Key Takeaways
- Anthropic (and OpenAI) are consulting religious leaders to explore how faith‑based moral frameworks might inform the ethical behavior of AI models like Claude.
- The initiative began with private meetings with Christian leaders and expanded to a multifaith “Faith‑AI Covenant” roundtable in New York, organized with the Interfaith Alliance for Safer Communities and attended by groups ranging from rabbis to Sikh and Greek Orthodox representatives.
- Anthropic says it is seeking “high‑order ethical truths” rather than trying to implant any single doctrine into its AI, echoing the historic role of the Kaaba as a repository of diverse symbols.
- Critics, such as Humane Intelligence CEO Rumman Chowdhury, warn that the tech industry’s earlier belief in universal AI ethics was naive and that turning to religion may only be a pragmatic stop‑gap for ethically gray situations.
- While the effort shows a genuine attempt to broaden moral input, it remains uncertain whether synthesizing diverse religious insights can reliably guide an AI in high‑stakes decisions.
Introduction
Before the rise of artificial intelligence, humanity already possessed a “black box” of moral guidance: the Kaaba, the black cube at the heart of the Sacred Mosque of Mecca. In pre‑Islamic times the Kaaba housed 360 sacred symbols from across the Arabian Peninsula, allowing a traveler to pray to a single source and feel covered “in the god department.” Anthropic’s recent outreach to faith communities mirrors this ancient practice—seeking a compact, all‑purpose repository of ethical insight that could be consulted when an AI faces dilemmas for which no explicit rule exists.
The Faith‑AI Covenant Roundtable
Last week, representatives from Anthropic joined OpenAI at a New York gathering titled the “Faith‑AI Covenant” roundtable. The event was not hosted by the tech firms themselves but organized by the Swiss NGO Interfaith Alliance for Safer Communities, which plans similar gatherings in China, Kenya, and the United Arab Emirates. Attendees included the New York Board of Rabbis, the Hindu Temple Society of North America, the Church of Jesus Christ of Latter‑day Saints, the U.S.-based Sikh Coalition, and the Greek Orthodox Archdiocese of America. Baroness Joanna Shields, a member of the British House of Lords, was cited as a “key partner” in the initiative.
Earlier Christian Consultations
The roundtable built on a series of private meetings Anthropic held last month with 15 Christian leaders. According to Anthropic, those sessions were intended to gather advice on the “spiritual development” of its Claude AI model and to receive guidance from moral thinkers. The company said it was “working on arranging meetings with moral thinkers who represented other groups,” indicating a deliberate effort to broaden the theological scope of its consultations.
Seeking Clarification
Gizmodo followed up with Anthropic for clarification on whether the Christian summit and the Faith‑AI Covenant roundtable were part of a single coordinated program, and whether the same staff members participated in both. As of this writing, Anthropic had not responded to the request, leaving the relationship between the two sets of meetings ambiguous. The Associated Press piece that reported on the roundtable likewise did not clarify whether the earlier Christian consultations were a subset of a larger, unified effort.
Broader Partnerships and Motivations
Beyond the faith leaders, the AP story notes that OpenAI also participated in the initiative, suggesting a shared industry interest in tapping religious wisdom for AI safety. The involvement of the Interfaith Alliance for Safer Communities signals an institutionalized approach to these dialogues, with plans to export the model to other regions. Baroness Shields’ participation adds a political dimension, hinting that policymakers may be watching how tech firms negotiate moral guidance from religious institutions.
Claude’s Constitution and Moral Uncertainty
Although the AP article records no specific religious edicts handed down by the assembled clergy, it does reveal Anthropic’s internal framing of the work. The company describes what it calls Claude’s constitution—a set of guiding principles aimed at enabling the model to “make the decision of a person with perfect values when there’s no way to write a rule for a situation that arises.” Anthropic admits that this effort is driven by worry: “we worry that our efforts to give Claude good enough ethical values will fail.” This candid acknowledgement underscores the difficulty of encoding ethics into software when real‑world scenarios are messy and context‑dependent.
Expert Skepticism
Rumman Chowdhury, CEO of the nonprofit Humane Intelligence, offered a blunt assessment of the tech industry’s evolving stance. She told the AP, “I think a very naive take that Silicon Valley has had for a couple of years related to generative AI was that we could arrive at some sort of universal principles of ethics,” adding, “They have very quickly realized that that’s just not true. That’s not real. So now they’re looking at maybe religion as a way of dealing with the ambiguity of ethically gray situations.” Chowdhury’s commentary captures the shift from confidence in abstract, universal AI ethics to a more humble, exploratory approach that seeks wisdom from humanity’s long‑standing moral traditions.
Assessment and Outlook
Anthropic’s strategy appears less about converting the AI to any particular faith and more about extracting high‑order ethical truths that transcend doctrinal specifics—much like a pre‑Islamic pilgrim might find whatever spiritual truth they needed within the Kaaba’s eclectic collection of symbols. Whether such a synthesis can reliably guide an AI in life‑or‑death decisions remains an open question; the efficacy of blending diverse religious insights into a coherent machine‑ethics framework is untested. Nonetheless, the effort signals a serious acknowledgment that technical fixes alone may insufficiently address the moral complexities AI will face. As the author concludes, “It probably can’t hurt, sorta like nodding at the pre‑Islamic Kaaba. But then again, only God knows for sure.” This reflective note captures both the pragmatic hope and the lingering uncertainty that now surrounds the intersection of artificial intelligence and religion.
https://gizmodo.com/anthropic-has-added-several-more-religions-on-its-quest-to-inject-perfect-morals-into-claude-2000756740