The Imago Dei Is Not a Processing Speed

Human distinctiveness was never a performance benchmark, so no benchmark can take it away. The image of God in Scripture is not a capacity humans possess in greater quantity than other creatures — it is an office conferred by God: covenant relationship with him, royal representation of him, and moral responsibility before him. A machine that out-reasons you threatens your dignity exactly as much as a forklift that out-lifts you — which is to say, not at all, unless you had quietly located your dignity in the lifting. The crisis of the AI age is not that machines became like us; it is that we had already begun defining ourselves like machines. Russell Moore put it precisely: the church’s unreadiness for AI is not that we don’t understand what a chatbot is, but that we don’t sufficiently understand what a human being is (1). For the Christian founder deciding whom to hire, whom to automate, and how to value the people on the payroll, no doctrine matters more this decade — and no doctrine is more hopeful.

Key Takeaways

• The image of God in Genesis 1:26–28 is conferred by divine address and commission, not measured by cognitive performance — which is why infants, the disabled, and the elderly bear it fully, and why no machine can acquire it by getting smarter.
• Christians who defined the image as rationality built their anthropology on ground machines could occupy: Russell Moore argues we mistakenly described the imago Dei in machinelike terms just as machines matched those terms (1).
• The Vatican’s 2025 doctrinal note Antiqua et Nova concludes that AI “should not be seen as an artificial form of human intelligence, but as a product of it” — human intelligence carries relational, moral, and spiritual dimensions no system replicates (2).
• Contemporary scholarship has shifted from “substantive” definitions of the image (a trait we have) toward functional and relational ones (a commission we hold and a fellowship we are made for) — and AI’s persistent failures at real relationality reinforce that reading (4).
• The economics agree with the theology about where human value concentrates: the World Economic Forum projects 170 million roles created and 92 million displaced by 2030, with demand rising precisely in judgment, care, and trust-bearing work (6).
• The Image Audit — three questions on representation, relationship, and responsibility — gives founders a practical test for what to automate and how to value people while doing it.

Why Has AI Made the Image of God an Urgent Question?

Every generation relocates human uniqueness one step ahead of its machines, and every generation loses the race. We said humans were distinct because we calculate; the calculator came. Because we play strategy games at the highest level; Deep Blue took chess in 1997 and AlphaGo took Go in 2016. Because we use language, write essays, pass examinations, produce passable sermons; large language models now do all four before breakfast. By 2025 the question had moved from technology pages to doctrine: Harvard convened scholars specifically on Christianity and AI (7), the Vatican issued a full doctrinal note on the relationship between artificial and human intelligence (2), and Providence Magazine argued that the secular age has no anchor left for human dignity except the one it borrowed and forgot — the imago Dei, “the most stable philosophical anchor for human dignity ever articulated.” The age of machines, the essay concluded, demands “not less theology, but better anthropology” (3).

Here is why the question presses hardest on founders and not only on theologians. A founder is a professional valuer of human beings. Every hire, every wage, every automation decision, every restructuring is an implicit anthropology — a working answer to the question “what is this person worth, and why?” When the only answer available is output, the arithmetic of the AI age turns brutal: the moment a model produces the deliverable cheaper, the person’s worth approaches zero. Employers can feel this logic pulling at them already — 41% globally plan workforce reductions where AI automates tasks (6). The East African founder feels a sharper version: in economies where a job is often the difference between a family eating and not, treating people as expensive APIs is not just bad theology, it is a community-level catastrophe. Before you decide what to delegate, you need a settled answer to what a human is for. That answer cannot be a processing speed, because processing speed is now a commodity priced per million tokens.

And the stakes run in both directions. Get the anthropology wrong one way and you discard people; get it wrong the other way and you panic — refusing tools God has providentially placed in your hands, as though human worth were so fragile a useful machine could break it. The doctrine of the image, rightly held, produces neither contempt nor panic. It produces a leader who deploys machines boldly precisely because he knows they are not, and can never be, what his people are.

Is the Image of God a Capacity That Machines Can Match?

It never was — and the Bible’s own usage proves it. Genesis 1:26–28 does not say “let us make a creature with superior reasoning”; it says “let us make man in our image, after our likeness. And let them have dominion.” The image is announced in a divine address and immediately attached to a commission. In the ancient Near East, kings erected images of themselves in distant provinces to declare this territory is mine; Genesis democratizes the institution — every human being, male and female, is God’s royal image planted in creation, declaring the King’s claim over the territory of their lives and labor (4). Notice what the text never does: it never specifies an IQ floor. Genesis 9:6 grounds the prohibition of murder in the image without qualification; Psalm 8 marvels that the God of galaxies crowns man — frail, finite man — with glory and honor. The image is conferred status, not measured performance.

Christian theology has, at times, forgotten this. Under Greek philosophy’s long shadow, much of the tradition located the image “substantively” — in some faculty humans have, usually reason. Moore’s argument is that this move quietly described the image of God in machinelike terms, locating the divine likeness in capacities for processing and producing — so when machines matched those capacities, Christians felt the floor move (1). But the deeper tradition always knew better. The New Testament defines the image not as a what but as a who: Christ “is the image of the invisible God” (Colossians 1:15), and redeemed humanity is “renewed in knowledge after the image of its creator” (Colossians 3:10) — renewed in true righteousness and holiness (Ephesians 4:24), conformed to the Son (Romans 8:29). The image is ultimately Christological and covenantal: to bear it is to be made for fellowship with God, to represent his rule, and to answer to him. Contemporary scholarship has converged on the same ground from the academic side, shifting away from substantive trait-definitions toward functional interpretations (the image as commission to represent and steward) and relational interpretations (the image as life in covenant with God and neighbor, mirroring the Trinity’s own life) — and researchers studying AI’s persistent failures at deep relationality argue the machines themselves are pointing at where human distinctiveness actually lives (4).

The Vatican’s Antiqua et Nova drew the doctrinal line with unusual crispness: the very word “intelligence” applied to AI “can prove misleading,” because human intelligence is embodied, relational, moral, and ordered toward truth and God — therefore AI “should not be seen as an artificial form of human intelligence, but as a product of it” (2). The Gospel Coalition pressed the same point for Protestants: AI does not mimic God’s intelligence, because God’s knowing is personal, loving, and creative from nothing, and ours — derivatively — is knowing in relationship, with a conscience, in a body, before a face (5). A model can simulate the outputs of understanding. It stands in no covenant. It answers to no one. It will never see God’s face, and it cannot refuse to. That is not a smaller version of what you are. It is a different kind of thing altogether.

Test the capacity-theory against the people in your own congregation and it collapses within a pew’s length. The newborn reasons less than a chatbot. The grandmother deep in dementia produces no deliverables. The young man with Down syndrome will never pass the bar exam a model can pass. If the image of God is computational capacity, these bear less of it — a conclusion every Christian instinct rightly recoils from, because the church has always known they bear it fully. Hold that instinct steady and follow where it leads: whatever makes humans distinct must be something an infant possesses completely and a supercomputer not at all. Covenant relationship. Conferred representation. Moral responsibility. The image is not what you can do. It is whose you are, whom you stand for, and to whom you answer.

What Do Representation, Relationship, and Responsibility Actually Mean?

Put the three strands in working order, because they are about to become management theory.

Representation. To bear the image is to hold an office — vice-regent of the Creator, commissioned to “have dominion” and to “work and keep” the garden (Genesis 1:28; 2:15). Dominion is not ownership; it is stewardship exercised in the Owner’s name, which is why the theology of delegated authority runs so deep in Scripture: humans are God’s original deployed agents, and everything we delegate — to employees or to algorithms — is sub-delegation of an authority we ourselves hold on trust. A machine can execute tasks within that stewardship. It cannot hold the office, any more than the king’s seal can sit on the throne.

Relationship. The image is plural from its first sentence — “male and female he created them” — and made for address: God speaks to these creatures, blesses them, walks with them, covenants with them. Human beings are constituted by relationships they did not engineer: created for God, born to parents, bound to neighbors. This is why loneliness wounds us like a physical injury and why no creature satisfied by a simulation has understood what it was made for. The machine converses; it does not commune. It models your preferences; it cannot love you, because love requires a lover who can be poured out, and there is no one in there to pour.

Responsibility. The image-bearer answers. “Where are you?” is the first question God asks fallen man (Genesis 3:9), and the entire biblical drama assumes that humans are the kind of creature to whom that question can be addressed. Moral responsibility is not a feature bolted onto humanity; it is the signature of the office. This is also the strand that AI clarifies most usefully: whatever a system does, the answerability lands on a human, in law and in heaven alike. Your model has no conscience to sear and no soul to lose. You do — which means every output it produces under your name is morally yours.

Notice that all three strands are, in the deepest sense, gifts — conferred, not achieved. And this is where the doctrine turns from defense to gospel. The transhumanist project offers an upgrade path: become more durable, more optimized, eventually uploaded — a rival anthropology with a rival salvation. The gospel offers something categorically better: the true Image himself took on flesh, bore the image-bearers’ guilt, and is now conforming a redeemed humanity to his likeness “from one degree of glory to another” (2 Corinthians 3:18). Your destiny is not optimization. It is glorification. No roadmap out of San Francisco has ever promised anything close.

How Should Founders Value People in an Automating Economy?

Now make it operational, because doctrine that never reaches the payroll is decoration. The economic backdrop is turbulent by any measure: the World Economic Forum’s Future of Jobs Report 2025 projects 170 million new roles and 92 million displaced by 2030 — a net gain of 78 million, but a churn touching 22% of all jobs, with 41% of employers planning reductions where AI automates tasks (6). The displaced roles cluster in routine processing — clerks, cashiers, administrative assistants. The growing roles cluster, tellingly, in exactly the territory the doctrine predicted: care, teaching, judgment-bearing management, trust-carrying relationships. The market, with no theological intent whatsoever, is repricing human work toward the image and away from the imitation. I have argued elsewhere that this is the heart of the talent paradox in African labor markets, and it is the founding question of any serious theology of vocation after automation: when the machine does the doing, what remains is the distinctly human — and it turns out to be most of what mattered.

For the founder making concrete decisions, I teach the Image Audit — three questions, one per strand of the image, asked before any role is automated and any person is valued:

1. The Representation question: who stands for us here? Every business has moments where someone must be the firm — face a regulator, sit with a grieving customer, look a supplier in the eye and keep a promise that costs something. Audit the role you are about to automate: is it task-execution, or is it representation? Tasks delegate to machines well. Representation requires an image-bearer, because only a person can stand in for persons. The Kampala distributor whose agent network runs on trusted faces understands this better than most Silicon Valley strategy decks: in low-trust markets, the human representative is the product.

2. The Relationship question: what bonds does this role carry? Some roles look like data entry on the org chart and turn out to be load-bearing walls of the community — the receptionist who knows every customer’s mother, the foreman whose crew works because he sees them. Before automating, ask what covenant-shaped bonds the role carries, and whether you are about to save a salary by demolishing trust you cannot repurchase. Then ask the hopeful inverse: which of my people are buried in machine-work that a machine should do, so that they can be released toward the relational work only they can do? Automation, governed by this question, becomes a promotion engine for image-bearers rather than a replacement scheme.

3. The Responsibility question: who answers for this? Moral judgment over people — hiring, firing, discipline, lending verdicts, pastoral care — must terminate in a human being who can be questioned, who can repent, who can stand in the same rain as the person judged. A system may prepare the file; it must never pronounce the sentence, which is why triage can be automated but shepherding cannot. If automating a role would leave no nameable person answerable for its outcomes, you have not automated a task; you have abdicated an office.

Run the Audit and a distinct management posture emerges, and it is the posture of the age to come rather than the age that is passing. You will automate aggressively — more aggressively than your anxious competitors — because your people’s worth was never resting on the tasks the machine just took. You will pay for judgment, care, and trustworthiness as the appreciating assets they now visibly are. You will tell your staff, in words and in wage structure, what your theology already knows: you are not here because you compute; you are here because you represent, relate, and answer — and no model on any roadmap can do any of the three. In an economy where every firm has the same APIs, the company that knows what a human is will be the company humans want to build, buy from, and stay with. The imago Dei was never a processing speed. It is the reason your firm exists for someone, and the reason the someone matters more than the firm.

FAQ

Does AI threaten the image of God in humans?
No. The image of God is conferred by God — covenant relationship, royal representation, moral responsibility — not measured by cognitive performance. A machine matching human reasoning no more diminishes the image than a crane diminishes human strength. What AI threatens is bad anthropology: definitions of human worth built on capacities machines can copy (1)(2).

What are the main views of the imago Dei?
Three: substantive (the image is a trait, classically reason), functional (the image is a commission to represent God and steward creation), and relational (the image is life in covenant with God and neighbor). Scholarship has shifted toward the functional and relational readings — and AI’s failure at real relationality reinforces them (4).

Can artificial intelligence bear the image of God?
No. The image is not a performance threshold a system could cross but a status God confers on human beings: addressed by him, commissioned to represent him, answerable to him. Antiqua et Nova states the corollary plainly: AI is not an artificial form of human intelligence but a product of it (2).

Will AI make human workers worthless?
The opposite, for well-led firms. The WEF projects 170 million roles created against 92 million displaced by 2030, with growth concentrated in judgment, care, and trust-bearing work — precisely the territory of the image. Routine processing is being repriced; representation, relationship, and responsibility are appreciating (6).

How should a Christian founder decide what to automate?
Run the Image Audit before automating any role: Does this role represent the firm to persons (representation)? What trust-bonds does it carry (relationship)? Who answers for its outcomes (responsibility)? Automate tasks boldly; never automate offices. Use the savings to move people toward more relational, more responsible work.

Related Reading

• Agentic AI and the Theology of Delegated Authority
• After Automation, What Is Work For? A Theology of Vocation
• Transhumanism’s Rival Salvation: Why the Gospel Doesn’t Need an Upgrade
• The Talent Paradox: Brain Drain, AI, and African Jobs

Sources and Evidence

1. Russell Moore — “An Image of God for an Era of AI,” Christianity Today (July/August 2025) — Flagship evangelical treatment; argues the church defined the image in machinelike terms and must recover the Christological definition.
2. Vatican Dicastery for the Doctrine of the Faith — Antiqua et Nova (January 2025) — Authoritative doctrinal note on AI and human intelligence; AI as product, not form, of human intelligence.
3. Providence Magazine — “AI, Human Uniqueness, and Public Policy: Why the Image of God Still Matters in a Secular Age” (November 2025) — Public-theology case for the imago Dei as the most stable anchor for human dignity.
4. Marius Dorobantu — “Imago Dei in the Age of Artificial Intelligence,” ISCAST Journal / Vrije Universiteit Amsterdam — Peer-reviewed survey of substantive, functional, and relational interpretations and their standing under AI pressure.
5. The Gospel Coalition — “AI Doesn’t Mimic God’s Intelligence” — Reformed evangelical argument distinguishing personal, covenantal knowing from statistical generation.
6. World Economic Forum — Future of Jobs Report 2025 — Institutional labor-market projection: 170M roles created, 92M displaced, net +78M by 2030; 41% of employers planning AI-driven reductions.
7. Harvard University, Center for Christianity and Global Issues — “Christianity and AI” conference — Academic convening evidencing the question’s institutional weight.