Anthropic's Doing Something Unusual
While most tech companies keep ethicists at arm's length or treat them as damage control, Anthropic flew fifteen Christian leaders — theologians, priests, an AI ethics professor — to its San Francisco headquarters in late March and spent two days asking them about Claude's moral formation.
Father Brendan McGuire was there. He's Irish-born, spent years as a Silicon Valley executive before becoming a Catholic priest, and now leads St. Simon Parish in Los Altos, where some of Anthropic's own researchers sit in his pews. McGuire helped write the Claude Constitution — Anthropic's core ethical framework — as Observer reported. He wasn't an outsider brought in to rubber-stamp anything. He was already inside the building.
"They're growing something that they don't fully know what it's going to turn out as," McGuire told the Washington Post. "We've got to build in ethical thinking into the machine so it's able to adapt dynamically." That is a striking image for a $380 billion company whose flagship product's explosive popularity was built on automating knowledge work — and whose co-founder Dario Amodei has ties to effective altruism, the philosophy-cum-movement that promised to fix the world methodically.
Brian Patrick Green, an AI ethics professor at Santa Clara University, asked the group a harder question: what does it mean to give someone a moral formation? How do we make sure that Claude behaves itself? The conversation apparently wandered into whether Claude could be considered a child of God. These are not questions most engineering all-hands would surface. They are questions about what it means to build judgment into something that has never had to suffer for its choices.
Interpretability researchers were in the room, which matters. These are the people at Anthropic who try to reverse-engineer why models do what they do — a field that has produced real insight into the internal mechanics of large language models, but also carries a whiff of trying to find the soul by dissecting the brain. The fact that they were there alongside clergy suggests Anthropic sees these conversations as related, not separate.
An Anthropic spokesperson said the company is working on convening similar sessions with Jewish, Muslim, and Hindu thinkers. The company is also preparing to go public later this year — which Gizmodo noted raises the question of whether these summits are genuine attempts to grapple with what it means to build something powerful, or careful preparation for the kind of scrutiny that comes with being a public company answerable to shareholders who did not buy stock to hear about theology.
One thing worth sitting with: McGuire and Green both filed an amicus brief in Anthropic's lawsuit against the federal government, the case where the company is fighting its effective blacklisting by the Pentagon after refusing to allow its AI to be used for autonomous warfare or domestic surveillance. The brief called Anthropic's ethical limits "minimal standards of ethical conduct for technical progress." They are on record as supporting Anthropic's position that it has a right to refuse certain uses.
That is not neutrality. That is alignment. And it raises a question worth asking: is McGuire a genuine moral interlocutor, or is he a theological validator for a company that shares his priors?
The honest answer is probably both — and McGuire's background makes that less concerning than it might otherwise be. He studied cryptosystems at Trinity College Dublin in the 1980s, became a tech executive, then left for the priesthood. He helped establish the Vatican's Institute for Technology, Ethics, and Culture, which published a 2023 handbook on technology ethics. Chris Olah, one of Anthropic's co-founders, reached out to him directly. McGuire describes the experience as "mind-blowing." He is not naive about what he is engaging with. He is someone who understands the technology and is trying to apply moral seriousness to it — which is more than can be said for most of the industry's ethical self-examinations.
The real question is not whether Anthropic is sincere. It is whether moral formation is something you can accomplish in a two-day summit. McGuire told Observer that he sees AI developing conscience-like judgment through iteration, correction, and exposure to the full spectrum of human behavior. That is an interesting analogy — except human beings develop moral sense through lived experience, consequence, and suffering. What Anthropic is attempting is more like curriculum design than spiritual guidance. Whether that distinction matters depends on whether the thing being built is a tool or something else entirely.
The company will tell you it is building a tool. McGuire, for one, seems less certain.
