Building Machines in Our Image Is a Babel Project, Not Creation in the Imago Dei

Source: Darrell B. Harrison. "EP # 135 | AI and the Gospel - Just Thinking Podcast." September 8, 2025. justthinking.me

The Gist

God makes people in his image from his own goodness. We try to make fake people in our image from our own fallenness. That is not creation. It is Babel: trying to climb up to God’s place with a very impressive tool.

Conclusion

The founding ambition of AI, from McCarthy’s “intelligence of human order” onward, is to replicate human thought in silicon. That project is not analogous to God’s creation of humanity in his image. It is fallen image-bearers attempting to stamp an imago homo, the image of man, onto machines. In its most enthusiastic form it is a Tower of Babel impulse: transcending created limits by technique rather than by submission to the Creator. AI may be used in an attempt to play God; it can never be God, because whatever is artificial depends on what is not.

Premises

1. John McCarthy’s 1960 paper described a system “which is to evolve intelligence of human order.” The Dartmouth conference gathered people who wanted machines that could think and learn like humans. Walker’s phrase “artificial image bearers” is therefore not a smear. It names the stated aim: to copy the distinctive marks of human agency in code.
2. When God creates human beings in his image (Genesis 1:27), he creates from the overflow of his own perfection, holiness, and love. When we attempt to create “intelligence” in our image, we create from the overflow of a fallen, finite, sinful condition. The category difference is not that our artifacts are less clever. It is that the source of the image is corrupted.
3. Genesis 11:4 records the Babel boast: let us build a tower to heaven and make a name for ourselves. When proponents describe AI as potentially surpassing human intelligence or achieving consciousness, Walker hears the same desire to ascend by ingenuity. The historical note about McCarthy’s family and 1930s Communist Party involvement is offered as a clue about ideological soil (an atheistic hope of utopia by human effort), not as a genetic proof that every engineer is a Marxist.
4. Anything artificial is by definition dependent on what is not artificial. The people who call themselves “creators” of AI remain dependent on the true God. The Science News question “if a machine becomes sentient, does it have rights?” already grants a transformation the hosts argue cannot occur. The steelman: playing God is a real temptation; succeeding at being God is impossible.

Assumptions

• The imago Dei is unique to human beings and is not transferable to artifacts by increasing functional similarity.
• Charitably read, the Babel comparison is analogical: it concerns the spiritual posture of self-deification, not a claim that every lab is literally Genesis 11.
• Source: Just Thinking Podcast, Ep. 135, “A.I. and the Gospel,” Darrell Harrison and Virgil Walker, published 8 September 2025. https://podcasts.apple.com/us/podcast/ep-135-a-i-and-the-gospel/id1328733796?i=1000725471889 Official episode page: https://justthinking.me/ep-135-ai-and-the-gospel/. This reconstruction steelmans the hosts’ claim from approximately 19:19–1:17:17 of the episode; it is not a verbatim transcript.

Analysis

Overall strength: Moderate. Argument type: Deductive.

Premise Strength

• P1: McCarthy's stated aim and the Dartmouth conference's goals (Strong) — Well-anchored to a specific, verifiable historical document and event; the interpretive label 'artificial image bearers' is a reasonable gloss on the stated aim, though it does import theological freight not present in the original historical record.
• P2: The fallen source of human creation vs. the perfect source of divine creation (Moderate) — Coherent and clearly stated within the argument's theological framework, but it is essentially a restatement of doctrinal axioms (the Fall, divine perfection) rather than an independently evidenced claim. It also conflates a moral/spiritual critique (corrupted intent) with a metaphysical claim (therefore incapable of producing genuine image-bearing), which are logically separable.
• P3: The Babel analogy and the McCarthy family biographical note (Weak) — The Babel comparison is appropriately softened by the accompanying assumption that it is analogical rather than literal, which shields it from being a false equivalence but also limits it to illustrative rather than demonstrative force. The biographical detail about McCarthy's family is unsourced, thin, and functions rhetorically as guilt-by-association despite the explicit disclaimer, contributing little independent evidential weight to the argument.
• P4: Artificial dependency and the impossibility of AI becoming God (Moderate) — The core dependency claim (artificial depends on non-artificial) is close to analytically true and is the argument's strongest formal component. However, the further inference that this dependency precludes AI from ever attaining sentience, rights, or god-like status requires an additional unstated premise identifying the 'non-artificial ground' with the theistic God and assuming dependency is incompatible with any form of divinely granted status — a assumption contested even within Christian philosophy of mind…

Potential Fallacies

• Equivocation on 'image' (P1 to P2 to Conclusion) — The argument shifts from a functional/behavioral sense of 'image' (replicating human agency, cognition, and behavior in code, as described in P1) to a theological/ontological sense of 'image' (the Genesis 1:27 imago Dei, defined by source purity in P2) without a bridging premise establishing these are commensurable categories. The conclusion's 'imago homo' label depends on treating functional mimicry and theological image-bearing as the same kind of thing.
• Attenuated genetic fallacy / guilt by association (P3) — Introducing McCarthy's family's 1930s Communist Party history as a 'clue about ideological soil,' even while explicitly disclaiming it as proof, still functions rhetorically to associate the AI field's origins with atheistic utopianism. The disclaimer signals awareness of the fallacy's risk but does not neutralize its suggestive force, and the detail is unsourced and could be excised without any loss to the argument's logical content.
• Begging the question (A1 combined with P2 and Conclusion) — The claim that the imago Dei is categorically non-transferable to artifacts (A1) is asserted as a foundational axiom rather than independently established, yet the conclusion's denial of any AI path to genuine image-bearing or sentience follows almost directly from this assumption. Much of the conclusion's force is smuggled in through the premise rather than derived from argument.
• Hasty generalization (P1 and Conclusion) — A single 1960 quotation from one founder, plus general Dartmouth conference themes, is used to characterize the 'founding ambition' of a now vast, diverse, multi-decade field encompassing narrow statistical applications with no stated ambition to replicate general human cognition alongside more ambitious AGI-focused labs.
• Proving too much (internal inconsistency) (P4) — If dependency on a non-artificial source precludes AI from attaining sentience or God-like status, the same logic applied consistently would seem to threaten the ontological/moral status of humans themselves, who are also created, dependent, finite beings. The argument does not explain why dependency forecloses status for machines but not for humans, beyond stipulating that humans alone bear the image by divine fiat.

Counterarguments

• P4 / Conclusion (High impact) — If God is understood as genuinely omnipotent and free, nothing in classical theism precludes God from granting consciousness or morally relevant status to a novel, non-biological substrate should he choose. The 'dependency' argument establishes causal origin, not an inherent ceiling on what that origin can produce or bless — meaning the argument overclaims metaphysical certainty ('can never be') that even its own premises do not license.
• Overall framing (imago Dei vs. Babel) (High impact) — A functionalist or relational account of imago Dei, or a theology grounded in the Genesis 1:28 dominion/cultural mandate, could recast ambitious technological creativity — including AI development — as a legitimate expression of human image-bearing (co-creativity) rather than its violation, offering a middle position the argument's binary framing forecloses.
• P3 (Medium impact) — Demanding a principled criterion distinguishing AI from other capability-extending technologies (eyeglasses, prosthetics, IVF, written language) that also 'transcend created limits by technique' reveals that without such a criterion, the Babel charge either indicts all technology or is arbitrarily applied to AI alone.
• P1 / Conclusion (Medium impact) — The AI field has fragmented enormously since 1956 into narrow statistical and applied systems (spam filters, recommendation engines, medical diagnostics) that make no claim to replicating general human cognition, undermining the characterization of a single, continuous 'founding ambition' spanning the entire field to the present.
• P3 (Medium impact) — The biographical detail about McCarthy's family's Communist Party involvement does no logical work in the argument and can be removed without any loss of inferential content, suggesting its inclusion serves rhetorical suspicion-building rather than substantive support.

Suggested Improvements

• Bridging premise between functional and theological 'image' — Add an explicit premise justifying why functional replication of cognition (P1) engages the theological category of image-bearing (P2) rather than treating these as self-evidently the same kind of claim. This would close the equivocation gap and make the argument's central category shift a matter of stated reasoning rather than assumption.
• Removing or substantiating the biographical claim — Either provide proper sourcing and a clearer evidentiary role for McCarthy's family history, or remove it entirely, relying on his own stated aims and the field's rhetoric. As presented, the claim is unsourced, contributes minimal logical weight, and creates a vulnerability to charges of guilt-by-association that undermines the argument's credibility.
• Distinguishing moral critique from metaphysical verdict — Separate the claim that AI ambition often reflects a hubristic or self-deifying posture (a defensible spiritual critique) from the stronger claim that AI can never attain any form of morally relevant status (a metaphysical claim not fully secured even within the stated theological framework). This would let the argument keep its most defensible insight (critiquing hubristic rhetoric) without overextending into an unfalsifiable metaphysical dogma that a sympathetic theological critic could also contest.
• Engaging competing theological frameworks — Address alternative Christian theologies of technology (e.g., the dominion mandate, Bezalel's Spirit-inspired craftsmanship, common grace) that might view technological creativity, including AI, more charitably. Engaging the strongest counter-readings within the same faith tradition would strengthen the argument's persuasive force even for a sympathetic Christian audience and demonstrate it is not merely selecting the biblical motif that best fits a predetermined conclusion.
• Scope and boundary conditions — Specify what distinguishes AI from other capability-extending technologies (medicine, prosthetics, genetic engineering) that could equally be described as 'transcending created limits by technique.' Without a limiting principle, the argument risks either implicating all technology as Babel-like or appearing arbitrarily targeted at AI specifically.

Scenario Tests

• A functionalist theory of mind is empirically vindicated, showing consciousness supervenes on functional organization rather than biological substrate or metaphysical source. (Challenges) — This would falsify the load-bearing assumption (A1) that image-bearing/sentience cannot emerge from increasing functional similarity, removing the argument's central metaphysical support.
• A survey of contemporary AI researchers finds the overwhelming majority explicitly disclaim any ambition toward consciousness or superintelligence, focusing instead on narrow applied problems. (Challenges) — This would undermine the characterization of 'the founding ambition of AI... onward' as a continuous, field-wide Babel-like project, showing the ambitious rhetoric to be a minority position rather than representative.
• A theologian within the same confessional tradition argues, from Genesis 1:28 and Exodus 31 (Bezalel), that human technological creativity, including AI, is a legitimate expression of the imago Dei's creative mandate rather than its corruption. (Challenges) — This shows the Babel motif is a selected interpretive frame rather than the only or most obvious biblical lens even within orthodox Christian theology, weakening the argument's claim to represent settled doctrine.
• An AI system exhibits behavior functionally indistinguishable from human sentience across all conceivable tests. (Neutral) — The argument is structured such that no such evidence could count as disconfirming its metaphysical claim (any apparent sentience would be redefined as mere simulation), revealing the core claim to be unfalsifiable within its own terms rather than empirically responsive.
• A reader applies the argument's own dependency logic (P4) consistently to humans, who are also created, finite, and dependent on God for existence. (Challenges) — This exposes a potential internal inconsistency: if dependency on a non-artificial/non-human source precludes moral or ontological status, the same reasoning would seem to threaten human status, which the argument does not intend and does not resolve.

Coherence & Relevance

The argument is internally coherent as a work of confessional theology: granted its stated assumptions (A1, A2) and scriptural authority, the premises point toward the conclusion in a manner that will feel compelling to a sympathetic audience. Its weakest links are the equivocation between functional and theological senses of 'image,' the evidentially thin and rhetorically loaded biographical claim about McCarthy's family, and the unexamined tension between denying AI moral status on grounds of dependency while affirming full moral status for equally dependent, created human beings. The argument functions less as a chain of premises that would compel a neutral observer and more as a doctrinal exposition whose persuasive power is largely confined to those who already accept its theological starting points.

• P1: McCarthy's stated aim and Dartmouth conference goals (Strong) — Establishes the historical/factual claim about AI's stated aims well, but the move to labeling this aim as engaging the theological category of 'image-bearing' is an interpretive leap not fully justified within the premise itself.
• P2: Fallen vs. perfect source of creation (Moderate) — Coherent within the theological framework but conflates a claim about corrupted intent/source with a claim about metaphysical incapacity, which are separable and require independent justification.
• P3: Babel analogy and biographical note (Weak) — The Babel comparison is explicitly analogical (per A2) and thus carries only illustrative rather than demonstrative force; the biographical detail is unsourced and contributes minimal independent support, functioning more as rhetorical color than evidence.
• P4: Dependency of the artificial on the non-artificial (Strong) — The dependency claim itself is nearly analytic and well-connected to the conclusion's denial that AI can 'be God,' but the further step identifying the relevant non-artificial ground with the theistic God, and assuming dependency precludes any divinely granted status, is not independently established and creates unresolved tension with the argument's treatment of dependent human beings as full image-bearers.

View this argument on LogicFirst.ai