Key Points:
- Anthropic co-founder Chris Olah spoke alongside Pope Leo XIV at the Vatican, warning that the future of AI cannot be left solely to tech companies.
- Olah conceded that every frontier AI lab operates under commercial and geopolitical incentives that can conflict with doing the right thing.
- The tech pioneer warned of a real possibility that AI will displace human labor on a massive scale, calling post-displacement support a historic moral imperative.
- The remarks coincided with the launch of “Magnifica humanitas,” the first-ever papal encyclical dedicated entirely to artificial intelligence.
In a historic and highly unusual convergence of religion and technology, Anthropic co-founder Chris Olah joined Pope Leo XIV at the Vatican to demand independent, external oversight of artificial intelligence. Speaking at the formal presentation of the Pope’s first encyclical letter on Monday, May 25, 2026, the tech pioneer delivered a remarkably candid assessment of the AI industry. Olah argued that the development of such a powerful technology cannot be left entirely in the hands of private technology companies, urging religious leaders, governments, and civil society to help guide its trajectory.
The landmark event took place inside the Vatican’s Synod Hall, where Pope Leo XIV unveiled his first encyclical, titled Magnifica humanitas (“Magnificent Humanity”). This more than 42,000-word document represents the first time in the Catholic Church’s 2,000-year history that a pope has dedicated an entire foundational teaching letter to an emerging technology. Signed on May 15—the 135th anniversary of Pope Leo XIII’s 1891 Rerum novarum on the Industrial Revolution—the document aims to establish a clear moral framework to protect human dignity in the digital age.
Sitting alongside the pontiff, Olah offered a rare inside perspective on the intense cultural and financial forces driving Silicon Valley. He admitted that every frontier AI laboratory, including safety-oriented firms like Anthropic, operates within a complex web of commercial, geopolitical, and personal incentives. These competitive pressures, such as pride and ambition, can easily conflict with doing the right thing for society. Olah argued that even the most well-intentioned computer scientists cannot fully escape these corporate forces, making unbiased self-regulation nearly impossible.
This candid admission comes as tech conglomerates invest more than $100 billion annually into building increasingly complex generative AI systems. To safeguard the public, Olah insisted that the AI research community needs earnest, thoughtful external critics who can challenge corporate decisions. He welcomed the Catholic Church’s active engagement, noting that the ethical questions raised by these deep-learning models extend far beyond the boundaries of computer science and computer engineering.
During his address, Olah highlighted three specific areas of AI development that require urgent international attention. First, he warned of a real possibility that AI will displace human labor on a massive scale. If this displacement occurs, Olah emphasized that supporting those who lose their livelihoods will become a moral imperative of historic proportions. His statement marks one of the most direct public acknowledgments by an AI safety pioneer that the technology could disrupt global employment markets faster than society can reabsorb displaced workers.
Second, Olah raised serious concerns about the global distribution of technology’s benefits. Currently, a handful of wealthy nations and elite tech firms monopolize the development of cutting-edge AI. This high concentration of wealth risks widening the digital divide, leaving developing nations on the margins of the technological revolution. To prevent this, the Anthropic co-founder urged international bodies to find concrete ways to share the economic and social gains of artificial intelligence globally, rather than allowing a small group of players to control the wealth.
Finally, Olah pointed to the unresolved and highly complex challenge of “interpretability” in modern neural networks. As the head of interpretability research at Anthropic, Olah spends his days trying to understand the inner workings of large language models like Claude. He explained that as these systems become more powerful, their internal reasoning processes become increasingly opaque and difficult to parse. Understanding how these models make decisions is vital to preventing unpredictable or dangerous system behaviors.
Pope Leo XIV echoed many of these concerns in his encyclical, taking direct aim at the monopolistic control of major tech platforms. The Pope warned that AI tends to amplify the power of those who already possess massive economic resources, data access, and technical expertise. This concentration allows small, highly influential groups to shape global information, influence democratic processes, and steer economic dynamics to their own advantage. To counter this, the pontiff called for the “disarmament” of artificial intelligence, which he defined as restoring the moral primacy of the human over the algorithm.
The joint presentation at the Vatican highlights a profound shift in the global debate surrounding AI safety and ethics. As Silicon Valley rushes to build more capable models, the industry’s own pioneers are looking to external moral authorities to help fill the governance vacuum. By combining the moral reach of the Catholic Church—which influences over 1.4 billion people worldwide—with the technical insights of leading research labs, advocates hope to steer the development of artificial intelligence toward the common good, ensuring that technology ultimately serves to elevate, rather than dominate, humanity.
