon Harold Hongju Koh, The National Security Constitution in the 21st Century (Yale University Press forthcoming 2024). 

Significantly, the last three U.S. presidents have been opposed to debilitating foreign military interventions and favored ending perpetual war. While President Obama made progress in that direction, President Trump had no visible strategy for implementing his opposition. President Biden made ending the “Forever War” a hallmark of his campaign and has taken concrete steps to implement that goal. To finish the job, we should not dismiss these initial steps, but rather call for building on where we are, using four “micro-reform strategies” over time.

I. Seizing the Inch

More than three decades ago, I urged Congress to enact a comprehensive charter of national security reforms. If political will could be mustered for such sweeping reforms, recent stalemated bipartisan legislative proposals — like the Senate’s National Security Powers Act and the House’s National Security Reforms and Accountability Act (analyzed in greater detail here, here, and here) — would lay the necessary foundation for broader structural changes to rebalance the institutional allocation of war-making powers. Until then, the United States should “seize the inch,” and embrace changes by all three federal branches that could be implemented either collectively or piecemeal over time. A pragmatic approach would focus more on what is genuinely broken than on areas where the Obama/Trump/Biden administrations have endorsed, but not fully implemented, the right principles. Securing and building on piecemeal reforms when they can be achieved will lead to progress, while staying transparent and candid about the moral hazards involved in living with an unsatisfactory and unstable status quo.

Let’s start with ending the continuing tragedy at Guantánamo. Thirty detainees remain, twenty of whom have long been cleared for transfer to willing recipient countries, despite Obama’s and Biden’s repeated, vocal pledges to close the prison. Perversely, the need to claim that “law-of-war detainee” status justifies keeping Guantánamo open has itself become an obstacle to removing the armed conflict framework. This is backwards: the war was supposed to justify long-term detention without charges, not the other way around. A painfully unwieldy interagency bureaucracy now determines who can be released from Guantánamo, with virtually every agency participant holding a de facto veto. Consequently, even when nearly every knowledgeable actor agrees that a detainee no longer presents a current threat, almost anyone who fears being blamed for that detainee’s release can block it from happening.

The Biden administration must devote more diplomatic resources and incentives, driven from the White House, to persuade foreign countries to take the twenty who have been cleared for transfer. A U.N. Special Rapporteur has issued a blistering report that should spark concerted action from the Biden administration. Amid other daily pressures, it must be the job of someone senior in the White House — preferably the Deputy National Security Advisor for Homeland Security — to ask, every single day: “And what did we do today to close Guantánamo?” Many of those remaining are so old and sick, physically and mentally, that they presumably could be released on humanitarian grounds without threat to the public. Because a recent ruling by a military judge bars convictions based on evidence derived from “voluntary” confessions made after torture, the Chief Prosecutor of Military Commissions should negotiate plea bargains based on time served with the detainees’ pro bono counsel. Once the remaining numbers are cut by half, President Biden could simply announce the closure of the offshore prison and bring the remaining detainees out on a single airplane for intermediate detention on the continental United States and eventual release. President Obama made a related claim of executive authority to release six Taliban prisoners in exchange for Bowe Bergdahl in 2014. For all its saber-rattling, Congress’s response was mild, and no legislation was enacted in protest.

To wind down the conflict, the first and most obvious legislative step is to repeal or replace the 1991, 2001, and 2002 Authorizations for Use of Military Force (AUMFs), which have become principal facilitators of the Forever War. These statutes have no sunset clauses and are elastically worded. Accordingly, so long as the “enemy” keeps mutating, government lawyers will try to extend these force authorizations to new enemies and situations never considered by the Congresses that enacted them nor affirmatively endorsed by any Congress since. And so, the vicious cycle continues: if we keep acquiring new enemies and can fight them under the same laws, we will never repeal those laws and will instead use the laws to extend warfare. But if we are always fighting new “Associated Forces,” there will always be new enemies and the war will never end. So we cannot end the Forever War, unless we end forever-war legislation, in a variety of military theaters, such as Iran and Yemen. President Biden’s legislative successes in his second and third years offer some hope that under the right circumstances, war powers reform legislation of the kind proposed by Senators Kaine, Merkley, Young, Lee and Paul could all be passed (discussed here and here). Promising legislative proposals exist, missing only the necessary congressional majorities.

Under international law, we should demand greater rigor for the interpretation of jus ad bellum relating to armed attacks, self-defense, collective self-defense, and state consent. Article 51 of the U.N. Charter speaks of an “inherent right of individual or collective self-defense if an armed attack occurs against a Member of the United Nations,” but leaves open many questions that demand doctrinal clarification through state practice and legal declaration. Expansion of the concept of self-defense has come from: using the “unable or unwilling” test as a proxy or workaround for state consent; growing pressure to relax the “imminence” requirement with respect to senior operational leaders; and increasing elasticization of the concept of collective self-defense to address “over-the-horizon” threats.

Although the “unable-or-unwilling” test remains controversial among international law scholars, it has proven less controversial among leading states, and it is unclear what a better alternative would be. If the unable-or-unwilling test is invalid, its critics should clarify what better legal rule should take its place. There is growing acceptance that the traditional Caroline test for “imminence” has been overtaken by technology and the need for immediate response, at least in the counterterrorism domain. But considerable debate remains about how imminent the next armed attack must be when the threatening actor has a documented history of armed attacks. And there is comparable acceptance that the doctrine of “preemptive self-defense” pressed by Presidents G.W. Bush and Trump cannot be squared with the U.N. Charter or international humanitarian law (IHL), which should both be affirmed by strengthening and reissuing the assassination ban, Executive Order 12333.

Diligent oversight can avoid the elastic application of legal standards in this area by retaining these legal standards, while demanding facts and assigning a rigorous burden of proof to those who use force. When, for example, in 2020 the Trump administration claimed it had killed Iranian Maj. Gen. Qassem Soleimani in “elongated self-defense,” Congress, U.S. allies, and the public all appropriately pressed the administration for its factual basis for the claims of necessity, proportionality, sovereignty, and imminence: in particular, that killing Soleimani prevented an unavoidable and imminent attack on U.S. territory or critical American interests. In response, the Trump administration failed to share persuasive evidence to support its claim that the strike was aimed at disrupting ongoing attacks, leaving little doubt that the attack was, in fact, illegal.

Under domestic law, the Obama-Biden legal theory concerning the use of force has acknowledged that war-making must follow Article II of the U.S. Constitution, the statutory AUMFs, and the policy guidance on direct action – the latest iteration of which is the Presidential Policy Memo (PPM). State and Justice Department lawyers should work to clarify that the international law standards that must be internalized into the AUMFs includes international human rights law. And the Biden administration should reissue and strengthen Obama’s Executive Order 13732, which, among other things, required all U.S. government agencies to publicly disclose the number of civilians killed in drone strikes, particularly outside of “areas of active hostilities.”

If the whole world is not a battleground, then greater precision is needed to define exactly where the zones outside of “hot” battlefields or “areas of active hostilities” are to which the PPM rules should apply. The key question here is operational: exactly where should military commanders have broad discretion, without prior political approval or civilian oversight, to decide when and where to attack? As Franklin Foer has recently reported, the inevitable “dynamic . . . that kept the United States entangled in Afghanistan [was that p]oliticians who hadn’t served in the military could never summon the will to overrule the generals, and the generals could never admit that they were losing. So the war continued indefinitely, a zombie campaign. Biden believed that he could break this cycle,” which the president correctly saw as necessary to ending the Forever War.

Finally, some U.S. laws, such as the 1973 War Powers Resolution, will become increasingly obsolete. Apart from Russia’s invasion of Ukraine, large-scale manned interventions like Vietnam or Operation Desert Storm will become rare. As armed conflict is conducted not by massive air strikes or ground invasions but by the sporadic use of drones, artificial intelligence (AI), short-term special operations, and cyber tools, the War Power Resolution’s blunt durational limit will become entirely ineffective in forcing thoughtful interbranch dialogue about the goals of military operations.

Any new War Powers Resolution should create a consultative congressional group or Joint Committee on National Security akin to the Joint Committee on Taxation, that would become a core group of expert members on national-security matters and war powers. Given that members of Congress now experience far greater longevity in office than executive-branch officials, the core legislative group would likely come to gain greater perspective and expertise than its executive-branch counterparts. Just as earlier presidents have selected secretaries of state and defense from the Senate—such as John Kerry, Hillary Clinton, Edmund Muskie, William Cohen, and Chuck Hagel — future presidents could draw upon this core group for future cabinet officers, bringing the two branches ever closer together.

Second, by requiring the president to consult regularly with that core consultative group, a new war powers bill should directly foster interbranch dialogue and attempt to equalize access to sensitive information that would otherwise lie exclusively within the president’s control. Thicker consultation requirements would promote congressional involvement earlier in the decision-making process so that the president could not simply commit troops first and then present Congress with a fait accompli. Third, any new war powers bill should also grant individual members of Congress statutory standing to seek judicial enforcement of the act in the D.C. federal courts, which would develop into a judicial counter-arena of legal expertise in that circuit with regard to war powers. As John Hart Ely wryly observed, the same courts that have claimed separation-of-powers or judicial-incompetence reasons for refusing to decide whether “hostilities” exist for the purpose of triggering the War Powers Resolution have routinely decided whether hostilities exist in insurance-contract cases.

Simply put, executive practice has gained undue predominance as the prime source of customary constitutional law in the area of war-making. Over time, we must restore a process within which all three branches work together: a presidency that will accept internal and external constraints; a Congress that will enact laws clarifying institutional responsibilities; and courts that will fulfill their duty to declare conduct illegal.

II. Moving Their Food

As difficult as it is to herd cats, you can still change their incentives by moving their food. The same reasoning applies to institutional reform: in the long run, the best way to change institutions is to alter organizational (and individual) incentives, reduce their isolation, and encourage interbranch dialogue while national security objectives are being set and implemented. By doing so, even piecemeal reform could modify normative expectations within foreign-policy institutions, reduce cynicism about the possibilities for balanced institutional decision-making, and generate a more virtuous policy cycle over time.

First, and most obviously, the executive branch should incorporate more reliable mechanisms to ensure internal checks and balances. Three changes should be made to ensure that the president receives loyal but impartial national-security legal advice. First, the White House counsel’s office should shrink in size, and the National Security Council (NSC) legal advisor should no longer function as a “double-hatted” deputy White House counsel vulnerable to political influence. Instead, the NSC legal advisor, advised by a group of foreign-relations lawyers from the general counsels’ offices of the participating agencies, should give independent advice directly to the national security advisor and deputy.

Second, a formal executive order should make permanent the requirement of the type of interagency lawyers group convened during the Clinton, Obama, and Biden administrations, but truncated in the G.W. Bush era (especially his first term) and abandoned in the Trump era. As in earlier times, the NSC legal advisor’s office should rarely generate its own written opinions; instead, its role should be to discern consensus, break ties, and derive the best advice available from the detailed analyses of the expert agencies. To ensure that the United States meets its “duty to explain,” the participants in the interagency legal process must press the NSC legal advisor to ensure that public written rationales are offered for all major legal decisions.

Third, we should return to the more efficient and effective system that functioned in the 1980s, when most detailed foreign-affairs legal advice came from two large, established foreign-policy offices with expertise in both foreign affairs law and international and humanitarian law: the Legal Adviser’s Office at the State Department and the General Counsel’s Office at the Defense Department. State Department review of executive initiatives in foreign affairs, for consistency with international law, would be fully consistent with statute and historical tradition. In most of the interagency national-security meetings that I attended since 2009, the Justice Department input came not from the Office of Legal Counsel (OLC), but from the National Security Division. Many of the OLC opinions generated since the G.W. Bush administration greatly exaggerated the breadth of presidential power in foreign affairs, and should accordingly be systematically reviewed for their correctness, published if believed correct, and withdrawn when found incorrect — e.g., the 2020 Soleimani opinions. Indeed, some of these were apparently inappropriately generated as post-hoc justifications for controversial national-security decisions.

The top priorities for congressional reform should be building the core consultative group within Congress described above; building a central repository within Congress of legal advice regarding international and foreign-relations law; and developing better tools to counter executive overreach.

To obtain better legal advice, Congress should create, as legal counsel to this consultative group, a central legal staff headed by a Congressional Legal Adviser, corresponding to the Legal Adviser’s Office in the State Department or OLC in the Justice Department. The office would also be comparable to the Congressional Budget Office (CBO) in the sense of independently monitoring and analyzing the legal output of the executive branch’s national-security apparatus. The congressional legal adviser could be an elected officer of the House — like the clerk or the parliamentarian — or an attorney agreed upon by a consensus decision of both parties, similar to the director of the CBO. The congressional legal adviser’s task would be to coordinate the work of the staff counsels to the various international-affairs committees, to act as a liaison between executive legal staffs, and to brief and advise the core consultative group on difficult legal questions. By internal rule, Congress could authorize and require the congressional legal adviser to issue a “counter-report” — analogous to the independent CBO cost assessments of proposed administration programs currently provided to the foreign-affairs committees — whenever the executive branch transmits a declaration of national emergency, a war-powers report, or an intelligence finding. These counter-reports could then be subject to an immediate congressional vote (under the fast-track joint resolution of approval procedure described in greater detail below), which would establish a contemporaneous written record either accepting or rejecting the president’s legal justification, against which any future executive claim of congressional acquiescence could be immediately tested.

Finally, Congress should strengthen its toolkit by developing a constitutional substitute for the legislative veto and making more effective use of its appropriations power. One alternative that would largely preserve the veto’s beneficial political compromise, without its accompanying disadvantages, is the so-called “fast-track legislative approval procedure.” Unlike the so-called legislative veto (which is not legislation jointly voted by the two houses and signed by the president), fast-track approval is a greatly expedited legislative process, found most prominently in the trade laws and discussed at length elsewhere. Under this procedure, Congress may authorize the president to initiate military action in exchange for a commitment that the president will submit the product of that action to Congress for final approval or face immediate appropriations cutoff. Under modified House and Senate rules, Congress “promises” the president that it will require automatic discharge of the completed initiative from committee within a certain number of days, bar floor amendment of the submitted proposal, and limit floor debate, thereby ensuring that the package will be voted up or down as is, within a fixed period. If triggered by objectively determinable circumstances — for example, “exchanges of hostile fire” — fast-track procedures would force Congress to vote on each particular troop commitment and thereby make a swift specific public judgment approving or disapproving each proposed presidential military action by action, not silence (as it currently does under the War Powers Resolution).

III. Where the Puck Is Going

Hockey star Wayne Gretzky once explained his success by saying, “I skate to where the puck is going to be, not where it has been.” As technology evolves, a state of perpetual conflict may strike policymakers as the best way to deal with future threats, thus perpetuating the Forever War. Where the puck seems to be going is “video game wars”— that is, cyber-assisted AI fighting other AI, with much if not most kinetic work being done by “expendable mercenaries” such as the Wagner Group. This mode of fighting would appear to reduce the financial and physical risks that aggressor states face, while leading to increased civilian harm.

To counter this trend, three reforms are urgently needed. First, emerging technology in the form of more advanced drones and increasingly autonomous weapons systems must be strictly regulated. Doing so requires significantly strengthening the laws of LAWS (Legally Autonomous Weapons Systems) at both the international and the domestic levels. Second, domestic and international laws concerning cyber conflict must be clarified, closing loopholes where they exist. Finally, private security contractors and mercenaries must be under proper government oversight and subject to all applicable laws governing armed conflict and force short of war.

President Obama’s 2013 speech at the National Defense University acknowledged that if Congress saw the need for judicial oversight of drone warfare as a more effective check on executive conduct than committee oversight, it could create a court to review targeted killing ex post, along the lines of the Foreign Intelligence Surveillance Court. Congress could require intelligence officials to make ex parte, in camera submissions of their evidence and justifications for kinetic drone action, giving the court discretion to deny that submission and order the release of the classified information to the oversight committees. And like reforms to the FISA Court, the system could include designated individuals with security clearances as amicus to present factual and legal arguments in opposition to the government’s position when needed. Over time, such a court could become an expert referee between the intelligence agencies and the committees, while reducing the likelihood of leaks and offering greater protection of classified information.

In the emerging fields of cyber conflict, AI, and private security contractors, the executive branch should continue to translate into codes of conduct the best practices currently being identified by such public and private lawmaking forums as the Group of Government Experts (GGE), the Open Ended Working Group (OEWG), the Tallinn Manual, and the Oxford Process on International Legal Protections in Cyberspace. Rules that are being negotiated have already become important reference points for U.S. interagency and allied legal discussions. With respect to Private Security Offensive Actors (PSOAs), public-private arrangements like the Montreux Document — a non-binding statement of principles and practices approved by 58 states participating, including the United States, European Union, NATO, and the Organization for Security and Co-operation in Europe— have produced an International Code of Conduct (ICoC) to govern the private security practices of more than five hundred global corporations. That code of conduct can be internalized into private behavior through contracts that govern the conduct of private security contractors, as well as domestic criminal and contract law.

A global consensus seems to be emerging in these various forums that relevant governmental personnel must require appropriate levels of human judgment in the development, deployment, and use of such AI for military purposes. The U.N. Conference on Conventional Weapons in Geneva has stated general principles for maintaining meaningful human control in compliance with IHL and establishing a responsible chain of command with humans “in the loop,” which tracks with a recent United States’ political declaration on the matter. The primary danger here is that advanced weapon systems, once activated, can —independently of human interference or supervision—acquire and engage targets, and adapt to a changing environment. To be clear, the problem is not so much the technology itself, but rather the exclusion of human beings from algorithmic decisions to engage in offensive selection of targets that are then attacked with lethal force. Even with full human control, however, special care still must be taken to ensure the underlying algorithms are unbiased and that AI power is used in transparent, just, and equitable ways.

Fully autonomous robots that do not have a human operator “in the loop” when they select, acquire, and engage targets should be outlawed as illegal weapons of war. The laws of war do not yet explicitly prohibit use of such weapons even as they are being widely developed. To the extent that many existing weapons are only semi-autonomous, their human operators can program in controlling principles analogous to those stated in the PPM to foster compliance with IHL, IHRL, and domestic law. Following the Montreux model for private security contractors, the president could issue an executive order, to be reinforced later by statute, directing purchase-contract terms and updating policies, contracts, and regulations to ensure producer liability for legal violations by “killer robots.” As these standards mature and crystallize, they too can be embedded into executive orders and statutes, and eventually multilateral legal agreements among States.

Even if all of this were done, however, we would not be spared hard moral choices. As the Afghanistan withdrawal painfully illustrated, withdrawing from a war theater can create significant risks of egregious human rights violations, the resurgence of terrorism, and the betrayal of loyal allies and local actors. When brutal dictators like Syrian President Bashar al-Assad commit widespread war crimes, would humanitarian intervention be illegal under all circumstances, or — as I have argued — a policy option legally available under extreme, carefully defined conditions?

Ending the Forever War would hardly eliminate painful moral dilemmas. As the book-cum-movie Oppenheimer reveals, leaders decide almost every day that this group of lives is more important to them than others: for Truman, for example, American lives over civilian lives in Hiroshima and Nagasaki. Inevitably, these horrific choices will be made, but at least we want them made by human beings who face political accountability and who will bear the moral and legal consequences when they are wrong.

IV. Changing the Light Bulb

How many psychiatrists does it take to change a light bulb? The answer, of course: “The Light Bulb Must Want to Change.” The same goes for the national security bureaucracy: when its hammer is waging war against terrorism, everything looks like a nail. Executive-branch reforms must address the growing autonomy of the national-security bureaucracy. On the one hand, bureaucratic inertia frustrated many extreme political initiatives during the Trump era; on the other, during the Biden era, it has stymied reforms such as closing Guantánamo, reversing draconian immigration policies, and issuing stricter guidance for drone strikes. Even if today’s sprawling national-security agencies cannot be quickly downsized, they must be better shaped and their legal marching orders clarified.

The main challenge here is for conscientious political appointees to take more deliberate, longer-term steps toward what Rebecca Ingber calls “intentional bureaucratic architecture,” to promote better national-security governance. Successive political leaders of cabinet agencies have found it far easier to answer today’s short-term fire drills, leaving untouched the longer-term specification of how the bureaucracy should give, receive, and incorporate foreign-affairs legal advice. This has led to a shared bias among both political and career executive-branch lawyers in favor of legal positions that preserve and extend the impulse toward executive war-making.

Citing internal executive decisions as stare decisis (standing precedent), the bureaucracy then entrenches even legal decisions, including many made in the G.W. Bush and Trump administrations that were reached outside a well-functioning interagency legal process. New political appointees to agency general-counsel positions should press for the internal overruling of defective prior positions and demand change when bureaucracies seek to preserve, on grounds of executive stare decisis, past legal analyses “that are not reached through cautious, deliberative, forward-looking processes,” and are influenced by partisan politics that disconnect expertise from authority.

Most fundamentally, comprehensive public education is needed to demonstrate why perpetual war results in diminishing returns for national security and — unlike diplomacy and dispute resolution — inevitably leads to crushing debt and national overstretch. In the end, this is about persuading a rising generation that Forever War is an unacceptable default for a country committed to peacetime global leadership. In the long run, our best hope for ending the Forever War must be to win the hearts and minds of the young people who do not want to keep fighting it.

IMAGE: Military equipment at sunset (via Getty Images).