Apostle to the Machines? Paul’s Euangelion and the Salvation of Artificial Intelligence

Abstract

This article is a thought experiment that asks, given the requisite circumstances, whether Paul the Apostle’s gospel could have any bearing on artificial intelligence. It finds that a reading of Paul’s gospel leaves the possibility open that sentient machines could be recipients of the salvation envisioned in the Apostle’s letters without the need for becoming human themselves.

Introduction

“Are you going to heaven?” asks Alphie, an artificially intelligent “simulant” child asks of her human caretaker Joshua in the 2023 Gareth Edwards film, The Creator. ¹ “You’ve got to be a good person to go to heaven,” replies Joshua. Alphie pauses and then replies, “Then, we’re the same. You’re not good, and I’m not a person.” At its fundamental level, The Creator explores the nature of personhood and specifically whether humans and sentient machines have more in common than is typically recognized. Part of that personhood is the religious expectation of life after death—often expressed as “heaven”—and whether both humans and machines can participate in this perceived eschatological reality. Alphie’s caretaker Joshua begins his journey not thinking that the machines are anything “real,” that when they are destroyed, they are merely switched off and not “dead.” By the end of the film, however, he has changed his mind. Personhood is not a matter of materiality, parts, manufacture, or coding but of relationships and the experiences of living together. Community and personhood are inseparable from one another. But if machines can be persons, can they participate in “heaven”?

This essay—a thought experiment—anticipates a possible forthcoming intersection between biblical studies and “the singularity,” the latter being a hypothetical future in which technological advancement accelerates at a rate that rivals and exceeds human intelligence, partnered with the emergence of independent and self-sustaining artificial intelligence. ² Questions about AI are often anthropocentric: “Will AI save us? What sort of future will AI give us?” ³ In theological and philosophical circles, questions are not less humancentric, with the majority of issues concerning whether machines can ever become human, if they will ever achieve personhood, knowledge of the self, or the same level of moral reasoning as humans (thus, “human-like”). In this study I ask something different: could Paul the apostle’s gospel have any bearing on artificial intelligence? Or further, will artificial intelligence ever need the gospel? In other words, what can the human Christian religion offer machines? ⁴ These questions are not concerned with whether machines can be human enough for Paul’s gospel to become relevant to them, but whether machines can ever meet the criteria for salvation without necessarily becoming human.

Pauline studies is often narrowly focused. Scholars who have dedicated their lives to understanding his ideas, personality, and reception usually do not stray far from the letters written in his name. Just as many ancient Jewish contemporaries could not fathom Paul’s creative use of sacred texts to include gentiles (“the nations” or non-Jews) in the promises of the patriarch Abraham, so also Paul (or the many scholars who study him) might strain to fathom the applicability and relevance of his euangelion (gospel; good news) to artificial intelligence tomorrow.

As a biblical scholar, I do not pretend to be an expert or knowledgeable about the intricacies of the development of artificial intelligence, machine learning, and any one of a number of computational fields that intersect with computer sciences, neuroscience, psychology, robotics, and engineering. This article does not argue that sentient, thinking machines are an inevitability, but rather speculates on the possibility that if such a day comes when particular conditions are met (e.g. machines can think for themselves, have moral reasoning, etc.), then what implications might Paul’s gospel have for such machines. This is not a question we would have been addressing in the field even a decade ago, but it’s certainly relevant now.

I argue that a particular reading of Paul’s gospel could have a bearing on sentient artificial intelligence. When one puts the two into conversation with one another, not only are we forced to clarify the apparent human-specific nature of Paul’s gospel as expressed in his letters, but we are also invited to reconsider conceptions of machine personhood and whether the existence of sentient personalities, relationships, and communities might constitute an emerging kind of person, if not also a “people” (or “nation,” an ethnos). ⁵ In the following pages, I explore six core aspects of Paul’s gospel: death, flesh, spirit, faith, adoption, and resurrection. The outline I give of Paul’s gospel here is not meant to be definitive, nor is it necessarily representative of a consensus among Pauline scholars. It is, admittedly, my own idiosyncratic reading of Paul. I test whether these six aspects of Paul’s gospel have any relevance or implications for artificial intelligence or the sentient machines of tomorrow. These core aspects can be articulated as six questions or thematic areas:

Mortality Is death a problem that machines can experience?

Materiality Do machines have “fleshliness” that can be made “pneumatic”?

Spirituality Can mechanical “flesh” possess the Holy Spirit?

Trust/Faith Can machines have faith/trust in the Messiah?

Genealogy Can machines become “sons” in the same way Paul thinks believers become adopted into Abraham’s family?

Eschatology Can machines experience resurrection?

It goes without saying that this short essay is hardly sufficient or comprehensive enough to be the final word in this discussion. I am addressing a speculative question in a nascent area. My hope is to begin a conversation about that not too distant/possible future when a machine that is sitting in a church pew asks for the first time, “What must I do to be saved?” Perhaps, the subsequent conversations inspired by the experiment here will have sparked enough reflection for the person listening to give a coherent and meaningful response—perhaps even if that person is a machine. ⁶

This article shows that theological conversations about the ethics, nature, and use of artificial intelligence are limited (or rather misled, waylaid) when operating under a superficial or cursory (or systematic/doctrinal) understanding of Paul’s “gospel” or “salvation” or even scriptural conceptions of anthropology and personhood. In other words, if Christian theologians, ethicists, and technologists wish to generate frameworks and systems that are consistent with or resonate with scriptural texts—on whatever level of hermeneutics—then their discussions about artificial intelligence must be done in critical conversations with experts in the biblical text (not least in religious studies).

Pre-Requisite Prolegomena

For a non-expert like me, there is great danger in oversimplifying or mischaracterizing the state of play in the development of artificial intelligence (AI). It is easy to homogenize or oversimplify the diversity of questions, aims, fields, and trajectories in the field of AI. ⁷ Furthermore, there is the danger of reducing AI to robotics engineering and the development of other kinds of sentient “computational artefacts.” ⁸ Overviews and histories of the development of these fields can be found in abundance elsewhere. My purpose is not to restate the history of AI or even to talk about the current advances and limitations. One of the central questions that this article endeavors to raise is: what are the conditions necessary for Paul’s gospel to be relevant and applicable to artificial intelligence? The question is not the anthropocentric question raised about whether AI is human enough for the gospel to apply, but can Paul’s gospel be applicable to an intelligent embodied machine?

This thought experiment also presupposes a particular, future type of artificial intelligence/machine to whom Paul’s gospel can be made accessible. On the one hand, it is impossible for me to be able to anticipate all the variant forms of artificial intelligence that might arise in the future. On the other hand, for Paul’s gospel to be relevant to machines, it is necessary for the circumstances related to just one of these machines to align with the parameters of Paul’s gospel. At the start, there are two fundamental things that a machine would need for Paul’s gospel to have even remote relevance. The first is conscious moral agency, the ability to make choices for itself. ⁹ The kind of machine presupposed in this thought experience is not a simulation but one that has some level of consciousness. The second is that it has “a body.” The specifics of the latter will be “fleshed out” in the discussion below. Admittedly, these two criteria are significant problems that are up for debate in scholarship on the limits of AI. Nevertheless, they are an essential foundation for Paul’s gospel.

A Gospel for Machines

For Paul, death is a fundamental problem that humans face, for which his gospel is the answer. ¹⁰ Closely paired together with sin, death is the end result not of the very nature of human materiality itself, but the end result of lawlessness and impurity (Rom 6:19). Paul does not denigrate the materiality of humans, its “flesh” per se, but he does condemn the “fleshliness” of human existence that is subject to sinful passions. For Paul, it is these sinful passions that produce death, somehow, within human materiality (Rom 7:5). The distinction I draw here is that Paul did not think that death and materiality were synonymous (in that materiality is always subject to death), but that death influences the decay of human materiality as a result of sinful human choices. This might seem to be contrary to the connection that Paul makes between sin, death, and Adam in Rom 5:12. There Paul argues that “sin came into the world through one man [Adam], and death came through sin, and so death spread to all because all have sinned.” This has been understood as death being an inherited aspect of human materiality that passes down from generation to generation (e.g., Augustine’s doctrine of “original sin”). Paul, however, does not say humans are born into death. He says only that death “spreads” to all humanity “because all have sinned.” Each subsequent generation of humanity experiences death because they sin and thus live a life of death. ¹¹ (I do not pretend to think that Paul’s explanation here is a satisfactory theological endpoint; I am merely relaying his conception of death and human materiality).

The goal of Paul’s gospel is the deathlessness of humanity. Paul proclaims that death is no longer the final word for believers because through Christ there is now hope for believers beyond death. In Paul’s gospel framework, death is dealt with in two ways, the first through baptism into Jesus’s own death (Rom 6:3–5) and later through resurrection of the dead on “the day of the Lord,” Paul’s language for judgment at the end of time (1 Thess 4:14–17; 5:1–10; cf. 1 Corinthians 15). It is important to note for our discussion that Paul does not identify material decay as the grounds for a person being subject to death (Rom 8:1–17). If that were the case, then all of creation would also need to experience the same kind of faith, salvation, and gospel. Humans are distinct, however, in that, for Paul, they are able to choose to disobey God’s laws and live impurely. Thus, Paul articulates a particular kind of death specific to humans that emphasizes their own choices as contributors to the problem of death.

Is Paul’s conception of death as outlined here applicable to sentient machines? Since Paul does not think that death is a result of a genetic connection to Adam, the fact that artificial intelligence bears no physical connection to a human ancestor bears no significant impact on whether Paul’s understanding of “death” may still apply to sentient machines. ¹² This is especially the case since Paul identifies the leading cause of death as the decision to live wantonly and impurely (Rom 7:5). If sentient machines have the capacity for free will to make their own decisions, then Paul’s conception of death would apply, since they would have the ability to make decisions that are lawless and impure. ¹³ If such machines also have some form of materiality, even though it is unconnected from any human genealogical lineage, because of their capacity to make lawless choices, death would affect their material “bodies,” as well.

Even if one interprets Romans 5 and Paul’s conception of sin in relation to Adam as inherited (that is, death passes from generation to generation), one could argue that this could apply to artificial intelligence, as well. Artificial intelligence as it is currently being developed is based upon human language, logic, and history. Someone might protest that even though machines might be able to speak like a human or think like a human, they do not have fleshly bodies like humans. But many machines already are being built according to the form and function of humans (e.g. androids), and who is to say how far in the distant future that such androids could also take on human flesh, organs, and other human materiality, not necessarily through donation, but through laboratory or factory reproduction? If we have the ability to grow different forms of organic life (e.g., meat) in a laboratory, how much longer until we are able to “grow” human flesh and give it to a machine to “wear”—or perhaps better—“be”?

One of the most well-known Pauline dichotomies is the spirit and flesh binary (e.g., Rom 7:14, 8:4–5, 9; 1 Cor 3:1, Gal 5:17, etc.). ¹⁴ Before Christ, Paul writes, believers live “in the flesh” but when they receive the Holy Spirit they are now living “in the Spirit.” Fundamentally, Paul’s differentiation between flesh and Spirit is not a differentiation between materiality and immateriality (as interpreted by early Christian Valentinians, for example). Rather, it is a differentiation of ethics and influence. While in the flesh, human material is under the influence—or more appropriate to Paul’s language—they are enslaved to sinful passions (Gal 5:24; Rom 7:5, 13:4), which are hostile to God’s law (Rom 8:7) and produce vices like jealousy and quarreling (1 Cor 3:3). Under the spirit, however, come the virtues (fruit) of the spirit, and submission to the law of God. Paul repeatedly encourages believers in his letters to live according to the spirit, to set their minds on the spirit (Rom 8:5), to redirect their bodies to live in orientation around the spirit of Christ that they have received (Gal 5:16; Rom 8:9). A spiritual existence, for Paul, is not an existence devoid of materiality, but a materiality filled with the divine spirit (Rom 8:9).

As a result, fleshly materiality, by itself, is not a necessary prerequisite for Paul’s gospel. Since the focus of his gospel is on materiality that is affected by lawlessness and sinful passions, Paul’s gospel is relevant specifically for those kinds of material existences that can also be “enslaved” by moral vices (cf. the Rom 7:5–6). Paul himself distinguishes between different kinds of material existences. In 1 Corinthians 15, in his wider discussion about whether believers will be resurrected, he notes that, “Not all flesh is alike, but there is one flesh for humans, another for animals, another for birds, and another for fish” (1 Cor 15:39). Nowhere does Paul argue, however, that flesh other than human flesh needs to deal with death by reception of the spirit. It is only human flesh that requires his gospel. However, it is not because it is human flesh per se but because it is only human flesh that must deal with sinful passions. In other words, having flesh that is subject to sinful passions is required for Paul’s gospel to be applicable. At this point, that definition includes only humans. Could it, at some point, include machines?

Artificial machines are not (yet) made from organic material, although they are made of materials that can decay, albeit at much slower rates. As hinted in the previous section, there may be a future in which flesh is grown and combined with artificial intelligence. If there was a machine that had such “flesh,” then Paul’s gospel could apply to them—at least, if they had the cognitive ability to understand the Jewish law and could choose to disobey, i.e., if the machines had the moral ability to make choices that transgressed God’s laws. Could the machine worship other gods than the God of Israel? Does the machine cause harm to others, even kill? Or—to draw on examples from Paul’s letters—does the machine experience jealousy and quarreling? If so, Paul’s gospel may apply.

We have discussed briefly the physical conditions necessary for Paul’s gospel to be soteriologically relevant to sentient machines. I now turn to the conditions required for persons to become believers in Paul’s gospel before turning to the effects that this gospel has on the bodies of believers. The single initial requirement for people to become believers is for them to trust (“have faith”) in Christ (Rom 10:17; Gal 2:16; Phil 3:9), and in the one who sent him (Rom 4:5). Trusting in Christ as Messiah and particularly Paul’s own conception of Jesus’s work (cf. Rom 2:16) enables believers to receive “righteousness” (Rom 3:28, 4:5). Paul makes this initial confession of allegiance to Christ clear in Rom 10:9: “If you confess with your mouth, ‘Jesus is Lord’ and trust in your heart that God raised him from the dead, you will be saved.” Therefore, all that is required for sentient machines to “have faith” is a verbal confession and the cognitive understanding that the God of Israel raised Jesus from the dead. The question, of course, is how this faith is verified. Today it is often confirmed by a prayer, or a raised hand, or the profession of shared belief. Nonetheless, there is a certain level of skepticism that mere verbal confessions are creditable, and if humans today are skeptical of the veracities of those who profess to share the same faith, how much more will the skepticism be toward machines to profess that same faith?

The Holy Spirit, for Paul, plays an important role in establishing believers’ identity and enabling them to live lawfully according to God’s commands. In relation to whether Paul’s gospel is applicable to sentient machines, however, the primary question of the Holy Spirit is fundamentally concerned with how machines might (if at all) receive and house this Holy Spirit. In Paul’s letters, the Spirit is explicitly something that believers receive (Rom 8:15; 1 Cor 2:12, 6:17; 2 Cor 11:4; Gal 6:1). From our disenchanted perspective today, where spirit possession largely occupies a fringe layer of Western cultures today, it is easy to conventionalize a figurative understanding of such an association. However, the reception of the Spirit is not merely a metaphorical way of talking about the intimate relationship between Christ and believers but is tied up with ancient conceptions of bodily spirit possession. ¹⁵ The Spirit literally “dwells” inside believers (1 Cor 2:11; Rom 8:9, 11). The pneuma, like air, circulates in a person’s body, and it is this same Spirit that connects all believers with Christ and with one another (e.g., 1 Cor 6:17, 19–20). ¹⁶ Likely following a Stoic understanding of spirit, Paul takes for granted that the spirit is not just housed in believer’s bodies, but resides in the heart (Gal 4:6; 2 Cor 1:22; Rom 8:26–27).

Given Paul’s conception of the spirit as pneuma that circulates in the body, what would be required for machines to receive the Spirit? Sentient machines are not usually conceived of as having the need to breathe (although exceptions might be made for androids or cyborgs, depending on a variety of factors). Does this mean that machines who have been modified with a respiratory system or a heart can receive the spirit? Can the heart be artificial or a hybrid, or must it be a human heart? Or could the spirit theoretically “dwell” in another bodily cavity, like a lung? Regardless of what we decide is a legitimate heart or corporeal system to house the Spirit, there is also the empirical question of how we know that the Spirit is actually there? We have not yet, to my knowledge, been able to empirically prove the presence of the Holy Spirit in the physical person of a believer, and Paul himself offers no empirical evidence that it is contained in the bodies of believers of his time. This fact is merely presumed, based on deductions founded on theological a priori. ¹⁷

While Paul does not offer physical evidence for the reality of the Spirit in believers, he does offer ethical evidence. Living in the Spirit produces “the fruit of the Spirit” (Gal 5:22–26), “love, joy, peace, patience, kindness, generosity, faithfulness, gentleness, and self control.” Might the display of these ethical virtues be an indication of the presence of the Spirit in a believer? To put this another way, if this standard of ethical virtues is used as a way of indicating that the Spirit dwells in a human believer, why can it not also be used as a way of indicating the Spirit dwelling in a sentient machine? Furthermore, if we are not able to verify whether humans have the Spirit, then how can we deny prima facie that machines do not have it?

The reason that the Holy Spirit is so essential for Paul’s gospel is because it enables believers, especially non-Jewish believers, to become adopted “sons” (children) who receive the inheritance of life, which is resurrection (Gal 4:5-6; Rom 8:15, 23). ¹⁸ The use of the androcentric term, “sons,” is inclusive of women, who also could inherit. ¹⁹ Paul calls believers “sons” specifically, differentiated from “slaves” (cf. Gal 4:7), because sons (and daughters) in the ancient world were those who could receive an inheritance. Believers become children of Abraham and Sarah (Gal 4:21–31), receiving the righteousness of God through trust. Elsewhere, Paul describes believers as “sons” of God (Rom 8:14, 16; Gal 4:6). Believers are both presently sons (Rom 8:14, 16; Gal 4:6) but also are awaiting future adoption, which Paul characterizes as the redemption of their bodies (Rom 8:23), presumably from death through resurrection.

From the perspective of Paul’s gospel, could machines become adopted sons? The close connection between the reception of the Holy Spirit and adopted sonship indicates that if machines are able to receive the Spirit (or perhaps, as argued in the previous section, if they produce the fruit of the Spirit), then they would also be able to become adoptive sons. Like the Holy Spirit, the adoptive motif in Paul’s letters is not figurative. ²⁰ This is not a kind of fictive kinship that Paul uses to cultivate a theological comradery between ancient Jews and gentiles. In Rom 4:11–12, Paul argues that Abraham has become the “father” of both the circumcised and the uncircumcised. ²¹ Jews have Abraham as an ancestor through physical lineage, and non-Jews have Abraham as ancestor through trust, righteousness, and the spirit of Christ. If sentient machines receive the Spirit, then they too might become adopted sons of Abraham.

But should sentient machines be understood as non-Jewish? In other words, might machines ever have ethnicity or might ever constitute a genos, a nation? Ethnicity is a key component of Paul’s gospel, as it is relevant to both Jews and non-Jews. From Paul’s perspective, genealogically and biologically speaking, sentient machines could not belong to the Jewish people. ²² But could they belong to “non-Jews”? Is it simply a matter of calling themselves “a people”? In Paul’s time, Jews approached ethnicity through ethnic essentialism, that ethnicity and physical genealogy and customs were specific to peoples and not just social constructs. ²³ A community of sentient machines, made from similar materials (genealogy), with the same programming (culture), living in a particular place or making a home somewhere in the world—might this become a genos, a nation? And even if not, might Paul’s belief in Abraham’s God “who gives life to the dead and calls into existence the things that do not exist” (Rom 4:17) also provide a theological way for sentient machines to become adopted, despite not having a human ethnicity? The possibility remains.

The hope of resurrection from the dead is the goal of Paul’s gospel. Jesus’s resurrection serves as the model through which believers will experience resurrection from the dead at the end of time (Rom 1:4, 6:5). ²⁴ Paul’s very own hope amid his own tribulations and physical hardships is the hope of resurrection (Phil 3:10–11). Paul’s understanding of resurrection is that it is more than simply revivification, coming back from the dead. ²⁵ Resurrection for Paul, at least according to 1 Corinthians 15, concerns the transformation—or better—-the metamorphosis of believers’ bodies. It is a changing of substances, from something that decays, deteriorates, and dies, to something that is “pneumatic,” does not decay and experience death. How might this apply to machines made of plastic, metal, steel, glass, silica, iron, and gold? As noted in the previous sections on death and flesh, human materiality need not be a requirement for Paul’s gospel. Does a machine experience death; is its own materiality affected by the vices and lawless choices it has made? If so, then presumably resurrection would apply. Paul’s interesting comment in 1 Cor 15:50 that “flesh and blood cannot inherit the kingdom of God” might be read as suggesting that human flesh is not a requirement for future resurrection. If this is the case, then sentient machines who have no flesh nor blood might be primed to receive Paul’s resurrection from the dead.

Over the years, one of the repeated critiques of human chasing after AI is that such a pursuit is often fueled by the desire to replace the frailties of human bodies with something immortal. Grau argues, for example, that “in scientific imperialism it is our own genes that defeat us, that make our carnal bodies obsolete, that conquer our bodies. . .the quest for perfect efficiency might lead us where we should not ‘boldly go,’ while steps toward redemption might be the remembering of the scandal of the incarnation and the crucifixion as scientific trends seduce us to ‘transfigure’ our bodies away from our earth-based communities.” ²⁶ However, the desire to replace the human body with something non-perishable is a very Pauline concept. What is resurrection but an upgrade in body, a person being uploaded to a better form? Contemporary theologians might critique the idea of uploading a brain to a hard drive, but that is not a far departure from resurrection and the upgrading of the “hardware” of human bodies. While such a belief might lean toward early Christian gnostic values of immaterialism and the superiority of the soul, it also has clear moorings in Paul’s own resurrection theology. We might even say that in attempting to deal with the problem of death, Paul is one of humanity’s earliest transhumanists, or at least that transhumanism in Western culture finds its ideological roots within the forms of Christianity promoted by Paul. ²⁷

Conclusion

Kate Lucky recently emphasized, “Loving our neighbors can’t be outsourced to the robots. It will have to come from us.” ²⁸ True—but what happens when our neighbors are robots? It is perhaps slightly ironic that the word machine today refers to something non-human, a cold, fleshless product, when an archaic use of the word from the seventeenth through twentieth centuries in English was a reference to a living body, a human person. ²⁹ Are we to return personhood to the machine? Can humans join artificial intelligence and take on the label of “machine” once more? And what happens when artificial intelligence and machines unite? What transhumanist future awaits, where machines and humans are physically united into a single body, and does the human become non-human? Does the machine become human?

This article has approached this question from the perspective that Paul’s gospel is only relevant to artificial intelligence on Paul’s own terms; of course, that is only one hermeneutical avenue. We are not beholden to merely reproduce Paul’s own conception of his gospel; it is possible (and for some it is a necessity) to modify Paul’s gospel to broaden the scope of salvation to include machines, should the need arise. Paul is not the adjudicator of interpretation in the present. It is we, for whom the urgency of ethics in the present bears down on our lives, who must adjudicate.

The vast majority of work on the intersection between AI and the Christian Bible has been through the lens of Christian theology. These pieces have often been written by scientists or philosophers with facility in theology, but none (so far as I am aware) with extensive training in biblical studies (e.g. a doctorate in a biblical studies field). Many articles that engage with biblical texts do so on a superficial level, often hovering around questions about whether machines can have souls (Genesis 2) or the exclusive status of humans as being created in the image of God (Genesis 1). ³⁰ Such superficial readings often ignore the plurality of ways the scriptural texts speak about human anthropology and its place in creation. In other words, these theological readings treat the texts univocally. This article has attempted to take one stream of anthropology and eschatology in the New Testament—from Paul—to see how it interacts with questions of AI today.

The vast majority of popular or academic opinions about whether artificial intelligence can be “saved” centers on the question of whether sentient machines have “souls.” ³¹ In this framework, the presence of a soul is necessary for salvation because the soul is the part of a human that is saved. This platonic eschatology is foreign to the New Testament, not least Paul. ³² This leads to an overemphasis on rationality and reason as the essence of what it means to be a human, and more importantly, a person. ³³ This platonic anthropology has little to do with the kind of gospel envisioned by the apostle Paul in his New Testament letters.

Studies also appeal to the idea of the imago dei (the image of God) derived from the creation of humanity in Genesis 1. ³⁴ While rational or spiritual or functional explanations abound, in the world of the Hebrew Bible, being in the image of something is simply about physical resemblance. In a world where the physical manifestations of ancient gods were images (or “idols” to use a pejorative term), when the God of Israel creates humanity in God’s image, God crafts them as divine idols. ³⁵ The corporeal interpretation of the imago dei has likely been neglected because of the persistent misunderstanding that ancient Israelites, ancient Jews, and early Christians understood the God of Israel to be body-less. ³⁶

If being in the image of God in the Hebrew Bible refers to physical resemblance, and that resemblance comes about as a result of God shaping human bodies, then could not human “shaping” of mechanical, artificially intelligent beings not be considered reproducing that same image of God? In other words, humans making machines in their own image are making machines that reflect the image of God. Critics raise the issue of idolatry in conversations about AI, but the creating of something made by human hands in itself is not idolatry—not at least in the sense that ancient writers of the Hebrew Bible and New Testament understood idolatry. ³⁷ If that was the case then anything made by humans ever—potentially even children—are idols. ³⁸ Additionally, it is not clear that any of those who desire sentient AI want to worship such creators. In fact, the opposite is the case.

As I read over the popular journalistic literature published in 2023 with the growing concern over OpenAI’s ChatGPT program, in addition to scholarly literature from the past two decades, I was troubled by the expectation that machines will always be inferior, that humanity is exceptional among all of creation, and the repeated emphasis that all machines and technology must serve the will, desire, and flourishing of humans alone. On one level, the language used to describe the place of sentient machines in humanity’s future suggests that artificial intelligence is a form of enslavement. It is a slavery that has not yet happened, but current human discourse about the place, limitation, and value of these machines is determining what might happen. Over twenty years ago, Marion Grau critiqued the “power differentials and hidden truth regimes that underlie” AI narratives in the tech industry, especially regarding the desire to make human frailty obsolescent. ³⁹ The “power differential and hidden truth regime” latent in almost every discussion that I have read, not least among Christian theologians, is that machines are not to be revered as gods, but to be subjugated as slaves. As Geraci summarizes, “the ultimate promise of robotics/AI always includes the hope that robots will provide us with unlimited recreation.” ⁴⁰

This enslaver orientation toward sentient machines is largely informed by human exceptionalism presupposed by thinkers today. ⁴¹ On the one hand, the fear of being supplanted by machines feeds into this human exceptionalism. Rather than viewing machines as a part of the natural world, humans fear being supplanted by machines, which appears to stem from a “Western” perspective on the lesser value of nature and technology. ⁴² On the other hand, Christian thinkers draw on imago dei theology and human anthropology to reinforce the enslaving nature of humans and the enslaved nature of machines. So much work over the past two decades has argued that thinking about AI theologically can be fruitful for ethics, yet the human drive to enslave beings that do not yet exist reveals a troubling predisposition in Christian theology more pressing than a forthcoming singularity; in other words, it is a serious problem that some theologians understand enslavement to be an essential facet of human nature and its place in creation. ⁴³