The Pope’s 42,300-word encyclical, Magnifica Humanitas, calls for protections against AI job losses and for children from AI-generated violent, hypersexualized, or fake information. It calls for algorithmic transparency, democratic oversight, and meaningful public participation in AI governance. Humans’ dignity must be preserved from a technology that threatens to replace human skills and tasks. Private companies must be constrained by regulation.
Many in Silicon Valley, far from rejecting these recommendations, endorse them. When the present Pope launched his encyclical, Anthropic co-founder Chris Olah stood next to him. Olah has warned that AI may displace human labor at a large scale, and Anthropic’s own researchers have found “internal states that functionally mirror joy, satisfaction, fear, grief, and unease” inside their models. Even the Trump Administration has been backing away from its hands-off policies, insisting that the government vet AI models for safety.
The Pope’s strong message draws on 135 years of Catholic Social Doctrine. Back in 1891, another Pope Leo issued the Rerum Novarum encyclical in response to the upheavals of the Industrial Revolution. Leo XIII looked at factory workers laboring in degrading conditions while industrialists accumulated wealth. In response, the Pope stressed the dignity of labor, the right to fair wages, and the right of workers to organize in unions.
The Industrial Revolution posed similar challenges to the AI Revolution. It concentrated productive power in machines and in the owners of machines, displacing artisans and agricultural workers at scale. Frontier AI concentrates value in algorithms and in the companies that own them, forcing workers, governments, and civil society to adapt.
The difference is that the Industrial Revolution encyclical landed in a world where the nation-state was the primary unit of economic governance. The mills were in Manchester and Pittsburgh. Their owners could be taxed, regulated, or confronted by organized labor.
In contrast, AI companies reshaping the global economy today are multinational. Their bytes and bits cross borders with the flick of a switch. The Internet’s infrastructure is distributed across jurisdictions. Regulations that reigned in industrialization — such as labor law, antitrust, and the welfare state — were national solutions. Today, national or even regional regulations risk impotence.
The Church and the frontier AI lab Anthropic share the same alarm and the same conviction that answers cannot come from the labs alone. Where they differ is in specificity: the encyclical names mechanisms, Anthropic names the urgency. So far, voluntary ethics frameworks that AI companies have signed onto have tended to bend with political winds.
In the present Pope’s view, genuine external accountability must come in the form of labor and workers’ representatives with real bargaining power, democratically mandated public hearings, antitrust tools capable of addressing the concentration of AI development, and shifting power to international treaty bodies with enforcement authority.
Another notable historical resonance concerns war. When confronted with the anarchic destructiveness of medieval warfare, St. Augustine and Thomas Aquinas came up with the theory of Just War that establishes the conditions under which a nation can morally justify military actions, as well as the ethical boundaries for how that war should be conducted.
But Just War theory still required princes, eventually parliaments, and ultimately international law to become anything more than moral doctrine. Adversaries and competitors had to buy into it and agree. Today, governments and companies must come together to respond to AI.
It will be difficult. While the encyclical and Olah’s speech share a broadly Western, liberal, and humanist tradition, in which individual dignity, democratic participation, and civil society oversight are the anchors, AI is not being built only within that tradition. China’s approach to AI development rests on different premises: collective harmony over individual rights, state authority over independent oversight, and technological capacity as an expression of national sovereignty.
Fragmented global governance risks creating a two-track scenario with a Western AI and a Chinese AI, diverging in values and safety standards. This is already happening. Our focus must be on preventing it from hardening into permanent incompatibility. So far, no institution has seriously attempted to champion that effort.
The irony is that the country most capable of building the required architecture is the one currently least willing to do so. The US has retreated from the multilateral engagement that AI governance requires. The EU has produced serious regulatory work. But Brussels lacks the geopolitical weight to compel Washington or Beijing. No combination of willing middle powers adds up to American leadership.
The Pope has stepped into this vacuum. He has a strong voice and enjoys widespread support. The Trump Administration has stepped up its own safety checks on Anthropic and OpenAI’s latest models. The unanswered question is whether Washington will turn these checks into a new international alliance devoted to both advancing — and taming — AI.
Bandwidth is CEPA’s online journal dedicated to advancing transatlantic cooperation on tech policy. All opinions expressed on Bandwidth are those of the author alone and may not represent those of the institutions they represent or the Center for European Policy Analysis. CEPA maintains a strict intellectual independence policy across all its projects and publications.
Tech 2030
A Roadmap for Europe-US Tech Cooperation
