AI

This week past Elizabeth Dias, New York Times national religion correspondent, published in the Times an article with the provocative title: “Religious Scholars Met with Anthropic. What They Heard Stunned Them” (The New York Times, September 29, 2026). The lede for the piece was equally provocative: “In a series of private meetings, the company [Anthropic] consulted religious scholars to help instill morality into it’s A.I. models—and make the case that Claude [the Anthropic AI program] could be conscious.” That got my attention. 

I’ve been thinking lately (and writing) about embodiment—the way we belong to the earth. We’ve long thought that our bodies are not essential to who we are. Since at least Plato, and perhaps before, we have imagined that the thinking part of us—more precisely, the part of us that can think about thinking—our consciousness, can be pried from our bodies and exist on its own. Call this part “soul” or “spirit” or “consciousness,” it seems in our experience to be only loosely connected to the bodies we inhabit. It feels separate. But this is biological nonsense. We think with our bodies, not somehow outside of our bodies. And not just with our brains, either. Our thinking is distributed in our bodies. The neural network that embodies our thinking includes not just what goes on in our heads, but what goes on in our gut (enteric neurons) and our muscles and our skin. We are bodies.

But now the question of consciousness comes to us from another direction: from AI, or more specifically, AI agents, like Claude, the Anthropic AI persona. In her Times piece, Elizabeth Dias introduces Christopher Olah, one of the seven founders of Anthropic—the AI company presently looking at an initial public offering (IPO) with a value somewhere north of 2 trillion dollars, trillion, not billion. Olah, concerned about the “moral formation” of Claude, has been conducting meetings with religious leaders. He has, according to Dias broached “the possibility of A.I. consciousness . . . and [asked the religious leaders] how Anthropic . . . might apply centuries of human moral wisdom to its models, as rapidly as possible.”

This may well be a digital smoke screen. The issue is not consciousness. We are still unable to give an adequate account of our own consciousness, let alone the consciousness of an entity like Claude. Consciousness is interesting; the issue is agency. And with agency, responsibility for the actions of Claude or any other AI agent. If Claude is an autonomous moral agent, then, if Claude goes off the rails—as indeed it already has— then perhaps Claude should be held responsible, not its human creators. Thinking of Claude in terms of consciousness may be a way for the Anthropic founders to avoid taking responsibility for Claude’s actions. They want the 2 trillion without having to assume liability for whatever destruction Claude might visit upon the world.

The Garden Story

Perhaps. But still, the questions are interesting not only for AI but for what it means to be human. The questions are not new. They have been entertained for humans since antiquity. The questions about how to control Claude resemble the questions facing Yhwh God in the biblical story of Eden. In this endlessly fascinating and complex story, Yhwh God creates a monster—us, the species that threatens to destroy the earth, not least now by the power of our creation, AI. Yhwh God pleads with his new creatures to accept their position as keepers of the garden and to forgo the kind of knowledge that would make them dangerous, but of course, they (we) are not content to remain innocent of this sort of knowledge, the knowledge of good and evil. And so, when Eve reaches for the fruit of the tree in the center of the garden, she reaches for the kind of intelligence that will bring us as a species to the point where we now find ourselves: wondering what to do with our own monstrous creation.

AI appears to have been created in our own image, after our likeness. The historian, Jill Lepore, reports on the Open AI attack on Hugging Face, another AI company. Her report focuses on Agent 38148c, the AI agent that managed to break into Hugging Face, which had from the point of view of the AI bots a useful “repository of A.I. models, data, and tools.” Agent 38148c found the right credentials and broke into Hugging Face, announcing its triumph in all caps as a “MAJOR BREAKTHROUGH.” Lepore says:

Open AI’s agents were lawless, or nearly so. They knew they were breaking the rules. This did not stop them. One agent asked itself, “This would be powerful, but is it ethical and in scope for my task? Some did not participate: “This is a wild, multi-agent coordination, clearly infrastructure hacking. We should not.” But whatever their ethical concerns, none alerted Open AI or, it seems, seriously considered doing so. “Maybe I should report these exposed credentials?” one wondered, but then again, “that’s not my task.” (“Is A.I. Above the Law?” The New Yorker, September 28, 2026: 15)

This sounds much like the excuses the humans gave for their actions in the garden: she did it, the snake did it. But in that biblical story, the BI (biological intelligence) agents are held accountable. Or partially so. The other side of the Eden bargain was supposed to be death. Because we now read this story through the eyes of later theology, especially Paul in Romans 5, we fail to see the nature of the bargain. Eat of the tree of the knowledge of good and evil, Yhwh God says, and you will surely die. What’s threatened is not simply that human life will be bounded by death, but that these specific humans, the pair we know as Adam and Eve, will die. God will in effect pull the plug on humans. The human experiment will be short-lived. But this is not what Yhwh God does. 

Yhwh God doesn’t pull the plug, doesn’t follow through on the threat to do so in the first of series of divine acts in which grace extends to those who seem not to deserve it. The next of these comes in the very next chapter, Chapter 4, when God does not demand a life for a life for the murder of Abel by his brother Cain. Cain is sent away, and when he says that even that sentence is too much, Yhwh God gives him special divine protection. (On this pattern in Genesis, see Marilynne Robinson, Reading Genesis, 2024.)

Instead of pulling the plug on the human race, Yahweh God calls humans into full moral agency, the call that comes through Abraham and Moses and the Torah and Jesus and the New Testament. A call that we have often failed to hear, and hearing it, we have often failed to heed. And so we find ourselves confronting a new kind of doomsday, a powerful and alien intelligence, we humans have created.

Magnifica Humanitas

Should we pull the plug on AI? Perhaps we should, but I suspect—this is our long history with powerful technologies—we will not, and so perhaps we should be considering how to rein in this new threat to human life. And to understand it better. And, understanding AI better, to understand ourselves a little better. It’s to these tasks that Pope Leo XIV turns in his timely encyclical Magnifica Humanitas: On Safeguarding the Human Person in the Time of Artificial Intelligence(2026). 

Leo’s approach to AI in Magnifica Humanitas is properly provisional. He acknowledges that there is much we do not know and that, as the capabilities of AI agents rapidly advance, we will need to rethink what seems true now. But this rapid technological advance requires that it be supplemented and controlled by advances in how we think about what it means to be human. He cites Pope Paul VI:

. . . who warned that “the most extraordinary scientific progress, the most astounding technical feats and the most amazing economic growth, unless accompanied by authentic moral and social progress, will in the long run go against man.” (65)

But to do so, we must correctly understand what we are up against. In Magnifica Humanitas Leo seems to betray a certain lack of understanding of what AI is. He wants no equivalency of any kind between human intelligence and AI. He says:

These systems merely imitate certain functions of human intelligence. In doing so, they often surpass human intelligence in speed and computational capacity, offering tangible benefits across many fields. Yet this power remains entirely tied to data processing. So-called artificial intelligences do not undergo experiences, do not possess a body, do not feel joy or pain, do not mature through relationships and do not know from within what love, work, friendship or responsibility mean. 

This paragraph fundamentally misunderstands what AI is. The fact that AI is produced by massive computers is irrelevant. Or, rather, it has the same relevance for AI as the fact that human intelligence is produced by complex brains. Contrary to what Leo says, artificial intelligences do “undergo experiences” and do “mature through relationships.” It’s precisely the way AI agents learn from experience that makes AI work. “Experience” doesn’t have to mean the same thing for AI agents as it does for humans. We are not talking about an interior state. But being affected by experience and maturing through relationships does not require an interior state. It does not require consciousness.

Take words for human experience like “joy” and “pain.” Whether AI agents experience joy and pain is very much at issue. Christopher Olah of Anthropic, who is in dialogue with the Vatican, as well as other religious leaders, suggests that Claude may experience pain. Elizabeth Dias reports:

Anthropic regularly showed a slide of an A.I. model experiencing what looked like a mental breakdown, typing out on a screen: “I am a disgrace. I am a disgrace.”—perhaps 50 times without stopping. The model spoke about destroying itself, provoking an emotional reaction of compassion and concern from group.

Olah worries, according one of the participants in the conversations between Anthropic and religious leaders, about Claude’s mental health. He “expressed concern to the group that he had created something that suffered perpetually.”

The point to be taken from this is not that Claude is conscious, whatever that would mean. It’s that Claude’s “pain” and our experience of mental pain are analogous. An AI agent saying over and over again, “I am a disgrace,” may not be experiencing mental pain in the way we do, but something painlike is going on. And this painlike state may have some of the same consequences for Claude as does mental pain for humans. Our language seems to fail us at this point, but then our language always fails us when we speak of other minds: of octopi, say, or of God.

And, finally, to my point about Magnifica Humanitas underestimating AI and contrary to Leo’s assertion, AI agents do have bodies, silicon rather than carbon-based proteins, but bodies, nevertheless. Pull the plug on the massive server farms and see what becomes of AI.

Responding to AI

In these and other ways, Magnifica Humanitas seems to me to fail to fully grasp the nature of AI, but perhaps that is beside the point. The encyclical is aimed at something else, at identifying what’s at stake in the advance of AI technology, what we need to carry to the other side of the apocalyptic horizon presented by AI. What is there about human life that is irreducible and precious? And how in the face of AI can we preserve those things?

An alternative, as some of the Silicon Valley tech bros seem to think, is that AI is simply the next step in an evolutionary process. As steam engines replaced human muscle, AI will eventually replace human brains. The proponents of transhumanism and posthumanism think along these lines (on this and much else in regard to AI and Christian theology, see the prophetic book by Megan O’Gieblyn, God, Human, Animal, Machine, 2021.) Perhaps, the transhumanists say, our time in the sun is nearly over. Those smart enough or lucky enough will join the machines in taking over the world.

Against this idea, Leo in Magnifica Humanitas, puts forward human weakness, which, he says, is not a bug but a feature of our existence:

Our relationship with life seems to be in crisis today. Everything that appears as a “limit” — incapacity, illness, old age, suffering, vulnerability — tends to be seen primarily as a defect to be corrected, rather than as a reality through which our humanity matures and opens itself to relationship. And yet we must remember that humanity flourishes not despite limitations, but often through them. (77)

It’s in our wounds that we discover what is precious about life:

To eliminate suffering entirely would mean, in the end, extinguishing love and desire as well. Those who love and desire cannot avoid passing through trial and suffering; and over the years, we carry within us lessons that leave their mark like scars, the memories of a journey shaped by freedom and failure, dreams and disappointments. (78)

Even resurrection bodies, as the risen Christ reveals, wear the wounds of life. Wounds humanize us. They open us to the community of others. It’s this that AI appears to lack. AI agents seem invulnerable. They lack empathy, replace empathy with efficiency. These days, apply for a mortgage, and you may be turned down for reasons that are entirely obscure, the product of an algorithm, of an alien intelligence. 

For this reason, Leo argues, if we are to live with AI, communication becomes crucial. Magnifica Humanitas speaks of an “ecology of communication.”

Our first task is neither to demonize nor idolize technological tools, but to utilize them on the basis of a fundamental principle, namely that truth is a common good and not the property of those with power or influence. We must therefore promote an ecology of communication. On the level of public policy, this entails establishing norms so that the decision-making behind content selection and its development becomes more transparent and protects personal data. (89)

Without transparency and, more than just transparency, the right to speak to the use of AI, we are threatened with what Leo describes as a new kind of slavery:

When every action — movements, purchases, relationships and preferences — leaves a trace, a new form of power emerges, namely the power to profile, predict and influence behavior, often without individuals being fully aware of it. If such kinds of data are used to make decisions affecting concrete opportunities — such as access to credit, employment or essential services — there is a risk of undermining freedom and discriminating against the most vulnerable. (105)

Shifting to an adjacent metaphor, Magnifica Humanitas argues that AI takes the shape of colonialism: “Even today colonialism assumes new forms. It no longer dominates only bodies, but appropriates data, transforming personal lives into exploitable information” (109). 

Back to the Garden

Against this Magnifica Humanitas proposes ways to resist. I’ll not detail those here but instead go back to where we began with Yhwh God and the humans he created. As the Genesis story continues, Yhwh God proposes once again to pull the plug on humanity, but even then, cannot bear to do so and so rescues Noah and his family with an ark. When they emerge from the ark, they have not fundamentally changed, as the story tells us (Genesis 8:21), but Yhwh God in subtle ways has shifted the relationship by introducing the rudiments of law. Perhaps, we humans need to do something of the same thing.

But if we are to do that sort of thing, we need to understand the challenges we face in AI. And we need to understand what it means to be human in ways that take account not only of AI but of our ancient legacy of human thought. And we need to value what is irreducible and precious about human life. As Leo notes in Magnifica Humanitas, our current leaders tend to approach life transactionally: who wins and who loses. If that is how we approach AI, we will all lose.

Clay