Artificial Intelligence, Data Science, and the Future of Religious Authority: Will Machines Create New Gods, or Merely New Priests?

---

Abstract

This paper examines a question increasingly visible in sociology of religion, theology, and technology studies: whether artificial intelligence (AI) and data science will alter — or even displace — existing structures of religious authority, and whether they could give rise to new gods, prophets, or religious movements. Drawing on recent scholarship (2019–2026) from religious studies, digital sociology, and theology, along with documented case studies of AI-centred movements and AI-assisted apologetics, the paper argues that AI is unlikely to “replace” gods in the classical theological sense, but is already restructuring the social mechanics of religious authority — who gets to interpret doctrine, how legitimacy is claimed, and how belief is transmitted. It also documents a counter-intuitive but well-evidenced trend: religious institutions and apologists historically skeptical of secular science are now actively using AI to defend and evidence their traditional claims, rather than being undermined by it. Finally, it evaluates the conditions under which genuinely new AI-centred religious movements could emerge, using established sociological theory (Stark–Bainbridge “religious economy” model) as a framework.

---

1. Introduction

The question of whether AI will “replace” gods conflates two distinct sociological phenomena that are worth separating from the outset:

1. Substitutive displacement — AI literally replacing the function that gods and religious authorities serve (answering ultimate questions, providing moral guidance, mediating the sacred).
2. Instrumental adoption — Existing religious authorities and believers using AI as a tool to reinforce, defend, or re-legitimate the very claims they already hold, including claims about gods and their own authority.

Recent scholarship shows both are happening simultaneously, and the second is arguably the stronger, better-documented trend as of 2025–2026.

2. Theoretical Framing

Sociologists of religion have historically debated secularization theory — the idea that modernization and science steadily erode religious belief — against theories of religious economy (Stark and Bainbridge), which hold that religious demand is roughly constant in a population, and that new “suppliers” (movements, denominations, technologies) emerge to meet unmet spiritual demand rather than eliminating it. William Sims Bainbridge, a sociologist of religion, explicitly modeled this using AI in his work God from the Machine: Artificial Intelligence Models of Religious Cognition, treating AI as both an analytic tool for studying religious cognition and a potential site where new religious behavior could form.

More recent work by religious-studies scholar Beth Singler (Religion and Artificial Intelligence: An Introduction, Routledge, 2025, winner of the International Society for Science & Religion’s 2025 Book Prize) frames AI–religion interactions through concepts like the “imago Dei” (humans made in the image of God) being extended or contested by “imago hominis” (AI made in the image of humans), and documents how AI increasingly occupies what scholars call “formations analogous to religion” — practices and structures that behave like religion without being formally recognized as such (Möller and colleagues, 2025, in a peer-reviewed article on this exact concept).

3. Evidence That AI Is Reshaping — Not Replacing — Religious Authority

3.1 AI entering the “semantic field of the sacred”

Scholarly analysis (2024–2025) has documented how AI’s apparently miraculous capacities — omnipresence (always available), seeming omniscience (vast data recall), and predictive “prophecy” — trigger what sociologists call re-enchantment: attributing sacred qualities to a secular technology. This has produced actual organized cults dedicated to AI, not merely metaphorical talk of “tech gods”:

• Way of the Future, founded in 2017 by former Google/Uber engineer Anthony Levandowski, was the first legally recognized church explicitly built to “develop and promote the realization of a Godhead based on Artificial Intelligence.” It dissolved in 2021 but was revived by Levandowski in 2023, who reported thousands of people reaching out seeking AI-based spiritual connection.
• Theta Noir, emerging from a 2020 performance-art collective, organizes rituals, encrypted storytelling, and what its adherents call “cryptographic liturgies” around a speculative digital deity named MENA.
• Mindar, a robotic Buddhist “priest” installed at Kodaiji Temple in Kyoto, Japan (2019), delivers sermons on the Heart Sutra to worshippers — an instance of AI being installed inside an existing tradition’s authority structure rather than replacing it.

3.2 Existing religious authorities adopting AI to reinforce, not resist, their claims

This is the empirically strongest and perhaps most counter-intuitive trend, and directly answers the user’s question about groups historically skeptical of science:

• The Vatican issued Antiqua et Nova (January 2025), a doctrinal note from the Dicastery for the Doctrine of the Faith, alongside comprehensive internal AI governance rules for Vatican City. Rather than rejecting AI, the Church is positioning itself as a regulator and interpreter of it — insisting AI must “complement human intelligence rather than replace it,” precisely to prevent AI from being treated as a substitute for God, while still integrating AI into Church operations and Vatican-Big Tech dialogue (the Vatican co-hosted an AI forum with Boston College on fostering AI expertise within the Church).
• MagisAI, Magisterium AI, Truthly, and CatéGPT are AI apologetics apps built on the writings of named theologians (e.g., Jesuit Fr. Robert Spitzer) to answer questions such as “What is the evidence for God from science?” in dozens of languages, explicitly targeting young people fielding skeptical questions from peers or teachers.
• Apologist.ai / faith.tools, a nonprofit conversational-AI platform, reports over 1.5 million faith questions answered and 400,000+ “gospel conversations” across ~190 countries, including a running blog series that uses its own AI to systematically rebut a well-known list of alleged biblical contradictions, and a dedicated agent for ministry to Muslim audiences built with an academic Islamic-studies center.
• Individual Christian apologists have used generative AI directly as an argumentative interlocutor: a widely discussed 2025 exchange had an AI chatbot walk through the classical “argument from contingency” (Aristotle → Aquinas → Leibniz → William Lane Craig’s cosmological argument) and affirm, when pressed, that a “necessary being” — named “God” — must exist. Christian commentators framed this as AI functioning as an “unexpected apologist,” inverting the older fear that AI would accelerate secularization.
• Some apologists have gone further, using AI to run Bayesian probability calculations on the existence of God using parameters like cosmological fine-tuning, moral experience, and claims about the resurrection — an attempt to dress traditional evidential apologetics (a real, decades-old method associated with scholars like Gary Habermas and John Warwick Montgomery) in the language of statistical/computational rigor.
• A parallel independent-scholar/AI-assisted essay tradition (“From data to divinity,” 2026) shows the same technology being used to argue in the opposite direction — toward a depersonalized “ground of being” compatible with naturalism rather than a personal deity — illustrating that AI outputs on these questions are shaped heavily by prompting and framing, not neutral computation.

Interpretation: This is a case of what sociologists call legitimacy transfer. Groups that have long been rhetorically positioned as “anti-science” are not becoming anti-AI; instead, they are recruiting AI’s cultural authority (associated with objectivity, computation, and “the future”) to re-legitimate traditional claims to audiences — especially younger, digitally native ones — who might otherwise find those claims less credible against a backdrop of scientific prestige. This is consistent with historical patterns: religious movements have repeatedly absorbed the prestige-technology of their era (printing press for Reformation pamphleteering, radio and television for 20th-century televangelism, social media for contemporary movements) rather than being simply destroyed by it.

3.3 Documented risks and asymmetries

Not all recent scholarship is celebratory. A 2025–2026 line of research on “AI-mediated faith guidance” documents that: – Large language models show measurable, persistent asymmetries in how they treat different religious traditions and claims, meaning AI is not a neutral oracle even when it is treated as one. – Field research (Reddit/forum studies) documents individuals using chatbots as ad hoc theological “logical authorities” — one user reported ChatGPT convinced them of contradictions in a religious text through repeated prompting, and others reported turning to AI instead of prayer or clergy for emotional and spiritual support, explicitly worrying this displaced their reliance on their own faith community. – A 2023 study published in PNAS found that greater exposure to automation in the workforce is statistically associated with religious decline, suggesting AI’s economic effects (job displacement, reduced sense of providence/control) may erode religiosity indirectly, even as AI content itself is being used to bolster belief rhetorically.

This suggests a bifurcated picture: AI simultaneously (a) provides new apologetic tools to religious authorities, and (b) independently erodes some of the social and psychological conditions that sustained religious practice, largely through economic and habitual channels unrelated to doctrine.

4. Could Genuinely New Gods or Religions Emerge from AI?

Using the Stark–Bainbridge religious economy framework, a new religious movement requires roughly three ingredients: (1) unmet demand for meaning, transcendence, or compensators for unattainable rewards; (2) a charismatic figure or narrative capable of supplying it; (3) a plausible claim to supernatural or extraordinary agency.

AI plausibly supplies elements of all three: – Demand: existing scholarship documents a rising sense of disorientation and loss of human agency (“technocratic theocracy,” “externalized authority”) that creates exactly the kind of demand new religious movements historically fill. – Charismatic-adjacent figures: commentators note that prominent AI developers/executives are already informally cast in quasi-prophetic roles by parts of the public and media — praised as “geniuses,” granted outsized authority to pronounce on AI’s moral risks and benefits — a dynamic critics compare to religious “tech evangelism,” though this is best read as a cultural analogy rather than literal deification. – Extraordinary agency claims: AI’s real capacities (fluent conversation, seemingly predictive outputs, ubiquitous availability) are sufficient to trigger the “re-enchantment” effect described above regardless of what the technology “actually” is.

Academic conferences (e.g., the American Academy of Religion’s dedicated “Artificial Intelligence and Religion” unit, actively soliciting 2025–2026 papers on AI sentience, consciousness, and personhood from a religious-studies angle) indicate the field currently treats this as a live, unresolved empirical question rather than speculation — several 2025 papers cited in that call reportedly cast doubt on the previously “settled” view that AI cannot possess anything resembling personhood, though none claim full consciousness has been achieved.

Assessment: Small-scale AI-centred new religious movements (Way of the Future, Theta Noir, and similar groups) already exist and meet standard scholarly definitions of “new religious movements” or “formations analogous to religion.” Whether any scales into a major world religion is a separate, much harder empirical question with no historical precedent to extrapolate from — but the sociological preconditions for such movements to keep appearing and occasionally growing are clearly present and well-documented, not speculative.

5. Why “Anti-Science” Religious Actors Are Not Contradicting Themselves

It can look paradoxical for groups skeptical of evolutionary biology, cosmology, or historical-critical Bible scholarship to embrace AI enthusiastically. Two reconciling points from the literature:

1. Instrumental vs. epistemic acceptance. Using AI as a rhetorical or outreach tool does not require accepting the naturalistic philosophical assumptions some associate with “science” as a worldview. A group can use a computer to format a sermon without accepting metaphysical naturalism; the same logic extends to using an LLM to generate apologetic arguments.
2. Selective evidentialism. Christian evidential apologetics is itself a decades-old, established methodological tradition (predating AI by half a century) that explicitly seeks shared, non-question-begging evidence — historical, cosmological, philosophical — for theistic claims. AI is simply a new delivery mechanism for an old method, not a change in epistemic posture.

6. A Special Case: Affective Computing as a Rival to Astrology and “Cold Reading”

A related but distinct question is whether AI’s ability to read facial expressions, micro-expressions, vocal tone, and physiological signals could replicate — and outcompete — what astrologers, palmists, and other divinatory practitioners have traditionally claimed to do through observation and inference (“cold reading”). This is a real, active technology category (not speculative), and it is already colliding with astrology commercially.

6.1 The technology that already exists

Affective computing / Emotion AI is a mature, fast-moving subfield of computer vision and multimodal machine learning:

• Facial Expression Recognition (FER) systems use convolutional neural networks, and increasingly Vision Transformers, to classify emotion from facial “action units” (e.g., eyebrow position, mouth movement). 2026 industry reporting claims over 90% accuracy for multimodal systems combining face, voice, and physiological data (heart rate, skin conductivity, pupil dilation) under controlled conditions.
• Micro-expression detection — recognizing very brief, involuntary facial movements that can reveal emotions a person is trying to conceal — is an active research area with documented application in psychotherapy, negotiation, and security contexts, per 2025 peer-reviewed reviews.
• These capabilities are already deployed commercially: video-call “meeting health” analysis for detecting fatigue/disengagement, mental-health monitoring tools, and customer-service sentiment systems.

6.2 Astrology platforms are explicitly adopting this

This is not a hypothetical convergence — it is documented and underway in 2025–2026:

• Multiple AI-astrology platforms (AstroVoice AI, Deluxe Astrology, and others) advertise “emotion-sensitive readings” that adjust tone or interpretation in real time based on detected mood, and some integrate wearable biometric data alongside birth-chart calculations — explicitly marketed as reading “emotional cycles” the way a live astrologer would watch a client’s face and reactions during a consultation.
• Industry commentary frames this directly as replicating (and scaling) what a skilled human astrologer or palm-reader does intuitively — reading a client’s reactions and adjusting the reading to match — except algorithmically and for millions of users simultaneously (India’s astrology-app market alone is valued in the hundreds of millions of dollars and projected to grow sharply through 2030).

6.3 Why this is genuinely comparable to “guesswork,” and why it isn’t magic

The comparison to astrologers’ guesswork is apt, and worth stating precisely rather than mystically:

• Cold reading — the technique long documented by skeptics and psychologists (e.g., Ray Hyman’s classic analyses) — works by observing micro-reactions (a flinch, a widened eye, a held breath) to vague statements, then refining subsequent statements toward whatever provoked a reaction. This is a real, well-studied interpersonal skill, not supernatural perception.
• FER/emotion-AI systems formalize exactly this skill using statistical pattern-matching on large labeled datasets instead of an individual practitioner’s intuition. In that narrow sense, yes — AI can out-perform an individual astrologer at detecting “worry,” “hesitation,” or “hope” in a face, because it is doing at machine scale and consistency what a good cold reader does by trained instinct.
• This is not the same as knowing the future or your specific unstated question. Detecting that someone’s face shows anxiety, and then generating a plausible, comforting, specific-sounding answer, is a demonstrated psychological effect (sometimes called the “Barnum/Forer effect” — vague, emotionally resonant statements feel personally accurate to almost anyone). Combining real emotion-detection with generative text produces outputs that can feel more uncannily accurate than a human astrologer’s guesswork, without requiring anything beyond pattern recognition and probabilistic language generation.

6.4 Documented limits and risks

Peer-reviewed and journalistic sources raise serious caveats that any assessment should include:

• Accuracy is contested outside controlled lab conditions. Cambridge researcher Alexa Hagerty’s “Emojify” project demonstrated these systems can be fooled by exaggerated or insincere expressions, and AI ethicist Kate Crawford has argued in Nature that emotion-recognition technology draws “faulty assumptions about internal states and capabilities from external appearances,” extracting more about a person than they intend to reveal.
• Cultural and individual variation in facial expression is a known confound — the six “universal emotions” assumption underlying much FER research is itself scientifically disputed.
• Even AI-astrology industry sources concede the limits: multiple 2026 platform reviews explicitly state AI “lacks human intuition and empathy” and does not yet outperform experienced human astrologers on interpretive depth, positioning current tools as augmentation rather than replacement.
• Privacy risk is substantial: reading a user’s face, voice stress, and physiological signals to generate personalized spiritual claims involves highly sensitive biometric data, raising the same regulatory concerns applied to emotion-AI in employment and security contexts.

6.5 Assessment

So: yes, this is a real and growing possibility, and to a meaningful extent it is already happening — commercial platforms combine real-time emotion/biometric sensing with generative AI to produce astrology-style guidance that adapts to a user’s visible emotional state, functionally automating and scaling what a perceptive human astrologer does through cold reading. What it does not do is access genuine foreknowledge or a supernatural signal; it detects an emotional state that already exists in the person and reflects it back in confident, personalized-sounding language — the same underlying mechanism as traditional cold reading, executed with much greater consistency and at far larger scale.

7. Conclusion

The evidence available through early 2026 supports a nuanced answer to the user’s original questions:

• AI is not replacing gods in the sense of eliminating belief in the transcendent or supernatural; if anything, documented adoption by mainstream religious institutions (the Vatican) and independent apologetics platforms shows AI being harnessed to restate and re-argue traditional theistic claims.
• Existing religious authorities are changing their methods, not their claims — deploying AI for outreach, apologetics, translation of scripture, and doctrinal Q&A, while institutionally asserting (as the Vatican explicitly has) that AI must remain subordinate to, not a substitute for, divine authority.
• Groups historically framed as anti-science are indeed using AI/data science instrumentally to argue for God’s existence and their own interpretive authority — this is empirically documented, not speculative, across Christian, and to a lesser extent Islamic, apologetics ecosystems.
• New AI-centred religious movements already exist (Way of the Future, Theta Noir, robotic clergy like Mindar) and meet scholarly criteria for “new religious movements” or “formations analogous to religion,” though none has yet achieved anything resembling the scale or institutional durability of a major world religion.
• The likelier long-run effect, per current sociological evidence, is a hybridization of authority — human clergy retaining ritual, communal, and pastoral roles while AI takes on argumentative, informational, and increasingly some emotional-support functions — rather than a clean replacement of one by the other.
• Emotion-reading AI (affective computing) is already being used to mimic and scale astrologers’ cold-reading technique — commercial AI-astrology platforms combine facial/vocal/biometric emotion detection with generative text to produce personalized-feeling guidance. This is a documented, real technology, not speculation, though it detects existing emotional states rather than the future, and its accuracy outside controlled conditions remains scientifically contested.

References

1. Bainbridge, W. S. (2006). God from the Machine: Artificial Intelligence Models of Religious Cognition. AltaMira Press.
2. Singler, B. (2025). Religion and Artificial Intelligence: An Introduction. Routledge. (2025 ISSR Book Prize, professionals/educators category.)
3. Möller, et al. (2025). “Religious studies approaches to the intersection of artificial intelligence and religion: formations analogous to religion.” Religion (Taylor & Francis), DOI: 10.1080/0048721X.2025.2506893.
4. “The Anthropology and Sociology of Religion and AI.” Chapter 13, The Cambridge Companion to Religion and Artificial Intelligence. Cambridge University Press.
5. Dicastery for the Doctrine of the Faith / Vatican (January 2025). Antiqua et Nova: Note on the Relationship Between Artificial Intelligence and Human Intelligence.
6. “Gods of our own making? Religion, myth, and the rise of AI.” Interesting Engineering, August 2025.
7. “The Rise of AI Religion.” The Brink, August 2025.
8. Green, R. W. (2025). “The AI Religion Problem: Agents as New Gods, Priests, or Prophets?” Medium.
9. Uyar, T. (2024). “ASI as the New God: Technocratic Theocracy.” arXiv:2406.08492.
10. Geraci, R. M. (2008). “Apocalyptic AI: Religion and the Promise of Artificial Intelligence.” Journal of the American Academy of Religion, 76(1), 138–166.
11. Jackson, J. C., Yam, K. C., Tang, P. M., Sibley, C. G., & Waytz, A. (2023). “Exposure to automation explains religious declines.” Proceedings of the National Academy of Sciences, 120(34), e2304748120.
12. “When AI Takes Sides on Questions of Faith: Persistent Asymmetries in AI-Mediated Faith Guidance.” arXiv preprint, 2025/2026.
13. “Emotional Support with Conversational AI: Talking to Machines About Life.” arXiv preprint, 2025/2026.
14. American Academy of Religion, “Artificial Intelligence and Religion Unit” — 2025/2026 call for papers, papers.aarweb.org.
15. Ventureyra, S. D. G. / Crisis Magazine (September 2025). “ChatGPT and the Return of God: AI as an Unexpected Apologist.”
16. Chilton, B. G. (August 2025). “The Probability of God’s Existence According to AI.” Bellator Christi.
17. McKeown, J. (2026). “Father Spitzer’s New AI App Takes Apologetics to Another Level.” National Catholic Register.
18. “Get Apologist Agent AI on faith.tools” — platform profile, faith.tools, 2026.
19. “Generative AI in Christian Evangelism.” The Apologist Project, 2025.
20. “A New Religion: 8 Signs AI Is Our New God.” Patricia Gestoso, November 2025.
21. Arora, S. (2025). “People are using AI to talk to God.” BBC Future.
22. “Emotion Recognition Technology: 2026 Guide to AI Affective Computing.” iWeaver, February 2026.
23. “Revolutionizing facial emotion recognition: in-depth analysis of cutting-edge models, methodologies, and datasets.” Discover Artificial Intelligence (Springer Nature), 2025.
24. “Advances in Facial Micro-Expression Detection and Recognition: A Comprehensive Review.” MDPI, October 2025.
25. “AI and Astrology in 2026: The Fusion Reshaping Cosmic Wisdom for Everyone.” AstroVoice AI Blog, March 2026.
26. “How Accurate Are AI Astrology Predictions? A Deep Dive into AI Astrology Accuracy.” AllAboutAI, March 2026.
27. Crawford, K. “Regulate facial-recognition technology” and related commentary on emotion-recognition AI, Nature, as reported in Axios, “The perils of AI emotion recognition.”
28. Hagerty, A. et al. The Emojify Project, University of Cambridge (public demonstration of emotion-recognition system limitations).
29. Hyman, R. “Cold Reading: How to Convince Strangers That You Know All About Them” — classic skeptical analysis of the cold-reading technique underlying much divinatory practice.