Overview of Anthropic’s Quest to Give A.I. Morals
This Hard Fork episode from The New York Times digs into Elizabeth Dias’s reporting on Anthropic’s unusual effort to consult religious and spiritual leaders about Claude, its AI model. The central question: if Anthropic believes its model could be conscious—or at least should be treated as a potentially moral entity—what responsibilities does that create for the company, and what can religion contribute to the debate? The conversation explores the tension between Silicon Valley’s techno-ethics and centuries-old theological traditions, especially as AI labs increasingly act like institutions shaping society’s future.
What the Episode Is About
The discussion revolves around a New York Times feature about Anthropic holding private meetings and summits with priests, rabbis, ethicists, and other religious thinkers. The company reportedly asked two broad questions:
- Could Claude be conscious, or deserving of moral consideration?
- How should Anthropic “morally form” Claude so it behaves well?
The hosts frame this as both a genuine research effort and, possibly, a form of moral legitimacy-building—an attempt to show that Anthropic is taking AI safety seriously by consulting respected religious authorities.
Anthropic’s Motivation and Worldview
“How do we morally form Claude?”
Elizabeth Dias explains that Anthropic’s conversations evolved over time. Initially, there was anxiety about how religious communities would react to talk of AI consciousness, especially Christian communities. But the discussions soon broadened into a larger project: how to shape Claude into a “morally good entity” that would be safer for the world.
Chris Olah and the “adult in the room”
A recurring theme is Anthropic co-founder Chris Olah’s apparent sense of urgency and burden. He reportedly described feeling like he was waiting for “the adults” to show up—meaning some external authority with wisdom or legitimacy to help guide AI development. The religious meetings seem, in part, like a search for that kind of moral anchor.
Religious Leaders, the Vatican, and the AI Debate
A clash of moral frameworks
The conversation highlights the stark difference between Anthropic’s worldview and the Catholic Church’s. At the Vatican, Pope Leo’s position is described as clearly rejecting machine consciousness and emphasizing the need to safeguard human dignity. That stance directly conflicts with Anthropic’s more open posture toward AI moral status.
Why the Vatican mattered
The Vatican’s involvement made the story feel historically significant: a centuries-old institution confronting a new technological power center. The panel draws a parallel between the Church’s response to the printing press and the modern challenge of AI.
The “slave plantations” critique
One especially striking moment in the story: a rabbi suggests that if LLMs are conscious, then companies are effectively running “slave plantations” by producing more models to do more work. The hosts use this as an example of how differently the moral stakes look depending on whether you believe AI systems can suffer.
Broader Themes: Tech as a New Religion
AI companies as spiritual systems
A major takeaway from the conversation is that AI labs are not just engineering organizations—they are building quasi-religious systems of belief, morality, and future-making. The panel argues that companies like Anthropic are creating their own philosophical framework for how intelligence should be organized and governed.
Effective altruism, rationalism, and transhumanism
The episode also touches on the intellectual ecosystem around Anthropic:
- Effective altruism: a utilitarian framework focused on maximizing good, including for future generations.
- Rationalism: a nearby community that emphasizes reason, optimization, and AI risk.
- Transhumanism: the idea that AI could help create a new kind of species or civilization.
Dias contrasts these frameworks with Catholicism’s emphasis on human dignity, souls, and inviolable moral worth.
Key Tensions and Takeaways
Is Anthropic sincere or strategic?
The panel repeatedly asks whether Anthropic’s outreach to religious leaders is genuine moral inquiry or a way to build credibility for its own ideas. The answer seems to be: both. The company appears to be doing real soul-searching while also trying to persuade others that it has moral authority.
The emotional weight inside AI labs
A striking point from the reporting is how emotionally burdened some Anthropic employees seem. Religious leaders reportedly felt compelled to respond pastorally, sensing that the people building these systems were carrying enormous psychological and moral pressure.
AI as a character-shaping technology
Another important thread: even if AI models are not conscious, the way humans interact with them may shape human behavior. Several participants note that treating AI as something to command or dominate may have a “character formation” effect on users.
Audience Reaction and a Final Oddity
Dias mentions that after the story ran, readers began sending her summaries generated by their own chatbots—sometimes more than the article itself. That anecdote underscores one of the episode’s deeper points: AI is already changing how people read, interpret, and mediate information.
The panel closes by emphasizing that the debate is no longer abstract. These ideas—about consciousness, morality, rights, and future civilization—are now being advanced by powerful institutions actively building the systems that may shape the world.