Christians, AI, and the Question Beneath the Question

Christians, AI, and the Question Beneath the Question

Christians, AI, and the Question Beneath the Question

I recently read Mike Duran’s article, On Christians Against AI, and found myself agreeing with quite a bit of it. Mike is right that Christians should take seriously what artificial intelligence means for creativity. He is right to emphasize the uniqueness of human beings as creatures made in the image of God. He is right that Christians should resist the temptation to exploit technology merely for profit. He is right that pastors, writers, musicians, and other creators remain responsible for what they put before others. And he is certainly right that our conversations about AI should be conducted without slander, self-righteousness, or technological witch hunts.

Where Mike and I differ is deeper. I have spent considerable time developing a theological position on artificial intelligence rather than simply deciding whether I personally like or dislike the technology. That distinction matters to me. My question is not merely, “Should Christians use AI?” I am asking something underneath that question: What does Scripture teach us about creation, human agency, tools, authorship, moral responsibility, attribution, Christian liberty, and stewardship—and what happens when we apply those principles consistently to AI?

I should also note that, like Mike, I have a personal stake in some of the legitimate grievances surrounding AI. My own published works were used by Anthropic in the training of its models. I am therefore not indifferent to the anger directed toward companies that have used copyrighted works without meaningful consent, transparency, or compensation. Those are real ethical and legal concerns, and acknowledging AI as a legitimate tool does not require minimizing them.

Nevertheless, the conclusions I have reached differ in important respects from Mike’s. Interestingly, I think my position is becoming increasingly countercultural. That alone does not make it correct. Christians can be contrarian and wrong. But it should make us pause. When, as Christians, we find ourselves adopting the same assumptions about humanity, creativity, and machines that increasingly dominate secular culture, we should at least ask whether these are actually biblical categories or whether we have unconsciously inherited the world’s categories and added Christian language to them.

One of the strongest arguments Mike discusses concerns the Imago Dei. Human beings are made in the image of God. I wholeheartedly agree. In fact, that doctrine is foundational to my own position. Human beings possess consciousness, rationality, intentionality, moral agency, imagination, responsibility, and the capacity for meaningful creative judgment in ways that our tools do not.

Where I differ is in what follows from that premise. Some Christians reason that human beings uniquely possess these characteristics, yet AI can now generate words, images, and music, and therefore allowing AI to generate those things threatens or replaces human creativity. I reach almost the opposite conclusion. Because AI possesses none of those characteristics, I do not recognize AI as an author in the first place.

AI has no desire to create. It has no intention to communicate. It has no aesthetic commitments. It has no beliefs. It has no moral responsibility. It does not wake up and decide to make a song. It does not decide to write a novel. It does not decide that the novel should be a detective story. It does not decide why that detective story should exist. A human being initiates that causal chain. Remove the intentional human agent and the work never exists. There is no literary or creative Big Bang that spontaneously happens apart from the intelligent designer.

This is where I believe much of the AI debate goes wrong. We have begun treating generation and authorship as though they were synonyms. They are not. A crayon generates the physical line on a piece of paper, but when a child draws a picture, we do not say, “The crayon created that.” We say, “The child drew that.” A guitar generates sound, but if someone lazily strums three chords, we do not say the guitar created the music. A camera mechanically produces an image through processes the photographer does not manually execute and, in many cases, does not fully understand, yet we attribute the photograph to the photographer.

Why? Because intuitively we understand the distinction between instrumental causation and intentional agency. The instrument produces something. The person intends something. That distinction matters enormously. AI can generate words. AI can generate pixels. AI can generate musical notes. But generation alone does not establish authorship. The theological question is: Where does the intentional agency reside? My answer is: in the human being.

Mike also draws an important distinction between what he calls “Assistive AI” and “Generative AI.” I think that distinction is useful insofar as it describes what different AI systems actually do. A spellchecker, research assistant, or editing tool is functionally different from a system generating prose, images, or music. Where Mike and I differ is in treating generation itself as the point at which the tool crosses some meaningful boundary into authorship.

That conclusion requires an argument.

A camera generates an image. A synthesizer generates sound. A drum machine generates percussion. Autocomplete generates words. Procedural software can generate landscapes, textures, characters, and entire virtual environments. We do not ordinarily conclude that these technologies become authors simply because they generate some portion of the resulting artifact.

So my question is not whether Generative AI does something substantially different from Assistive AI. Clearly it does. My question is why that functional difference should establish a new category of creative agency.

This brings me back to the distinction I have been making throughout this essay: generation is not the same thing as intentionality. An AI system can generate a paragraph just as a synthesizer can generate a sound. But neither instrument wanted the paragraph or the sound to exist. Neither conceived its purpose. Neither decided to communicate something through it. Neither possesses an aesthetic commitment to the result. Neither assumes responsibility for what happens next.

The human does.

This is why I am comfortable distinguishing between Assistive AI and Generative AI technologically while remaining unconvinced that the distinction establishes an authorship boundary theologically. If we are going to say that a tool ceases to be merely a tool once it generates expressive material, we need to explain why. We cannot simply build the conclusion into the word “generative.”

For me, the relevant boundary remains where it has always been: Where does intentional agency reside?

Until an AI possesses genuine intention, purpose, moral responsibility, and the capacity to independently choose to create, I see no reason to relocate authorship from the human being to the instrument through which that human acts.

This becomes especially interesting when someone raises the objection of the “lazy prompter.” Suppose someone types, “Write me a detective novel.” And let’s say eventually the AI can produce 70,000 words. The person publishes it. Surely that person is not an author?

Maybe the book is terrible. Maybe the person’s craftsmanship is nonexistent. Maybe I would not pay fifty cents for it. But none of those observations actually answer the philosophical question. We have quietly introduced another standard: How much effort must someone expend before something becomes art?

Where does that standard come from? Is a child’s drawing not art because it took thirty seconds? If someone lazily strums three chords, does the resulting sound cease to be music? If a photographer captures an extraordinary image with one press of a button, must we calculate how difficult pressing the shutter was before deciding whether photography occurred?

Effort may tell us something about craftsmanship. It may tell us something about excellence. It may affect artistic value. It may affect what someone is willing to pay. But those are different questions from whether something is art and who bears responsibility for creating it.

The person who typed “Write me a detective novel” made several determinations the AI never made. There should be a work. It should be written. It should be fiction. It should be a novel. It should be a detective novel. And ultimately, this result should or should not be published. The AI spontaneously determined none of those things.

That does not mean every AI-assisted work represents equal craftsmanship. Of course it does not. But poor craftsmanship does not magically transfer agency from a person to a machine.

A critic may respond that the AI still determined all the details. Yes, computationally. Not intentionally. And that distinction is important. An AI system may generate a detective named Sarah rather than John, but it does not care whether the detective is Sarah or John. It has no aesthetic preference between them. It has no purpose it hopes Sarah will accomplish. It has no emotional investment in Sarah’s journey. It has no message it wants Sarah to communicate. It has no reason for wanting the story to end one way rather than another.

The human can look at Sarah and say, “No.” And that one word reveals something extraordinarily important. The human can reject, revise, regenerate, edit, accept, abandon, or publish. The AI cannot decide that its own creation should even enter the world.

So, when someone tells me that AI has “written my book” or “created my music,” I think they are inadvertently granting the machine far more agency than I do. Ironically, some of the strongest critics of AI simultaneously call it a “soulless machine” and then proceed to describe it as though it possesses creative agency. I agree with the first description, but that is precisely why I reject the second.

In studying this issue, Romans 1 has significantly influenced my thinking. Now, I am not claiming that Romans 1 is about AI. It obviously is not. But there is a theological principle within the passage that I believe matters enormously: proper attribution.

Paul describes humanity’s catastrophic reversal of proper attribution, giving to created things what belongs to their Creator. At the highest possible level, Scripture insists that we correctly distinguish the Creator from His creation. That has caused me to become extremely sensitive to the language we use about AI and to the awareness of how limiting English is, in particular, to the task.

God is the ultimate Creator. Human beings, made in His image, are genuine secondary creators. Our tools are instruments through which human beings act. So, I am reluctant to attribute to an instrument what properly belongs to the intentional agent employing it.

Hebrews 3:4 expresses a related principle: “For every house is built by some man; but he that built all things is God.” The building points toward a builder. The hammer does not receive authorship.

For me, this is not simply semantics. Language eventually shapes philosophy. When we continually say, “AI decided,” “AI thinks,” “AI wants,” “AI created,” or “AI wrote,” we gradually begin assigning person-like categories to something that possesses no personhood. I believe Christians should be especially cautious about that.

What is particularly striking to me is that Christians readily appeal to intelligent design—the principle that evidence of order, complexity, and apparent design points beyond the artifact to an intelligent agent—yet we seem willing to abandon that same reasoning when discussing AI. We encounter an extraordinarily sophisticated artifact, observe what it can produce, and begin attributing intelligence, intention, and creative agency to the artifact itself rather than to the human intelligence responsible for designing, training, directing, and employing it. In doing so, I fear we have too readily surrendered to the world’s categorization of what constitutes “intelligence” rather than carefully examining that category through a biblical worldview.

This becomes particularly important when Christians discuss AI-generated Christian art. I have frequently encountered the objection, “If a soulless machine writes your Christian music, where is the Holy Spirit?”

My answer is straightforward: in the Christian.

I do not expect the Holy Spirit to indwell my guitar. I do not expect Him to indwell my microphone. I do not expect Him to indwell my keyboard. I do not expect Him to indwell Microsoft Word. And I do not expect Him to indwell an AI model.

The relevant question is whether I, as the moral and spiritual agent using those things, am walking according to the Spirit. Did I exercise discernment? Did I test what was produced against Scripture? Did I reject error? Did I correct bad theology? Did I faithfully steward the message? Did I take responsibility for what I ultimately placed before others?

If AI generates heresy and I knowingly publish it, AI will not stand before Christ and answer for that decision.

I will.

That is why I believe Christian stewardship actually reinforces human agency rather than diminishes it.

This also informs how I think about claims that AI is inherently spiritually dangerous or “of the devil.” Christians should be careful here. Created things can certainly be involved in sinful acts. Food can. Money can. Weapons can. Printing presses can. Musical instruments can. Computers can. AI can.

Humanity’s fall itself involved eating something God had forbidden. But the fruit was not a moral agent. Adam and Eve were. A gun can murder or defend. Money can be used to feed the hungry or finance evil. A printing press can be used to reproduce Scripture or pornography. AI can be used to educate, deceive, manipulate, create, plagiarize, encourage, exploit, or communicate truth.

The moral question concerns what moral agents do with created things.

Acts 10 has also influenced my thinking here. Again, I am not claiming Peter’s vision was about technology. Peter himself explains its immediate application to the inclusion of Gentiles. But the correction God gives Peter contains a principle I take seriously: “What God hath cleansed, that call not thou common.”

I want to be extremely cautious about assigning spiritual categories to things that God has not assigned to them. Calling something inherently spiritually corrupt requires more than my discomfort with it. When we begin attributing spiritual agency or inherent spiritual power to created things without biblical warrant, we move dangerously close to the very kind of superstitious thinking that Scripture warns against—assigning to the creation powers and properties that properly belong to spiritual persons or ultimately to the Creator. In that sense, we risk adopting a worldview closer to witchcraft than biblical Christianity: treating an object or instrument as though spiritual agency resides within the thing itself rather than in the moral and spiritual agents who employ it.

Romans 14 therefore becomes extraordinarily important. A Christian may prayerfully conclude, “I cannot use generative AI with a clear conscience.” I respect that. Do not use it. Another Christian may prayerfully conclude, “I understand AI as an instrument and can use it faithfully while remaining responsible before God for everything I produce.” I believe Christian liberty requires room for that Christian too.

Paul asks, “Who art thou that judgest another man’s servant? to his own master he standeth or falleth.” That does not mean everything is permissible. It does not eliminate discernment. It does not prohibit Christians from correcting actual sin. But before we elevate something into a universal Christian prohibition, we need to actually establish that prohibition rather than simply assume it.

My conscience cannot become God’s commandment for someone else. And the stronger brother is not permitted to despise the weaker any more than the weaker is permitted to condemn the stronger. Both belong to Christ.

Christian liberty, however, does not require belligerence. There is another passage that shapes my thinking: Matthew 17:24–27. Jesus establishes that, as the Son, He is free from the temple tax. “Then are the children free.” But He immediately says, “Notwithstanding, lest we should offend them,” and He pays it.

That gives me an important principle for navigating AI disclosure requirements. I may disagree philosophically or ontologically with the way a publisher, distributor, retailer, platform, or government defines “AI-generated.” I may believe their terminology incorrectly attributes creation to the tool rather than the human agent.

Nevertheless, if they ask, “Does this work contain AI-generated prose?” I will answer truthfully. If they ask, “Does this recording contain AI-generated vocals?” I will answer truthfully. If they ask about images, lyrics, music, or other content, I will disclose it according to their requirements.

Why? “Notwithstanding, lest we should offend them.”

My compliance does not require me to surrender my theological understanding of authorship. Christian liberty does not require me to exercise every liberty available to me. Sometimes wisdom, transparency, love, and respect for legitimate earthly authority call me to voluntarily comply even when I disagree with the categories underlying the requirement.

So, I suspect Mike Duran and I agree about far more than we disagree about. We agree that humanity matters. We agree that the Imago Dei matters. We agree that creativity matters. We agree that Christians should exercise stewardship. We agree that AI can be abused. We agree that exploiting the Gospel for profit is wrong. We agree that Christians should be truthful. We agree that AI must never relieve human beings of moral responsibility. We agree that Christians should be charitable toward those who reach different conclusions.

Our fundamental disagreement appears to concern where authorship resides.

Mike’s position appears to locate authorship substantially in the human production of expressive material. His distinction between Assistive AI and Generative AI follows naturally from that understanding: once the machine begins generating the expressive material itself, something uniquely human has been surrendered. I think this is actually a commonly held view.

My position begins one level deeper. I locate authorship in intentional human agency. Generation is something an instrument can perform. Authorship requires an agent. Therefore, the fact that an instrument has become vastly more capable of generating expressive material does not, by itself, transform that instrument into an author. And that distinction changes almost everything downstream.

This is ultimately where I think Christians should pause. The dominant culture increasingly anthropomorphizes AI. AI “thinks.” AI “knows.” AI “hallucinates.” AI “decides.” AI “creates.” AI “writes.” AI becomes increasingly described as though it were another kind of person.

And ironically, Christians who strongly oppose AI sometimes adopt precisely the same ontology while condemning it. They subconsciously concede that the machine creates. They simply believe humans should not let it.

My position is considerably more skeptical of AI than that.

I do not believe the machine is an author at all. I do not believe it possesses creative intention or consciousness. I do not believe it possesses moral agency. I do not believe it possesses spiritual agency. I do not believe it possesses a soul. I do not believe it bears God’s image.

It is an extraordinarily sophisticated instrument created and operated by human beings.

That is it.

Perhaps, then, the genuinely countercultural Christian response is not simply, “AI is dangerous because machines can now create like humans.”

Perhaps we should first question the premise.

Can they?

Can something without consciousness, intention, moral agency, responsibility, desire, or purpose actually be an author? Or has our culture become so impressed by the artifact that we have begun attributing to the instrument what properly belongs to the human agent behind it?

I do not claim that every Christian must reach my conclusion. But I do think Christians should do the theological work before adopting the culture’s categories.

Because our task is not simply to decide whether we are “for AI” or “against AI.”

Our task is to think Biblically.

And before we declare another servant’s use of a new technology sinful, counterfeit, soulless, or incompatible with human creativity, perhaps we should ask ourselves a more difficult question:

Is my judgment actually arising from Scripture, or have I accepted the world’s definition of what AI is, what creativity is, and what authorship is, and then assumed that definition must be Christian?