Good intentions may not be enough when financial pressures mount.
At the Vatican ceremony unveiling Pope Leo XIV's sweeping moral indictment of artificial intelligence, Anthropic co-founder Chris Olah stood as a guest speaker — a moment that crystallized one of the defining tensions of our technological age: the proximity of those who build transformative systems to those who warn of their consequences. The encyclical's concerns about worker displacement, environmental harm, and automated warfare are not abstract to Anthropic, whose own research identifies entire professions at risk and whose infrastructure investments run counter to the pope's calls for restraint. Whether this alliance represents a genuine reckoning or a carefully managed performance of conscience is a question the partnership has raised but not yet answered.
- A 42,000-word papal encyclical condemning AI's threat to human dignity landed with unusual force when one of AI's most prominent architects was standing on the same stage to receive it.
- Critics immediately named the contradiction: a company whose business model depends on automating human labor cannot simply absorb a moral framework built around protecting it without something having to give.
- The charge of 'Vatican-washing' cuts deep — suggesting the church may have lent its moral authority to the very actors whose accountability it should be demanding, rather than to the workers and communities already bearing AI's costs.
- Anthropic's genuine refusal to enable autonomous weapons offers a rare example where its stated values have actually cost it — triggering government blacklisting and legal battles — lending at least partial credibility to its presence at the ceremony.
- With $50 billion in infrastructure commitments expanding the data centers the pope explicitly warned against, the company is navigating a narrowing corridor between moral positioning and the imperatives of scale.
- The partnership has created pressure, but pressure is not yet accountability — and observers are watching closely to see whether this dialogue will constrain what Anthropic builds, or simply adorn it.
Pope Leo XIV's first major teaching on artificial intelligence arrived as a moral indictment: AI threatens to displace workers, accelerate warfare, and strain the natural world. The document ran to roughly 42,000 words. At the ceremony marking its release, the Vatican welcomed as a guest speaker Chris Olah, co-founder of Anthropic — one of the most powerful AI companies in the world, and a principal architect of the technology the encyclical warned against.
The tension was immediate. Anthropic's own labor market research had identified coders, customer service workers, and data-entry professionals as professions at acute risk. One in five American full-time workers already report that AI has absorbed parts of their jobs. Anthropic's CEO has himself warned of sweeping white-collar displacement in the years ahead. Yet the company was being welcomed into the Vatican's moral conversation as a partner.
Critics named what they saw. A Notre Dame law professor warned of 'feelgood' discourse — alliances that allow both parties to claim alignment without requiring either to change. Timnit Gebru, founder of the Distributed AI Research Institute, called it 'Vatican-washing,' arguing the church had chosen the wrong partners. The Vatican's moral weight, she suggested, belonged with exploited data workers and communities whose water supplies were threatened by data centers — not with the companies driving those harms.
There is a coherent logic to Anthropic's positioning. The company has built its identity around safety and responsibility, lobbies actively for AI regulation, and refused to allow its models to be used for fully autonomous weapons or mass surveillance — a stance that triggered a feud with the Trump administration and an ongoing legal battle. Google and OpenAI were not on that Vatican stage. The implication was deliberate.
Yet the contradictions hold. Leo's encyclical calls for environmentally responsible AI development, citing the enormous energy and water demands of current systems. Anthropic has committed $50 billion to AI infrastructure, including the data centers the pope is asking to be made more sustainable. Olah himself acknowledged the bind in his remarks, admitting that even sincere intentions are shaped by financial incentives — an unusual moment of candor from inside the industry.
The clearest point of genuine alignment between the two parties is warfare. Leo warned that AI lowers the threshold for military action and dissolves human accountability in conflict. Anthropic's refusal to enable autonomous weapons reflects the same concern — and has actually cost the company, in contracts and legal standing. It is the one place where stated values have met real consequence.
Whether the broader partnership deepens into accountability or remains a sophisticated form of brand management is the question the alliance has opened but not resolved. Some observers see the very tension between the pope's words and Anthropic's actions as productive — a source of pressure. Others are less hopeful. The Vatican has opened a door. Whether Anthropic walks through it, or simply remains visible in the doorway, is what the coming years will reveal.
Pope Leo XIV released his first major teaching on artificial intelligence with a stark warning: the technology threatens to displace workers, accelerate warfare, and ravage the environment. The encyclical, roughly 42,000 words, laid out a moral case against AI's unchecked expansion. Then, at the Vatican ceremony marking its release, something unexpected happened. Standing beside the pontiff as a guest speaker was Chris Olah, co-founder of Anthropic, one of the world's most powerful AI companies—and one of the architects of the very technology Leo had just condemned.
The optics created an immediate tension. Here was a company whose business depends on building systems designed to automate human work, standing shoulder-to-shoulder with a religious leader whose encyclical emphasizes the dignity of human labor. Anthropic's own labor market analysis, released in March, identified entire professions at risk: coders, customer service representatives, data-entry workers. A survey by the nonprofit Epoch AI found that one in five American full-time workers already report that AI has taken over parts of their job. Dario Amodei, Anthropic's CEO, has himself warned of catastrophic job losses among white-collar workers in the coming years. Yet there he was, or his company was, being welcomed into the Vatican's moral conversation.
Critics were quick to name what they saw happening. Paolo Carozza, a law professor at Notre Dame and co-chair of the Meta Oversight Board, called it a risk of "feelgood" discourse—the kind of alliance that allows both parties to claim alignment without forcing either to change. Timnit Gebru, founder of the Distributed Artificial Intelligence Research Institute, was blunter. She called it "Vatican-washing," suggesting the church had partnered with the wrong actors. The Vatican should have aligned itself, she argued, with exploited data workers fighting for their rights, with communities whose water supplies were threatened by data centers, with the actual victims of AI's expansion.
There is a logic to Anthropic's positioning here. The company has built its brand around safety and responsibility—a deliberate contrast to competitors like OpenAI. It spends heavily on lobbying for AI regulation. It refused to allow the U.S. government to use its models for fully autonomous weapons and mass surveillance, a stance that led to a bitter feud with the Trump administration and an ongoing court battle. By appearing at the Vatican, by engaging with the pope's moral framework, Anthropic signals that it takes these concerns seriously. Google and OpenAI were not on that stage. The implication is clear: we are different.
Yet the contradictions remain. Leo's encyclical calls for sustainable, environmentally responsible AI development. It notes that current systems consume enormous amounts of energy and water, driving up carbon emissions and straining natural resources. Communities across the United States have already begun pushing back against data centers, worried about industrial emissions and skyrocketing energy bills. Anthropic, meanwhile, has committed to investing $50 billion in AI infrastructure, including data centers—the very facilities the pope is asking to be made more sustainable. The company has pledged to cover electricity price increases for consumers affected by these facilities and to reduce power usage during peak demand, but the fundamental tension remains: Anthropic's growth depends on the expansion of the infrastructure Leo is warning against.
Olah himself acknowledged the bind during his remarks at the ceremony. Every AI lab, he said, operates within incentives and constraints that can sometimes conflict with doing the right thing. "No matter how sincerely any of us intend to do the right thing—and I believe many of us do—we will always be influenced by those incentives," he said. It was a moment of unusual candor, an admission that good intentions may not be enough when financial pressures mount. Anthropic spent a record $1.6 million on lobbying in the first quarter of 2026 alone, outpacing OpenAI. The company is navigating a path between moral claims and business imperatives.
There is one area where the Vatican and Anthropic do genuinely align: the dangers of AI in warfare. Leo wrote about how artificial intelligence can lower the threshold for military action, shield people from responsibility, and reduce enemies to statistics and victims to "collateral damage." He called for the most rigorous ethical constraints to protect human life. Anthropic's refusal to enable fully autonomous weapons reflects this same concern. It is perhaps the clearest point of agreement between the two parties, and the one where Anthropic's stated values have actually cost the company—triggering government blacklisting and legal battles.
But alignment on one issue does not resolve the larger question. Is this partnership a genuine commitment to accountability, or a sophisticated form of brand management? Some observers, like Pete Furlong of the Center for Humane Technology, are cautiously optimistic. The conflict between the pope's words and Anthropic's actions, he argues, is actually a good sign—it creates pressure for change. Others remain skeptical. What matters now is whether the dialogue continues, whether it deepens, and whether it actually constrains what Anthropic builds and how it builds it. The Vatican has opened a door. The question is whether Anthropic will walk through it, or simply stand in the doorway.
Notable Quotes
All of these companies are building technology that is designed to replace people. That's very much at odds with the pope's words.— Pete Furlong, Center for Humane Technology
No matter how sincerely any of us intend to do the right thing—and I believe many of us do—we will always be influenced by those incentives.— Chris Olah, Anthropic co-founder
The Hearth Conversation Another angle on the story
Why would the pope agree to share a stage with someone from the company he's warning against?
Because the alternative is silence. If the Vatican only talks to people who already agree with it, the message never reaches the people making the decisions. Olah's presence is uncomfortable—that's partly the point.
But doesn't that discomfort just disappear once the ceremony ends?
It might. That's what the critics are worried about. The Vatican gets to say it engaged with AI leaders. Anthropic gets to say the pope trusts them. Both sides claim victory and nothing changes.
Olah admitted that incentives can override good intentions. Isn't that a confession?
It's an acknowledgment, which is different. He's saying the problem is structural, not personal. That's honest. But honesty without action is just performance.
What would real change look like?
The pope would have to say no to something Anthropic wants. Anthropic would have to build less, or slower, or differently. Right now they're both talking about the same problems while moving in opposite directions.
On weapons, they actually agree.
They do. And Anthropic paid a real price for that stance—blacklisting, legal battles. That's where you see whether the values are genuine. When it costs something.
So the question is whether the Vatican can make it cost something too?
Exactly. Whether the moral weight of the church can actually bend the trajectory of a $50 billion infrastructure investment. That's what we're watching.