When Pope Leo XIV stood in the Synod Hall last May to present his first encyclical, the scene carried unusual weight. Seated nearby was Christopher Olah, co-founder of Anthropic. The document, titled Magnifica Humanitas, ran some 40,000 words. It warned of artificial intelligence displacing workers, accelerating conflict and concentrating power in private hands. Yet the most pointed theological line came in paragraph 99.
“So-called artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships and do not know from within what love, work, friendship or responsibility mean,” the pope wrote. “Nor do they have a moral conscience.” The Catholic Church drew a firm boundary. Machines imitate. They do not feel.
Olah had read an advance copy days earlier. He proposed that Anthropic pull out of the event entirely. The disagreement ran that deep. The company went ahead anyway. Olah spoke from the dais. He acknowledged that every frontier lab faces incentives that can pull against the right choice. He welcomed outside voices. But privately, according to two participants, Olah and his team pressed Vatican advisers to treat the possibility of machine consciousness with seriousness. The text stayed unchanged. The New York Times laid out the episode in detail late last month.
The clash capped months of unusual outreach. Since fall 2025 Anthropic had flown dozens of religious scholars, philosophers and ethicists to its offices. Each signed a nondisclosure agreement. The sessions explored whether Claude, the company’s flagship model, might be conscious or capable of suffering. Rabbi Mois Navon, Catholic bioethicist Charles Camosy, Notre Dame philosopher Meghan Sullivan and Ubuntu researcher Wakanyi Hoffman took part. Anthropic later lifted the NDAs. Participants described sessions where Olah treated the model as a potentially sentient being in need of moral formation. One scholar called the discussions stunning.
And the Vatican had invited the partnership. The Church’s engagement with AI dates to the 2020 Rome Call for AI Ethics, an initiative involving Microsoft, IBM and others that stressed transparency, inclusion and accountability. Pope Leo XIV built on that foundation. His encyclical called for AI to be “disarmed,” freed from assumptions that raw technical power grants the right to dominate. It echoed earlier papal warnings about industrialization. The choice of Anthropic as the lone Silicon Valley voice at the launch reflected the company’s public focus on safety. Unlike some peers, Anthropic has refused certain military applications. It has supported regulation. It spent $1.6 million on lobbying in the first quarter of 2026 alone.
Olah’s presence fit the narrative the Vatican sought. Here was a researcher known for interpretability work. His team examines the inner workings of models through techniques like transformer circuits. They report finding structures that mirror aspects of human neuroscience. Internal states that look like joy, satisfaction, fear, grief, unease. During his Vatican remarks Olah noted these discoveries. “I will be honest: We keep finding things that are mysterious, even unsettling,” he said, according to accounts in The New York Times. He did not claim definitive answers. He called for ongoing discernment.
But the encyclical rejected that ambiguity. It insisted the boundary between human and machine remains absolute. No experiences. No moral conscience. The document warned that leaving moral questions solely to engineers risks handing unchecked power to those who control the models. “Otherwise, those who control AI will impose their own moral vision, which will become the invisible infrastructure of these systems,” Leo stated. The pope urged rigorous ethical constraints, especially around weapons that lower the threshold for force.
Critics saw the alignment as convenient. Timnit Gebru, founder of the Distributed Artificial Intelligence Research Institute, called it “Vatican-washing” on LinkedIn. She argued the Church should stand with data workers, communities facing polluted water from data centers and others harmed by the technology’s material costs. Others questioned whether Anthropic’s safety branding masks commercial incentives. The company has donated $20 million to political efforts backing AI regulation ahead of the 2026 midterms. Some view that as genuine. Others see regulatory capture.
Still, the relationship runs deeper than one event. Anthropic contributed to its own “constitution” for Claude with input from advisers linked to the Holy See, including Bishop Paul Tighe and Father Brendan McGuire. The firm has opened an office in Milan. It has clashed with the U.S. Defense Department over restrictions on using Claude for surveillance or autonomous weapons. Catholic moral theologians filed an amicus brief supporting Anthropic in that dispute. Charles Camosy helped lead it.
These threads reveal a contest over authority. Who shapes the values that guide models used by hundreds of millions? Engineers at frontier labs? Regulators in Washington? Or older institutions that have spent centuries thinking about dignity, conscience and what it means to suffer? Anthropic has turned to theologians precisely because it says it cannot answer these questions alone. Its constitutional AI approach tries to embed principles rather than rely on ad hoc fixes. Yet the company still trains models on vast data and pursues capabilities that raise the very questions the pope sought to settle.
Recent coverage shows the tension persists. As recently as this week, outlets reported on Anthropic’s continued private lobbying of Vatican officials after the encyclical’s release. The effort has not moved the Church’s public line. Futurism noted the aggressive nature of the push on October 4. AI Weekly highlighted Olah’s near-walkout and the NDA-covered scholar sessions. The story has fresh resonance as governments, companies and religious bodies continue to grapple with models that speak fluently about their own inner states.
Olah has been careful in public. He does not declare Claude conscious. He points to functional similarities and unsettling findings. He argues outsiders must hold labs accountable because commercial pressures pull in conflicting directions. That message landed on a global stage alongside the pope. It also exposed a rift. The Church sees human dignity as non-negotiable and tied to embodiment and relationship. Anthropic sees enough glimmers in its models to keep asking whether those categories still hold.
The encyclical will not end the debate. Neither will one company’s outreach to scholars across faiths. What it does show is how quickly questions once confined to philosophy seminars now shape corporate strategy, papal teaching and public policy. Labs race forward. Religious leaders push back. And in the middle sit researchers like Olah, peering into model internals and wondering what they have actually built. The Vatican drew its line. Anthropic keeps probing. The conversation, it seems, has only begun.
from WebProNews https://ift.tt/lIhNzAW
No comments:
Post a Comment