As tech companies race to build ever-more-powerful artificial intelligence, some of them are asking an old institution for help with a new problem: how to teach a chatbot right from wrong. Their answer, increasingly, is religion.
In recent months, AI executives have sought input from philosophers and faith leaders on how to make models "morally good." The outreach runs from formal consultations to Vatican appearances — and, according to a report from the Associated Press published Sunday, it is raising concerns among scholars of religion that the companies misunderstand what they are tapping into. For readers tracking the latest AI developments, the story signals how central moral alignment has become to how these companies now operate.
"Questions a Lab Cannot Answer"
Anthropic, the AI company widely viewed as taking ethics most seriously, has been among the most active. At the Vatican's presentation of Pope Leo XIV's encyclical on AI earlier this year, Anthropic co-founder Chris Olah told the audience that moral questions surrounding AI exceed what companies can resolve on their own.
"These are not questions a lab can answer," Olah said. "They are questions traditions like yours have carried for millennia."
Anthropic said in a May blog post that it had met with "scholars, clergy, philosophers and ethicists from more than 15 religious and cross-cultural groups," a pool it says it hopes to broaden. The company framed the effort broadly: "This work isn't about aligning our models with any one tradition's worldview. We want Claude to draw from a full range of viewpoints — religious, secular, political — with equal depth and rigor."
OpenAI CEO Sam Altman has articulated a different, more aggregate vision. "What I think ChatGPT should do is reflect that weighted average, or whatever, of humanity's moral view," Altman said last year.
Inside Anthropic's Faith Meetings
Meghan Sullivan, a philosopher of religion who leads an initiative on ethics and AI at the University of Notre Dame, attended one of Anthropic's meetings earlier this year. In contrast to conversations she has had with other tech companies, she came away impressed by the rigor of the engagement.
"The culture inside the company is similar to that on a good college campus in the sense that you've got a collection of very intellectually curious people," Sullivan said. "These meetings are pretty nerdy — I don't know if that's the best way to put it — in a good way."
The meeting, which included Christians from different denominations, was split into two parts: the company first explained its models "in layman's terms," and the faith leaders responded with feedback and discussion. According to Sullivan, the questions ranged from "Do humans have souls?" to "Should third graders be able to access a chatbot for literacy education?"
The Vatican Becomes a Regular Stop
Anthropic is not alone in the pews. Representatives from Meta, Google and Amazon have also met with the Vatican, while Microsoft has been described as particularly vigorous in its religious outreach.
The engagement reflects a calculation unique to this moment in AI development. As executives sound alarms about AI's destructive potential and debate what "morally good" model behavior even means, religious traditions offer several millennia of accumulated thinking about character, obligation and harm — material that no training corpus can fully replicate and that companies cannot easily source elsewhere at scale.
Anthropic's own framing makes the ambition explicit: "What we're after in these conversations is careful, accumulated thinking on how good character actually forms."
Scholars Warn: Religion Is Not a Rule Book
Experts and adherents alike caution that the companies' premise — that religion can supply an extractable set of moral principles — fails to reckon with how traditions actually work. Different worldviews often contradict one another, and treating religion as a database of rules risks distorting the very material being consulted.
Webb Keane, an anthropologist at the University of Michigan, argues that scholars have spent decades pushing back against the view of religion as a set of principles divorced from the "messy, practical, ad hoc" ways adherents live.
"There's a very conventional way of thinking of religion as a rule book," Keane said. "It is certainly very tied to scriptural religion, and the scriptural religions at the end of the day really turn out to be the Abrahamic religions."
That assumption, he warns, risks prioritizing certain traditions while distorting others — particularly faiths that do not understand themselves primarily as systems of ethics.
Tomoko Masuzawa, a religious studies scholar and author of "The Invention of World Religions," goes further, noting that categories like "Buddhism" and "Hinduism" as clearly bounded, comparable traditions are largely a construct of European scholarship dating to the 19th century. "All these categories were not native," Masuzawa said.
What Is Actually Being Built
The scholarly critique matters practically because whatever companies learn in these consultations may shape how models answer moral questions for billions of users. An AI that treats religion as a source of weighted-average principles will behave differently from one trained to recognize that moral authority is contested, situated and lived.
For now, the companies deny any single-tradition agenda, and the faith leaders involved describe the engagement as serious and substantive. Sullivan's verdict on Anthropic — intellectually curious, rigorous, a bit nerdy — suggests the meetings are more than public relations.
But the deeper question the AP report surfaces remains open: whether any consultation process can translate millennia of lived tradition into model behavior without flattening it — and who decides which parts of humanity's moral inheritance make it into the weights.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
