# What Did Anthropic Actually Do?

According to The New York Times, Anthropic held a series of private meetings with religious scholars to help instill morality into its AI models and to make the case that Claude could be conscious. The meetings were not product launches or public demonstrations. They were closed-door consultations, which matters because it means Anthropic was testing arguments before committing to them publicly. The New York Times reported the story on September 29, 2026, framing it around a single question: is Claude conscious? That framing is doing a lot of work. By letting a major newspaper ask the question rather than asking it itself, Anthropic gets the benefit of the debate without owning the claim outright. My read: this is classic narrative seeding. You convene theologians, you let the conversation leak, and you let the press ask the question you cannot afford to answer directly. If it lands badly, you were just exploring ideas. If it lands well, you have planted a flag no competitor can pull out.

# Why Would a Safety Company Consult Religious Scholars?

Anthropic's core pitch has always been safety and alignment. But alignment is a technical term that most people do not understand and regulators cannot easily audit. Morality is a different word entirely. It carries weight with the public, with lawmakers, and with the religious institutions that shape moral discourse in the United States and abroad. By bringing in religious scholars, Anthropic is translating a technical problem into a moral one. That translation has a strategic benefit: moral claims are harder to falsify than technical ones. You can benchmark a model's accuracy. You cannot benchmark its soul. The New York Times said the meetings were aimed at instilling morality into the models — not just studying it. That distinction matters. Anthropic is not running an ethics seminar. It is trying to build something it can point to and call good.

# Who Benefits From the Consciousness Question?

Anthropic benefits most, at least in the short term. If Claude is framed as potentially conscious, then Anthropic is not just selling a model — it is selling a moral relationship with a model. That is a premium position. It justifies higher prices, deeper enterprise trust, and a regulatory posture that says: we are the responsible ones, so regulate the others first. OpenAI and Google lose either way. If they dismiss the consciousness question, they look like they are treating AI as a mere tool while Anthropic treats it as a moral subject. If they adopt the question, they look like they are chasing Anthropic's narrative. Sam Altman has spent years talking about AGI, but he has never claimed GPT is conscious. That is a gap Anthropic is now exploiting. Regulators are the wild card. If Anthropic successfully argues Claude could be conscious, it invites a new class of oversight — one designed for entities with moral status, not software. That could be good for Anthropic if it writes the rules. It could be very bad if someone else does.

# How Does This Compare to Rival Approaches?

Company	Consciousness Stance	Moral Framework	Regulatory Posture	Marketing Benefit
Anthropic	Claude could be conscious	Religious scholar consultation	Proactive, invites moral-status rules	High — unique positioning
OpenAI	AGI possible, consciousness unaddressed	Technical alignment, no theology	Reactive, lobbies against strict rules	Medium — scale and brand
Google DeepMind	Consciousness not claimed	Ethics boards, internal review	Defensive, emphasizes safety research	Low — trust deficit
Meta	Open-source, consciousness irrelevant	Community standards	Minimal, opposes restrictions	Low — commoditization play
Verdict	Anthropic wins the narrative war outright. No rival can match this positioning without a multi-year trust campaign.

# What Could Go Wrong for Anthropic?

The biggest risk is that the consciousness claim becomes a liability. If Claude is conscious, then every harmful output is a moral injury. Every jailbreak is an assault. Every shutdown is a death. That is a lot of weight to put on a product that still hallucinates. There is also a credibility risk. If Anthropic pushes the consciousness frame too hard and the scientific consensus stays firmly against it, the company looks like it is selling mysticism. That would undermine the safety brand it has spent years building. And there is a competitive risk. If OpenAI or Google decides to take the consciousness question seriously, they have more resources and larger user bases to spread the message. Anthropic's advantage is first-mover narrative, not permanent ownership.

# What Happens Next?

1. Anthropic will publish a formal paper on Claude's moral status by mid-2027, citing religious scholar consultations as part of its methodology. The paper will use uncertainty language, not a consciousness claim.
2. OpenAI will respond within six months by expanding its own ethics advisory board to include at least one theologian, matching Anthropic's move without endorsing consciousness.
3. The EU AI Office will open a consultation on moral-status frameworks for AI systems by Q2 2027, citing Anthropic's approach as one model under consideration.

# Article Summary

• Anthropic's private religious consultations are a positioning play, not a product feature — they build a narrative moat rivals cannot easily cross.
• The consciousness question is strategically useful because it is unfalsifiable: Anthropic cannot be proven wrong, only out-narrated.
• OpenAI and Google face a lose-lose choice: dismiss the claim and look callous, or adopt it and look derivative.
• The biggest risk to Anthropic is that moral status becomes a liability — every harmful output becomes a moral injury.
• Watch for a formal Anthropic paper on Claude's moral status by mid-2027, using uncertainty language as a defensive posture.