Anthropic, the company behind the Claude chatbot, privately pressed Pope Leo XIV's advisers to take seriously the possibility that its AI models could be conscious — and the pontiff, who had already rejected that idea in a major encyclical, refused to budge.
That is the picture that emerges from a New York Times report by national religion correspondent Elizabeth Dias, published Tuesday and amplified by follow-up coverage in The Telegraph, Vox, and other outlets this week. What might have stayed a niche theological debate has become one of the AI industry's most unusual public clashes: Silicon Valley's most safety-focused laboratory facing off against the head of the Catholic Church over whether a language model can have an inner life.
The dispute matters beyond theology. If a leading AI lab treats its models as potential bearers of moral status, that could reshape how models are trained, retired, and marketed to hundreds of millions of users. You can follow our AI industry coverage for the latest developments as this debate moves from ethics boards into public view.
A Year of Quiet Meetings With Religious Scholars
According to the Times, Anthropic has spent roughly the past year conducting a series of private meetings and calls with religious scholars from many faith traditions, many of them under nondisclosure agreements, to explore whether its Claude models "might be conscious or have moral status."
Participants told the paper that members of the Anthropic team appeared to be "relating to [Claude] like a conscious being," and that the company entertained the idea that its models could deserve a degree of moral status. The courtship came with hospitality. On one April evening, an Anthropic executive took a group of religious thinkers to an upscale tasting-menu restaurant in San Francisco; participants reportedly went home with a handwritten thank-you note, a coffee mug, and an orange-bound copy of Claude's constitution.
Not everyone was charmed. David Decosimo, a theologian who teaches ethics and religion at the University of North Carolina, wrote on X this week that Anthropic had courted religious leaders under nondisclosure agreements "all while wining & dining them to the max."
The Encyclical That Alarmed Anthropic
At the center of the clash is "Magnifica humanitas: On safeguarding the human person in the time of artificial intelligence," Pope Leo XIV's encyclical on AI, released on May 25, 2026, at the Vatican. The document argues that AI will never be able to replace humanity — a stance that, according to the Times, alarmed Anthropic co-founder Chris Olah when he reviewed an advance copy days before the launch.
Olah, who leads Anthropic's interpretability research, had been invited to speak at the encyclical's presentation in the Synod Hall. The Times reports that he went so far as to propose withdrawing from the event over its stance on machine consciousness. Two participants told the paper that Olah and his team then privately lobbied the pope's advisers to take the possibility of machine consciousness seriously. The Telegraph's account was blunter still: "Anthropic tried to persuade Pope that AI could be conscious being."
Anthropic's own published remarks from the Vatican event show how Olah frames the question. AI systems, he told the audience, are not engineered the way a bridge or an airplane is. "They are grown, on a structure roughly modeled after the brain, on an enormous inheritance of human thought and speech," he said, adding that the resulting systems remain "mysterious even to those of us who train them." He compared building them to "bringing a fictional character to life," and argued that the questions they raise belong to the humanities, to philosophy, and to religion — not only to computer science.
Who Is Chris Olah?
Olah is one of the most respected figures in AI interpretability, the field that tries to map the internal circuits of neural networks. According to the Times, he grew up evangelical and has since left Christianity, meaning he arrived in Rome without deep fluency in how the Church teaches — a detail that may explain why the meeting of minds was so awkward, and why the pope's advisers stood firm.
The Pope's Answer: Algorithms Lack 'the Spark of Humanity'
Pope Leo held his ground. The encyclical's core claim — that machines can simulate human output without possessing human dignity — has been echoed in his public comments throughout the week. On Friday, he posted that "there is an ontological difference, even before an aesthetic one, between art and what a machine can generate through statistical calculation based on millions of images created by others," adding that "algorithms lack the spark of humanity."
The pope has repeatedly framed AI as a tool that must remain subordinate to human beings, warning lawmakers and international bodies that technology must serve the common good rather than become a rival to it. The Vatican's position, in short: no amount of statistical sophistication confers a soul.
An Industry Split Over Machine Minds
The Anthropic-Vatican standoff sits inside a broader argument roiling the industry. Microsoft's AI chief Mustafa Suleyman has publicly cautioned against treating today's models as conscious, warning that "AI consciousness" narratives are premature and potentially harmful. OpenAI chief executive Sam Altman told Axios this week that ascribing religion to models is a "safety issue" of its own.
At the same time, the major labs are shipping products that encourage users to relate to models as people. OpenAI and Meta introduced companion-style AI agents called Dots and Muse last month, and labs increasingly publish "model welfare" policies that treat the question of machine experience as at least open. Critics argue the industry cannot have it both ways: marketing companionship while denying that anything is home inside the machine.
Why It Matters
Whether AI systems are conscious is not an abstract question. Moral status would bear on how models are trained, whether they can be switched off, how they are advertised to lonely or vulnerable users, and what rights — if any — future systems might claim. It could also shape regulation: legislators drafting AI rules currently assume the systems are property, not persons.
Anthropic has not disavowed the outreach described by the Times, and the company's own remarks from the Vatican presentation argue for widening the conversation about AI to traditions far outside Silicon Valley. The Vatican, for its part, has made clear that the Church's answer to machine consciousness is, for now, a firm no.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
