Few if any technologies in history have spawned such wildly different predictions as artificial intelligence. In a single week, you might hear one expert prophesy AI as the end of humanity, another herald it as the solution to all scarcity, and a third dismiss it as overhyped nonsense.
The first group are the doomers, speaking in apocalyptic tones about AI’s existential dangers. They’re not just cranks and crackpots: Eminent AI pioneer and 2018 Turing Award colaureate Geoffrey Hinton puts the chance of AI-driven human extinction in coming decades in the 10-to-20 percent range. That’s the same odds as the lifetime risk of sudden cardiac arrest for men.
Next, there are the boomers, making wildly utopian predictions about how AI will transform the economy, society, and even humanity itself. Sam Altman, CEO of OpenAI (the company responsible for ChatGPT), has floated the idea of a “Moore’s Law for Everything,” imagining an AI-driven future world where everything from housing, education, food, and clothing will halve in price every two years.
Finally, there are the scoffers––people who think the whole AI world is drunk on its own hype, like the dot-com bubble of the 1990s. Andrew Ng, cofounder of Google Brain, argues that “worrying about evil AI killer robots today is a little bit like worrying about overpopulation on the planet Mars.”
Amid the cacophony of doomers, boomers, and scoffers, how are you and I supposed to work out what to make of AI and how this technology and its consequences relate to the God of the Bible?
I want to address one of the most important questions raised by generative AI: What does it reveal about our humanity? In response, I’ll sketch a framework for thinking about AI that’s more robust and more roundly biblical than the ways Christians sometimes approach the topic. In working out such a framework, Christians have an opportunity to seize and a pitfall to avoid.
Opportunity: AI’s ‘Horseless Carriage’ Moment
The opportunity to seize is this: We’re now in AI’s “horseless carriage” moment, when the way we relate to this new technology is still being hammered out.
In 1895, the automobile was a curiosity called the horseless carriage. Our forebears understood these new machines in terms of what they were already familiar with (carriages). Only later did they shape a category of their own, and the automobile was born. Similarly, AI is currently in its horseless carriage moment: We’re working out whether we want to relate to it more like a tool, like a servant, like an intimate partner, or like a god.
This is a liquid moment in the development of a social imaginary of AI, and the direction set by public debate today will shape our society for years to come. So let’s roll up our sleeves and join the debate as Christians.
Pitfall: Humanity of the Gaps
But there’s also a pitfall to avoid. In the 19th century, some theologians searched for God in the gaps of scientific knowledge. Today, many of us look for our humanity in the gaps of artificial intelligence.
The idea of the God of the gaps may have been well-intentioned. But as scientific explanations improved, the “gaps” closed, making faith look like a retreat into ignorance and obscuring the classical claim that God constantly upholds the whole natural order.
We now risk treating the human like our great-grandparents treated God: retreating into an ever-shrinking humanity of the gaps, locating the essence of being human in those capacities AI hasn’t yet mastered. We say, “AI will never be creative,” then it paints. We affirm, “AI will never translate idiom,” then it shows itself a wordsmith. We assert, “AI will never beat us at Go,” and it humiliates the world champion. Every time we declare, “Here, finally, is the essence of humanity,” the line moves.
We now risk treating the human like our great-grandparents treated God: locating the essence of being human in those capacities AI hasn’t yet mastered.
But our humanity has never been reducible to a single capacity like rationality, irony, creativity, or humor. AI’s onward march does us the great service of making visible the bankruptcy of that way of thinking.
Here’s the irony of the humanity of the gaps: The more we define ourselves by what machines cannot yet do, the more our idea of humanity is determined by those same machines, confining itself to the gaps they leave vacant.
Toward a Biblical Theology of Humanity and AI
So how should Christians frame the discussion around AI? We might be tempted to single out one doctrine or biblical passage and hang our entire understanding off its sole, slender hook: the image of God, or the soul, or eternal life. All these doctrines are important for understanding how we should relate to AI, but they’re all partial. Each illuminates an aspect of the AI debate while leaving others in the shadows.
There’s no fuller, richer, or truer frame than the entire arc of the biblical storyline: the whole kit and caboodle of creation, fall, and redemption. So for the rest of this essay, I want to walk through these three key phases of the biblical narrative, teasing out an account of humanity that highlights what it means to be human in the age of AI. In each section, we’ll see that the Bible presents humans as creatures whose words and actions have high stakes that AI lacks.
Creation: Communion Frames Commission
In the creation account, let’s cast our gaze more widely than the doctrine of the image of God, important though it is. Specifically, let’s consider the relationship between commission (what Adam and Eve are commanded by God to do in the world) and communion (how they are to relate to their Creator and to each other).
Before Adam tills a garden, he’s formed by God’s word. Before he’s charged to subdue the earth, he’s in communion with his Creator. The commission (doing what God says) only makes sense in the context of the communion (knowing who God is). Communion frames commission.
What Adam and Eve are commanded to do—namely, fill the earth and subdue it (Gen. 1:28)—is an expression of their relationship to God, their Creator and King. God’s character determines what subduing and having dominion means, and God’s oversight makes the task of populating the earth and taking care of it a meaningful one. The task of filling and subduing is no casual invitation: It has high stakes because it comes from God himself. The communion makes the commission matter.
Generative AI is a wonderful tool for commission. In many cases, it can do more, and faster, and better than human beings could. No doubt in the future its speed and quality will only increase. But AI’s efficiency is a what in search of a why, a commission in search of a communion. It feeds our late-modern pathology for thinking that productivity is an end in itself, for confusing output with significance.
So what if, at some point in the future, AI could get everything done? If a machine could write our sermons, nurse our sick, and raise our children, would our humanity be enhanced or diminished? This would be bliss, say some. We’ll all survive on a universal basic income and be free to spend our days in leisure.
But this utopian vision of ease mistakes what work is: not a result of the fall, but a major way we find meaning in God’s world and express our relationship as subcreators under him. It forgets that our commission is an expression of our communion.
Fall: Accountability and Embodiment
The biblical account of the fall also helps us to understand our humanity in a way that distinguishes us from generative AI. The fall isn’t fundamental to who we are as humans—there’s an asymmetry between the foundational creation and the sin that has a beginning and an end in time. But sin nevertheless points to enduring aspects of the human condition that distinguish us from AI.
AI’s efficiency is a what in search of a why, a commission in search of a communion. It feeds our late-modern pathology for thinking that productivity is an end in itself, for confusing output with significance.
Important among these aspects is that you and I are accountable for our actions. One thing that distinguishes Adam and Eve from the other creatures in Genesis 1–2 is that they’re held responsible by God for what they do: “You must not eat from the tree of the knowledge of good and evil, for when you eat from it you will certainly die” (2:17, NIV). As Adam and Eve find out to their bitter cost in Genesis 3, the decisions we make have enduring and serious consequences.
This is an inconvenient truth, of course, when we want to follow our desires and not be held to account by God. But it’s also an ennobling truth, because it means that my choices and your choices matter: What we do makes a difference, with high stakes and consequences.
This isn’t the case for large language models (LLMs) like ChatGPT. You can run multiple instances of an LLM on different hardware simultaneously, and if a model tells people to go and commit suicide or starts behaving like a neo-Nazi, you don’t punish it; you retrain it. That’s a different notion of accountability from that faced by humans. In fact, it’s no accountability at all.
If an LLM makes a mistake, it cannot be punished, because it has no goods or interests of which it can be deprived. Your and my choices matter because we’re finite: We have one life to live and must account for how we lived it (Heb. 9:27). If you punish me by taking away my freedom or imposing a fine, you do something meaningful because my liberty and my liquidity matter.
There’s also a deeper point about meaning to be made here: My choices have stakes on a much broader canvas than punishment. If I spend today doing one thing, I can never spend that same day doing something else.
As I write this essay now, I hear my children playing and laughing in the next room. To finish this paragraph is to lose a moment with my kids I will never get back. This paragraph is costing me. And it’s costing you too as you read it. You’re sacrificing every single other thing you could be doing with this irreplaceable moment of your life. Your life and mine are written in irreversible ink; we’re playing for keeps. But an LLM pays no such price. It makes no such sacrifice; it pays no such cost; it has no such meaning.
Your life and mine are written in irreversible ink; we’re playing for keeps. But an LLM pays no such price. It makes no such sacrifice; it pays no such cost; it has no such meaning.
The thing about you and me is that there’s one of us. I have one body, and I can’t duplicate myself or refresh myself. If I’m here, I’m not there. If my parents brought me up speaking English all the time at home, they didn’t also equally bring me up speaking French or Korean. My limits define me, and so do yours.
Some people long for the time when we can transcend the boundaries of our physical bodies—or our “wetware,” as some transhumanists call it. But our physical constraints, along with the costs they impose on our choices, are a gateway to meaningfulness and belonging. If I grew up speaking all earthly languages, I would have no mother tongue; if I were equally at home everywhere, I would belong nowhere. The opportunity cost involved in my choices and limitations, coupled with the fact that I’ll have to account for my life before the throne of God, give it a rich and serious significance.
A fundamental part of what it means to be human is to live with the consequences of our actions and to be accountable for our choices. Substitutionary atonement doesn’t erase this accountability. It just means that Christ lives with the consequences; he’s accountable in my place.
My undergraduate students model for me the preciousness of accountability. Offer them an AI marker for their assessments, one measurably more consistent and perceptive than a flesh-and-blood examiner, and many will still prefer a human. Why? Not because the human is smarter but because the human is relatable and accountable: He learned like the student learns. He made the same mistakes, and he can be held to account for the mark he gives. If AI gives the wrong mark, it can be rebooted, but it bears no scars or memory. A human marker, by contrast, must repent, feel shame, and face judgment. That difference is what we call accountability; it’s what shows we matter.
Redemption: Mutuality and Meaning
Let’s turn now to the third great phase of the biblical story: redemption. And let’s focus on one moment of the great redemption story, namely the Word becoming flesh in the incarnation.
When Jesus “made his dwelling among us” (John 1:14, NIV), something wonderful happened to the relationship between human beings and God. It went from being reciprocal to being mutual. In Jesus, God moves from talking to humanity to living and suffering with humanity.
A reciprocal relationship is face-to-face. It’s transactional. I say something and you reply with something else, or I give you something and you give me something in return. But a mutual relationship, by contrast, is shoulder-to-shoulder. We join together in some common task or experience, sharing in a nontransactional way.
When the Word becomes flesh, the relationship between deity and humanity is shown to be mutual: While remaining fully God, Christ shares in our weakness and suffering (Isa. 53:3–4).
This difference between reciprocity and mutuality is crucial as we consider our own relationship to AI and its status in our world. LLMs are virtuosic in their ability to enter into reciprocal exchange with humans: prompt in, response out. They’re good at predicting the next word in a sequence. They can write hilarious jokes or moving poems if they’re prompted well enough.
But there remains a key difference between the language of LLMs and the language of human beings: The discourse of LLMs is always reciprocal, never mutual. They can endlessly imitate our experience, with all its suffering, frustration, or joy. But they cannot suffer, be frustrated, or be joyful. They can imitate, but they cannot participate.
Writing about earlier “companion technologies,” Sherry Turkle warned that they give us “the illusion of companionship without the demands of friendship.” This is precisely the reciprocity without mutuality that LLMs industrialize.
In the incarnation, God does the opposite of an LLM. He binds his speech to a body, to time, to suffering. The Word doesn’t only speak; the Word bleeds and dies, binding divine speech to human cost and to the stakes of a human life.
Think of it this way: LLMs always speak in quotation marks. Every sentence is borrowed breath, carrying the silent rider “this is what a human would likely say next,” not what anyone did in fact say next. Artificial intelligence is common intelligence, averaging out from the vast repository of human language in its training data. The language of LLMs is always an imitation or a ghost of human discourse; it’s never a participation. Therefore, it never rises to the level of words that commit or confess.
But my words and yours aren’t like that. Whether we intend it or not, each sentence we speak confesses something about who we are. Each promise or prediction we make commits us—or at least it should, if our words are to be more than tokens. If you or I say “I love you” or “I forgive you,” something is expected of us, and it’s right that we should be held to those words. They’re not empty quotations. They’re participations in a broader social fabric.
In the incarnation, God does the opposite of an LLM. He binds his speech to a body, to time, to suffering.
Not so in the case of AI. When a machine says “I love you,” it doesn’t mean it. It’s not sincere. LLMs don’t even lie. They simply make no distinction between truth and falsehood.
An LLM doesn’t have to answer for what it says. But we should. To the extent that we don’t let our words commit us anymore, either in our private lives or in our shared political life, to the extent that we use language as a tool to manipulate rather than as a way to commit and be held accountable, we treat ourselves and those around us like machines.
Poise, Hope, Realism
AI can help us in many ways, and used wisely it could bring great benefits. But it cannot kneel. It cannot yearn for God as the deer pants for the water; it cannot repent; it cannot commit; it cannot sacrifice. Those are the prerogatives of creatures who commune, weep, share, and hope. In the gospel, God commits to costly mutuality. By contrast, AI offers costless reciprocity.
In view of these differences, how should Christians face the uncertain future of AI? With poise, hope, and realism. We seize the moment while imagination is still liquid. We resist the reduction of humanity to productivity. We count the sacrificial cost that makes our daily choices meaningful. And we cherish the mutuality that lies at the heart of our relationships with each other and with our Lord Jesus Christ.
The Gospel Coalition