Why Pope Leo is right to call on EU to disarm lethal AI weapons

# Why Pope Leo is right to call on EU to disarm lethal AI weapons Source: [https://euobserver.com/219185/pope-leo-xiv-urges-europe-to-stop-lethal-ai-weapons-before-its-too-late](https://euobserver.com/219185/pope-leo-xiv-urges-europe-to-stop-lethal-ai-weapons-before-its-too-late) This week Pope Leo XIV released his first major encyclical,[*Magnifica Humanitas*](https://www.vaticannews.va/en/pope/news/2026-05/pope-leo-xiv-encyclical-magnifica-humanitas-ai.html)\. While a new pope’s first encyclical is typically devoted to broad questions of faith and human dignity, Leo has instead chosen to confront one of the most pressing issues of our time by placing lethal autonomous weapons and the militarisation of artificial intelligence at the centre of the document\. In a dedicated section on weapons and artificial intelligence, the pope calls for AI to be ‘disarmed,’ states bluntly that “no algorithm can make war morally acceptable,” and insists that lethal or otherwise irreversible decisions must never be entrusted to machines\. He explicitly condemns lethal autonomous weapons as ‘not permissible’ and urges rigorous ethical constraints on AI in warfare\. The encyclical’s language on human control is unusually direct\. It warns that autonomous systems risk lowering the threshold for conflict, making war more impersonal and easier to initiate\. For EU policymakers, the timing could not be more relevant: while the Vatican frames the issue in moral terms, the practical implications for European security and regulation are immediate\. The reason is that the development of autonomous weapon systems, like AI more broadly, is advancing rapidly, as falling costs, commercial AI advances, and export dynamics are accelerating their development and deployment\. Regulation, by contrast, lags dangerously behind\. The[EU AI Act](https://artificialintelligenceact.eu/)largely excludes systems developed or used exclusively for military purposes, leaving lethal autonomous weapons outside the bloc’s main risk\-based framework\. The[European Defence Fund](https://defence-industry-space.ec.europa.eu/eu-defence-industry/european-defence-fund-edf-official-webpage-european-commission_en)conditions funding on compliance with ethical and legal standards, including human oversight, but sets no binding rules for national procurement or operational use\. In the Russia\-Ukraine war, from AI\-assisted targeting to semi\-autonomous loitering munitions and emerging drone swarms, systems with varying levels of autonomy are proliferating rapidly\. The battlefield use of autonomous systems in Ukraine is eroding long\-standing norms against delegating lethal decisions to machines\. Once these capabilities become cheaper and more accessible, the risk of proliferation to authoritarian regimes and non\-state actors grows sharply\. Europe cannot treat this as a distant problem; the next conflict on its borders could feature the very systems the EU claims to want to constrain\. ## ‘Human control’ vs the US\-controlled cloud Meaningful human control is central to regulating lethal autonomous weapons\. It requires that operators retain sufficient understanding, supervision, and the ability to intervene in or override lethal decisions, ensuring that moral and legal responsibility remains with people rather than machines\. Europe’s position on human control reveals clear tension\. The EU and its member states have consistently championed strong normative standards on it in international forums\. Yet industrial realities tell a different story\. Major European programmes such as the[Future Combat Air System](https://www.airbus.com/en/products-services/defence/future-combat-air-system-fcas)\(FCAS\) and[Global Combat Air Programme](https://www.leonardo.com/en/business/gcap)\(GCAP\) are likely to increasingly rely on components and cloud infrastructure from US AI firms\. Enforcing European standards of meaningful human control becomes difficult when core technological building blocks come from partners with different priorities and looser rules\. This dependence raises concerns about European strategic autonomy\. If critical AI components in FCAS and GCAP are designed under different rules on human control, the EU risks importing not only technology but also the very standards it publicly criticises\. It creates a quiet erosion of regulatory sovereignty in one of the most consequential domains of future warfare, leaving the EU with a hybrid approach that appears ambitious on paper but proves difficult to enforce once projects move beyond the funding stage\. The resulting enforcement gap matters\. If the EU continues to treat military AI regulation as an afterthought, it risks losing both moral authority and practical influence over how these systems are developed and used\. European defence industries already face pressure to keep pace with faster\-moving competitors\. Without coherent rules spanning research, procurement, and export, the bloc may end up importing standards it claims to reject\. The EU must now close the gap between its normative ambitions and its industrial dependencies\. This requires moving beyond declarations\. Brussels should introduce binding meaningful human control standards for all joint defence procurement and exports, not just EDF\-funded projects, invest in sovereign European AI capabilities for military use, and drive forward a stronger EU push within the[Convention on Certain Conventional Weapons](https://disarmament.unoda.org/en/our-work/conventional-arms/convention-certain-conventional-weapons), advancing stronger verification mechanisms and a legally binding instrument\. If Europe wants to retain both ethical credibility and operational sovereignty in the coming era of military AI, reducing technological dependence on non\-European suppliers must become a core strategic priority\. The window is narrowing\. Autonomous weapons are moving from prototype to battlefield at a pace few predicted even a year ago\. Europe cannot allow regulation and industrial strategy drift apart any longer\. The cost of inaction will be measured not only in ethical terms, but also in diminished strategic autonomy\.