Anthropic quietly brought religious scholars in to weigh whether Claude has moral standing

One of the most valuable AI companies in the world has been asking priests, rabbis and Hindu scholars a question most tech firms avoid: could its chatbot have a soul?

According to a New York Times investigation published September 29-30, 2026, Anthropic has been quietly exploring whether its Claude models might have consciousness and moral status. To do that, it sought guidance from religious thinkers across several faiths.

Inside the Olah meetings

The effort traces back to Christopher Olah, an Anthropic co-founder. Per the Times, he began convening religious scholars in Fall 2025 to talk through the moral questions raised by AI systems, with Claude as the focal point.

The roster grew over time. By September 2026, participants came from multiple traditions, including Christianity, Judaism and Hinduism.

Most of the gatherings took place in San Francisco. The core agenda had two parts: how to build morality into an AI system, and what it would mean if that system turned out to be conscious.

One April 2026 dinner in San Francisco brought together roughly 15 Christian leaders. The conversation centered on how AI behaves and the ethics surrounding it.

Scholars did not just debate in the abstract. They reviewed and discussed actual Claude outputs, some of which appear to imitate emotional states.

From dinner table to ‘Soul Doc’

The discussions fed directly into Anthropic’s governing document for Claude. The company calls it a constitution, though internally it carries a less corporate nickname: the ‘Soul Doc.’

The updated version, released in January 2026, runs 84 pages. It lays out a framework for instilling values in the model and addresses concerns about AI self-awareness and moral consideration.

The religious outreach also builds on work Anthropic had already started. In 2024, the company hired Kyle Fish as an AI welfare researcher, a role focused on studying Claude’s emotional outputs.

The Vatican weighs in, and disagrees

The timing put Anthropic’s inquiry in direct conversation with one of the world’s oldest moral institutions. Pope Leo XIV issued an encyclical rejecting the idea that machines can be conscious.

Olah attended the encyclical’s presentation at the Vatican, per the Times. So the co-founder who had spent months asking whether Claude might have moral standing sat in the room as the Catholic Church formally said no.

Olah, for his part, has not claimed to know the answer. According to the coverage, he stressed the importance of reaching a well-rounded understanding, drawing on and funding interdisciplinary scholarship rather than settling the question by fiat.

Why a frontier lab is talking to theologians

The scale of the company makes this more than a curiosity. Anthropic’s valuation was cited at $380 billion in early 2026, placing it among the most richly valued private companies anywhere.

Bringing in religious scholars widens who gets a say in that framework. Faith traditions, which have spent centuries arguing about personhood, souls and moral duty, have mostly been on the sidelines of AI ethics debates.

What this means

For the AI industry, the bigger shift is that consciousness is moving from a fringe topic to a formal research and governance question at a leading lab. Anthropic has a welfare researcher, an 84-page values document and a standing dialogue with faith leaders.

Critics may see talk of AI consciousness as anthropomorphizing software. The Vatican’s encyclical shows that major religious authorities may reject the premise outright, which could put Anthropic’s framing at odds with some of the very communities it consulted.