| Trae Stephens is the executive chairman and a co-founder of Anduril Industries and a general partner at Founders Fund. Katherine Boyle is a general partner and co-founder of Andreessen Horowitz’s American Dynamism initiative. How should the United States use artificial intelligence to deter war, and if necessary, wage war in a just and moral manner? There is deep disagreement on this question. Employees at Google and Microsoft protested AI projects with military customers long before Claude and ChatGPT were in everyone’s pockets. Anthropic is now engaged in its version of this dispute with the Pentagon, opposing how the technology might be used in combat operations. These concerns are understandable. No one should want a future where automated armies — without human oversight or decision-making — determine war and peace. But one must face the world as it is. The U.S. is in a technological arms race with its adversaries, and AI is the cornerstone of that race. For companies and individuals working to build the next generation of American military power, AI and autonomy are not abstractions; they are decade-defining innovations that will shape whether future wars are deterred or fought. The debate is not new. Last month, Pope Leo XIV, invoking centuries-old Catholic teachings about “just war” theory, warned of the dangers of a “technological arms race.” Can Christian ethics be reconciled with this new geopolitical reality? We both take this question seriously. One of us is a defense technology founder and a Christian. The other is an early investor in several defense technology companies and a practicing Catholic. Informed by our work and religious principles, we believe there is a moral argument for the U.S. to create a superior technological arsenal first: to wage wars of precision rather than wars of indiscriminate strikes, and to protect human life, civilian and combatant alike. Just war theory, like any enduring operating system, can be refreshed for new conditions. But an update is not an abandonment. Hypersonic missiles do not look like King Henry V’s bows and arrows at the Battle of Agincourt. But the use of arms must still rest on enduring principles: proportionality, right intention, legitimate authority and last resort. America’s approach to combat has long been deeply informed by this tradition, even during previous periods of technological disruption on the battlefield. The U.S. established some of the first official codes of just warfare during the Civil War, when the Gatling gun and repeating rifle changed the battlefield. Christian realism took root after atomic bombs ended World War II. In each age, elements of just war theory were adapted to establish principles for how to wield new armaments. This is no less true today. Pentagon rules on the employment of autonomous weapons date back to the Obama administration and are continuously revised. Cruder unmanned weapons have been in the American arsenal for decades. Though their capabilities might be different from those being developed today, all have always required oversight through an informed and democratically accountable chain of command. AI-powered weapons make the theory’s continuing relevance only more urgent. The worsening global security environment is bringing these questions to the fore. Pope Leo may call “dialogue,” “diplomacy” and “forgiveness” the essential tools of modern statecraft, but there is little evidence that Vladimir Putin, Xi Jinping or the Islamic Revolutionary Guard Corps agree. In the past century, as many political and church leaders argued for détente, Europe and the Soviet Union endured war, enslavement and mass murder. In hindsight, few would disagree that America was on the right side of a protracted conflict that ultimately made the world freer. Even as we pray for peace, America must always prepare for adversaries who seek war. Consider two cases of how autonomous weapons could be employed. An American air-defense battery under Iranian missile and drone salvos might use autonomous systems to task interceptors in an act of self-defense. The U.S. and Taiwan might develop robotic submarines, unmanned fighter jets, one-way attack drones and other strike weapons to deny an invasion of the island. In either case, would anyone deny the operational effectiveness — or ethical use — of automated weaponry? Deterrence has an indisputable record. Disarmament has a poorer one. If the West lays down its arms or pauses building critical technologies, it will do so alone. Russia is building drones capable of autonomous target selection. China is racing toward the “intelligentization” of its military using large language models, intending to surpass the United States. America’s rivals may already have a head start: While U.S. software companies spent the 2010s debating whether it was ethical to work with the Pentagon, Chinese military documents referred to “AI weapons” as early as 2011. If the U.S. falls behind in this race, what could it mean for tomorrow’s battlefield? Chinese gun-mounted robotic dogs could storm Taiwan’s beaches with AI-powered unmanned aircraft overhead. Fully autonomous Russian drones could crowd the skies over Ukraine. No plea for pacifism will stop this kind of war. Beijing’s generals are blunt: These weapons are designed to “win future wars.” The U.S. and its allies must be equally blunt. The task is to build new military technologies to prevent the wars Beijing, Moscow, Tehran or any other adversary may seek. In the context of today’s great-power rivalry, the obligations of a commander in chief and a technologist — to defend our families, our communities and the nation — meaningfully differ from those of church leaders or the lay faithful. But just war theory does not imply that God sanctions violence. It recognizes the authority of the state to determine the moral necessity of force. Even when force is just, wise Christians should lament its necessity. There is still much the U.S. can do to fortify its sinews of ethical doctrine and military effectiveness at once. The Trump administration is already moving to ensure that precision accuracy and machine-speed information processing are governed by the strict legal mechanisms the American way of war demands — making war more ethical and less likely. A recent national security memorandum highlights America’s “unbroken chain of command and accountability” in the deployment of AI-powered weapons. It is not clear that any analogous documents exist in the Kremlin or China’s Ministry of National Defense. If the U.S. fails to build better and faster, it will not inherit a world without war. It will inherit a world where adversaries wage war on their terms — using technology divorced from human judgment and accountability, exactly as the pope and others rightly fear. To worry or not to worry?The Financial Times reported last week that OpenAI has proposed giving the U.S. government a 5 percent ownership stake in the company. CEO Sam Altman has reportedly discussed the idea directly with President Donald Trump and a couple of top economic advisers. Other leading AI companies, reporting has it, would be expected to follow suit. (The Washington Post has a content partnership with OpenAI.) OpenAI had previously suggested that the government should set up an AI-related “Public Wealth Fund” that pays out to the American public, similar to how Alaska’s Permanent Fund pays dividends from the state’s oil wealth to residents. There’s cause for worry. As Richard Wells argues in Tech Times, this would make the government both a shareholder of AI companies and their regulator — a recipe for unintended consequences. Others legitimate worries: Government influence might stunt innovation, especially if an administration more hostile to tech companies wins power in 2028. But even as a recipe for sharing in upside, the scheme is suspect. It assumes that today’s major players in the AI industry will remain so in perpetuity. Remember Sun Microsystems? Where are Nokia and BlackBerry now? Tech moves quickly, and often in unpredictable ways. Palantir CEO Alex Karp last week sharply criticized frontier labs for syphoning off users’ data and vowed to disrupt their business model by offering clients cheaper alternatives with privacy guarantees. Which AI titan will outcompete the others? It’s early days — still too soon to place definitive bets. Essential reading More from The Post | 
