Pope Leo XIV has put AI-directed warfare on the moral and political agenda at a moment when defense technology is moving from software into life-and-death decisions.
Pope Leo XIV did not speak to founders, investors or defense contractors when he addressed La Sapienza University in Rome on May 14. But the audience that should read the speech most carefully may be the people building autonomous systems, battlefield analytics and agentic AI tools that governments are now eager to buy.
The message was simple enough. Artificial intelligence cannot become a way for humans to step back from responsibility. As AP reported from Rome, Leo warned that rising investment in AI and high-tech weaponry is pushing conflicts in Ukraine, Gaza and the Palestinian territories, Lebanon and Iran into what he called a spiral of annihilation. That is not a narrow theological complaint. It is a direct challenge to the idea that faster targeting, faster escalation and more automated decision-making automatically make the world safer.
For the AI industry, the timing matters. This was not an isolated line in a campus speech. Axios reported that Leo is expected to sign his first encyclical as soon as May 15, with artificial intelligence framed as a moral and labor challenge for a new industrial era. If that document arrives as expected, AI will no longer be just a regulatory issue for Brussels, Washington or Beijing. It will be treated by one of the world's largest global institutions as a test of human dignity, work and peace.
The Catholic Church cannot write procurement rules for the Pentagon or stop a startup from selling software to a defense ministry. That is not the point. Institutions like the Vatican operate through pressure, language and legitimacy. They shape what becomes acceptable, what becomes embarrassing and what investors are forced to explain in public.
That is why Leo's speech should matter even to people who have no religious interest in it. AI governance has often been treated as a technical problem: model evaluations, safety thresholds, audit trails, red teaming and compliance. Those are all necessary. But the pope is pushing the argument into a different lane, where the question is not only whether a system works, but whether it lets people hide behind the machine when the consequences are brutal.
This is especially uncomfortable for defense AI companies because many of their products are sold in the language of precision and restraint. Better intelligence can reduce civilian harm. Faster analysis can help commanders avoid bad decisions. Autonomous navigation can keep soldiers out of danger. Those arguments can be true. They are also incomplete if the technology shortens the path from suspicion to strike, or if human review becomes a ceremonial checkbox rather than a real decision.
Leo's intervention lands at a time when war has already become a proving ground for drones, sensors, automated targeting support and information warfare. Ukraine has shown how quickly civilian technology can become military infrastructure. Gaza has intensified scrutiny of how advanced systems are used in dense urban conflict. Iran and Lebanon sit inside a wider regional environment where miscalculation can spread quickly. In that context, moral language becomes strategic language.
Defense founders need to read the room
For founders building in defense technology, the practical takeaway is not to avoid the sector. Democracies will continue buying AI tools for security, logistics, intelligence and battlefield protection. The question is whether companies can show that human accountability is built into the product and the business model, not added later for a policy page.
That means clearer lines around what systems are allowed to recommend, what they are allowed to execute and where a human must remain meaningfully in control. It also means keeping records that can survive public scrutiny. If a company says its tool supports decision-making, it should be able to explain how the tool handles uncertainty, what data it depends on and how operators are trained to challenge its output.
Investors should pay attention as well. The defense AI boom has been driven by the belief that Western governments will modernize procurement and spend heavily on software-defined military capability. That thesis may still be right. But capital will increasingly face a reputational filter. A company working on battlefield autonomy may need more than technical excellence and government contracts. It may need a credible ethical architecture that customers, lawmakers and limited partners can defend.
The same applies beyond weapons. Leo's reported encyclical focus on labor suggests a broader critique of AI systems that replace human judgment with automated control. Civilian AI agents that approve loans, screen workers, manage classrooms or influence public debate may not carry missiles, but they raise a similar concern: who is responsible when the machine acts and people are harmed?
That is where the Vatican's role becomes interesting. It is unlikely to slow the AI race by itself. But it can help make human responsibility the center of the debate at a moment when markets are rewarding speed, scale and automation. For startups, that changes the conversation with customers. It is no longer enough to say the technology is powerful. The stronger question is whether the company can prove that power remains accountable.
The next thing to watch is whether Leo's expected encyclical turns this warning into a durable framework. If it does, AI companies may find that governance is not just a legal burden or a brand exercise. It may become part of how buyers, investors and the public decide which technologies deserve trust.
Also read: Stealth clipping campaigns are making organic virality harder to trust • Orthrus makes local AI inference economics look worth rechecking • AI data centers are pushing power prices higher on the PJM grid
