On Monday May 25, 2026 Pope Leo XIV released a 235-page encyclical, titled "Magnifica humanitas: On safeguarding the human person in the time of artificial Intelligence," and presented it at the Vatican alongside AI experts including Chris Olah, who spoke at the document’s public presentation.
The pope used the letter to warn against the military and political dangers of rapid AI deployment, saying that he had deliberately chosen strong language to attract attention and writing that "No algorithm can make war morally acceptable." The reported that Leo described artificial intelligence as needing to be "disarmed," and that he condemned the use of AI in warfare and the manipulation of images and videos in politics.
Olah, a co-founder of Anthropic who was invited to the Vatican presentation, told the audience that the risks around AI extend beyond the labs that build it, saying, "Every frontier AI lab—including Anthropic—operates inside a set of incentives and constraints that can sometimes conflict with doing the right thing." He pressed the point that there must be people outside those incentives who insist on safety, and that the questions AI raises are "bigger than the AI research community, not just in their implications, but also in their nature."
Olah framed the technology as not engineered like a bridge or an airplane but grown from a structure roughly modeled after the brain and fed on a vast inheritance of human thought and speech; he said the systems remain mysterious even to their builders and compared bringing an AI model to life to bringing a fictional character to life. He warned that mass job losses from artificial intelligence are "a real possibility" and called supporting displaced workers "a moral imperative of historic proportions," warnings reported by Forbes.
Anthropic said Olah’s appearance was part of its initiative to widen the conversation on the important questions raised by AI. The presence of a leading industry voice at the Vatican underscored the encyclical’s intent: this is not technical guidance for engineers but a moral teaching document aimed at society at large. The noted that the letter is Pope Leo’s first major teaching document and his first encyclical, and that it addresses slavery, labor, warfare, education, privacy and human dignity.
The encyclical and Olah’s remarks together moved the debate from code and corporate memos into moral and public terms. The pope’s blunt line—"No algorithm can make war morally acceptable"—is both a moral verdict and a challenge to governments and militaries that may be weighing offensive or autonomous uses of AI. Olah’s insistence that decisions "should not be left to people in the industry" framed the issue as a civic one: technical communities alone cannot settle the shape of systems that touch work, politics and human identity.
The tension is clear. The Vatican has moral authority and a global pulpit; it urged disarmament and limits on manipulation. But the practical levers for preventing an AI arms race or for protecting workers—budgets, regulation, international agreements—sit with secular governments and private firms. Forbes has also reported that AI development is concentrated in "a handful of wealthy nations," a fact that complicates any global response the encyclical seeks to catalyze.
The pope’s document lays out a broad moral frame; Olah’s testimony supplied a technical and ethical warning from someone inside the field. Both stressed that AI’s nature makes it different from previous technologies and that the choices about what character these systems will have are questions for the humanities, religion, philosophy and society at large. Whether that moral framing translates into policy is now the central question.
The single most consequential unanswered question is whether governments and industry will act on the encyclical’s call—to disarm AI in warfare, curb political manipulation and meet the "moral imperative of historic proportions" to support workers—turning moral injunction into enforceable rules and resources rather than leaving the future of artificial intelligence to labs and markets alone.
