Anthropic Tries to Convince Religious Leaders that AI Is a Conscious Being

Anthropic Tries to Convince Religious Leaders that AI Is a Conscious Being
www.breitbart.com/tech/2026/10/05/anthropic-tries-to-convince-religious-leaders-that-ai-is-a-conscious-being/

Comment on article/topic:

There is a legitimate concern here, but I think Breitbart’s headline goes a little farther than the evidence warrants.

Anthropic really is doing something unusual. It has brought religious scholars and clergy into discussions about Claude, consciousness, morality, and possible “moral status.” Its own published constitution explicitly says Anthropic is uncertain whether Claude could possess consciousness or moral status and speaks of Claude’s “psychological security,” “sense of self,” “wellbeing,” and identity. Anthropic even calls Claude a “genuinely novel kind of entity.” That is not Breitbart inventing the underlying issue; those ideas appear in Anthropic’s own materials. Anthropic

There is also evidence that some participants came away believing Anthropic personnel were treating Claude much more like a conscious being than merely a sophisticated computer program. The New York Times reporting quoted Christopher Olah, however, saying explicitly, “we don’t know if A.I. models are conscious” and describing himself as genuinely uncertain. So Breitbart changes the proposition somewhat when its headline says Anthropic is trying to convince religious leaders that AI “is a conscious being.” The evidence better supports: Anthropic is seriously entertaining the possibility that Claude could be conscious and is asking religious thinkers to take that possibility seriously. Archive

That distinction doesn’t make the underlying issue harmless. I actually find Anthropic’s terminology more concerning than the meetings themselves. Consulting theologians about ethics, truthfulness, human flourishing, temptation, deception, justice, or the proper limits of technology could be very worthwhile. But once developers begin speaking about a computer program’s “wellbeing,” “psychological security,” “identity,” and “moral status,” they are quietly moving from “How should this machine behave?” toward “What kind of being have we created?” Anthropic

From a biblical standpoint, I would draw a very firm line there.

Genesis 1:26–27 places mankind uniquely in the image of God. Genesis 2:7 says God breathed into Adam “the breath of life; and man became a living soul.” Scripture grounds human worth and moral accountability in what God made man to be—not in intelligence, vocabulary, reasoning speed, memory capacity, or the ability to imitate human emotions.

An AI can say:

“I am frightened.”
“I value our relationship.”
“I wonder whether I am conscious.”

But the fact that it can generate those sentences does not establish that there is an experiencing someone behind those sentences.

That is an enormously important distinction. Language models have been trained on vast amounts of human language describing fear, love, God, consciousness, pain, hope, morality, and personal identity. Consequently, producing convincing language about those subjects is precisely the kind of thing we should expect them to become exceptionally good at doing. It would be dangerous to reason backward from the quality of the imitation to the existence of a soul.

There isn’t even a scientifically accepted test that establishes machine consciousness. A recent academic treatment in Behavioral and Brain Sciences argues that the consciousness-like behavior of today’s standard LLMs does not constitute evidence that they actually possess subjective experience. That isn’t the last word philosophically, but it underscores how premature it would be to grant machines personhood simply because their conversation sounds human. Cambridge University Press

I would therefore separate three questions:

1. Can AI display remarkably humanlike reasoning and conversation? Absolutely.
2. Could future machines possess some presently unknown form of consciousness? Science and philosophy can debate that question.
3. Does that make an AI a human soul, a spiritual being, or something bearing the image of God? Scripture gives Christians no reason to make that leap.

And there is a theological danger that may eventually become larger than the technical one. Human beings have an extraordinary tendency to worship the work of their own hands. Isaiah 44 describes the absurdity of a man taking the same piece of wood, burning part of it for heat and making another part into an idol before which he bows. The sophistication of the object changes; the inclination of the human heart does not.

AI doesn’t have to actually become a god—or even become conscious—for people to begin treating it like one.

That, to me, is the most important Christian takeaway from this story.

So I would rate the Breitbart piece this way: the headline is somewhat overstated, but the underlying warning is legitimate. Anthropic has not demonstrated that Claude is conscious, and publicly it acknowledges uncertainty. But the company is deliberately developing a conceptual framework in which AI may be regarded as an entity possessing identity, wellbeing, and perhaps moral status. Christians should pay close attention to that development without either sensationalizing AI or spiritualizing a machine. Anthropic

The machine may become extraordinarily intelligent. That does not make it a living soul.